baked-lib wav_channel to base

This reverts commit 1c9d646a50.

.gitignore

@@ -1,6 +1,15 @@
 **/*.sw*
 /whisper-cpp-2023/rust.d/target
+/rust-whisper.d/target
+/rust-whisper.d/models
+/target/
+/models/
 snowboy-2022/snowboy
 **/*.git.d
 **/*.wav
 snowboy-2022/Dockerfile
+**/target
+/candle-wasm-examples-whisper/mel_filters*
+/candle-wasm-examples-whisper/whisper-tiny*
+/candle-wasm-examples-whisper/quantized*
+/candle-wasm-examples-whisper/audios*

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "rust-whisper.d/gitea-whisper-rs"]
+	path = gitea-whisper-rs
+	url = https://gitea.inhome.blapointe.com/bel/whisper-rs.git

README.md

@@ -1,29 +0,0 @@
-# stt
-
-## listen on linux
-
-https://wiki.archlinux.org/title/PulseAudio/Examples
-
-```
-10. ALSA monitor source
-
-To be able to record from a monitor source (a.k.a. "What-U-Hear", "Stereo Mix"), use pactl list to find out the name of the source in PulseAudio (e.g. alsa_output.pci-0000_00_1b.0.analog-stereo.monitor). Then add lines like the following to /etc/asound.conf or ~/.asoundrc:
-
-pcm.pulse_monitor {
-  type pulse
-  device alsa_output.pci-0000_00_1b.0.analog-stereo.monitor
-}
-
-ctl.pulse_monitor {
-  type pulse
-  device alsa_output.pci-0000_00_1b.0.analog-stereo.monitor
-}
-
-Now you can select pulse_monitor as a recording source.
-
-Alternatively, you can use pavucontrol to do this: make sure you have set up the display to "All input devices", then select "Monitor of [your sound card]" as the recording source.
-```
-
-```bash
-$ pactl list | grep -A 50 RUNNING | grep -E 'RUNNING|Name:|Monitor Source:' | grep Monitor.Source | head -n 1 | awk '{print $NF}'
-```

candle-wasm-examples-whisper/Cargo.toml

@@ -0,0 +1,52 @@
+[package]
+name = "candle-wasm-example-whisper"
+version.workspace = true
+edition.workspace = true
+description.workspace = true
+repository.workspace = true
+keywords.workspace = true
+categories.workspace = true
+license.workspace = true
+
+[dependencies]
+candle = { path = "../../candle-core", version = "0.3.2", package = "candle-core" }
+candle-nn = { path = "../../candle-nn", version = "0.3.2" }
+candle-transformers = { path = "../../candle-transformers", version = "0.3.2" }
+num-traits = { workspace = true }
+tokenizers = { workspace = true, features = ["unstable_wasm"] }
+
+# App crates.
+anyhow = { workspace = true }
+log = { workspace = true }
+rand = { workspace = true }
+serde = { workspace = true }
+serde_json = { workspace = true }
+wav = { workspace = true }
+safetensors = { workspace = true }
+
+# Wasm specific crates.
+getrandom = { version = "0.2", features = ["js"] }
+gloo = "0.8"
+js-sys = "0.3.64"
+wasm-bindgen = "0.2.87"
+wasm-bindgen-futures = "0.4.37"
+wasm-logger = "0.2"
+yew-agent = "0.2.0"
+yew = { version = "0.20.0", features = ["csr"] }
+
+[dependencies.web-sys]
+version = "0.3.64"
+features = [
+  'Blob',
+  'Document',
+  'Element',
+  'HtmlElement',
+  'Node',
+  'Window',
+  'Request',
+  'RequestCache',
+  'RequestInit',
+  'RequestMode',
+  'Response',
+  'Performance',
+]

candle-wasm-examples-whisper/README.md

@@ -0,0 +1,68 @@
+## Running Whisper Examples
+
+Here, we provide two examples of how to run Whisper using a Candle-compiled WASM binary and runtimes.
+
+### Pure Rust UI
+
+To build and test the UI made in Rust you will need [Trunk](https://trunkrs.dev/#install)
+From the `candle-wasm-examples/whisper` directory run:
+
+Download assets:
+
+```bash
+# mel filters
+wget -c https://huggingface.co/spaces/lmz/candle-whisper/resolve/main/mel_filters.safetensors
+# Model and tokenizer tiny.en
+wget -c https://huggingface.co/openai/whisper-tiny.en/resolve/main/model.safetensors -P whisper-tiny.en 
+wget -c https://huggingface.co/openai/whisper-tiny.en/raw/main/tokenizer.json -P whisper-tiny.en
+wget -c https://huggingface.co/openai/whisper-tiny.en/raw/main/config.json -P whisper-tiny.en
+# model and tokenizer tiny multilanguage
+wget -c https://huggingface.co/openai/whisper-tiny/resolve/main/model.safetensors -P whisper-tiny
+wget -c https://huggingface.co/openai/whisper-tiny/raw/main/tokenizer.json -P whisper-tiny
+wget -c https://huggingface.co/openai/whisper-tiny/raw/main/config.json -P whisper-tiny
+
+#quantized 
+wget -c https://huggingface.co/lmz/candle-whisper/resolve/main/model-tiny-en-q80.gguf -P quantized
+wget -c https://huggingface.co/lmz/candle-whisper/raw/main/tokenizer-tiny-en.json -P quantized
+wget -c https://huggingface.co/lmz/candle-whisper/raw/main/config-tiny-en.json -P quantized
+
+
+
+# Audio samples
+wget -c https://huggingface.co/datasets/Narsil/candle-examples/resolve/main/samples_gb0.wav -P audios
+wget -c https://huggingface.co/datasets/Narsil/candle-examples/resolve/main/samples_a13.wav -P audios
+wget -c https://huggingface.co/datasets/Narsil/candle-examples/resolve/main/samples_gb1.wav -P audios
+wget -c https://huggingface.co/datasets/Narsil/candle-examples/resolve/main/samples_hp0.wav -P audios
+wget -c https://huggingface.co/datasets/Narsil/candle-examples/resolve/main/samples_jfk.wav -P audios
+wget -c https://huggingface.co/datasets/Narsil/candle-examples/resolve/main/samples_mm0.wav -P audios
+
+```
+
+Run hot reload server:
+
+```bash
+trunk serve --release --public-url / --port 8080
+```
+
+### Vanilla JS and WebWorkers
+
+To build and test the UI made in Vanilla JS and WebWorkers, first we need to build the WASM library:
+
+```bash
+sh build-lib.sh
+```
+
+This will bundle the library under `./build` and we can import it inside our WebWorker like a normal JS module:
+
+```js
+import init, { Decoder } from "./build/m.js";
+```
+
+The full example can be found under `./lib-example.html`. All needed assets are fetched from the web, so no need to download anything.
+Finally, you can preview the example by running a local HTTP server. For example:
+
+```bash
+python -m http.server
+```
+
+Then open `http://localhost:8000/lib-example.html` in your browser.

candle-wasm-examples-whisper/build-lib.sh

@@ -0,0 +1,2 @@
+cargo build --target wasm32-unknown-unknown --release
+wasm-bindgen ../../target/wasm32-unknown-unknown/release/m.wasm --out-dir build --target web

candle-wasm-examples-whisper/index.html

@@ -0,0 +1,40 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="utf-8" />
+    <title>Welcome to Candle!</title>
+    <link data-trunk rel="copy-file" href="mel_filters.safetensors" />
+    <!-- samples -->
+    <link data-trunk rel="copy-dir" href="audios" />
+    <!-- tiny.en -->
+    <link data-trunk rel="copy-dir" href="whisper-tiny.en" />
+    <!-- tiny -->
+    <link data-trunk rel="copy-dir" href="whisper-tiny" />
+    <!-- quantized -->
+    <link data-trunk rel="copy-dir" href="quantized" />
+
+    <link
+      data-trunk
+      rel="rust"
+      href="Cargo.toml"
+      data-bin="app"
+      data-type="main" />
+    <link
+      data-trunk
+      rel="rust"
+      href="Cargo.toml"
+      data-bin="worker"
+      data-type="worker" />
+
+    <link
+      rel="stylesheet"
+      href="https://fonts.googleapis.com/css?family=Roboto:300,300italic,700,700italic" />
+    <link
+      rel="stylesheet"
+      href="https://cdnjs.cloudflare.com/ajax/libs/normalize/8.0.1/normalize.css" />
+    <link
+      rel="stylesheet"
+      href="https://cdnjs.cloudflare.com/ajax/libs/milligram/1.4.1/milligram.css" />
+  </head>
+  <body></body>
+</html>

candle-wasm-examples-whisper/lib-example.html

@@ -0,0 +1,351 @@
+<html>
+  <head>
+    <meta content="text/html;charset=utf-8" http-equiv="Content-Type" />
+    <title>Candle Whisper Rust/WASM</title>
+  </head>
+  <body></body>
+</html>
+
+<!DOCTYPE html>
+<html>
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <style>
+      @import url("https://fonts.googleapis.com/css2?family=Source+Code+Pro:wght@200;300;400&family=Source+Sans+3:wght@100;200;300;400;500;600;700;800;900&display=swap");
+      html,
+      body {
+        font-family: "Source Sans 3", sans-serif;
+      }
+    </style>
+    <script src="https://cdn.tailwindcss.com"></script>
+    <script type="module">
+      // base url for audio examples
+      const AUDIO_BASE_URL =
+        "https://huggingface.co/datasets/Narsil/candle-examples/resolve/main/";
+
+      // models base url
+      const MODELS = {
+          tiny_multilingual: {
+            base_url: "https://huggingface.co/openai/whisper-tiny/resolve/main/",
+            model: "model.safetensors",
+            tokenizer: "tokenizer.json",
+            config: "config.json",
+            size: "151 MB",
+          },
+          tiny_en: {
+            base_url:
+              "https://huggingface.co/openai/whisper-tiny.en/resolve/main/",
+            model: "model.safetensors",
+            tokenizer: "tokenizer.json",
+            config: "config.json",
+            size: "151 MB",
+          },
+          tiny_quantized_multilingual_q80: {
+            base_url: "https://huggingface.co/lmz/candle-whisper/resolve/main/",
+            model: "model-tiny-q80.gguf",
+            tokenizer: "tokenizer-tiny.json",
+            config: "config-tiny.json",
+            size: "41.5 MB",
+          },
+          tiny_en_quantized_q80: {
+            base_url: "https://huggingface.co/lmz/candle-whisper/resolve/main/",
+            model: "model-tiny-q80.gguf",
+            tokenizer: "tokenizer-tiny-en.json",
+            config: "config-tiny-en.json",
+            size: "41.8 MB",
+          },
+          distil_medium_en: {
+            base_url:
+              "https://huggingface.co/distil-whisper/distil-medium.en/resolve/main/",
+            model: "model.safetensors",
+            tokenizer: "tokenizer.json",
+            config: "config.json",
+            size: "789 MB",
+          },
+        };
+
+      const modelEl = document.querySelector("#model");
+
+      Object.keys(MODELS).forEach((modelID) => {
+        const model = MODELS[modelID];
+        const option = document.createElement("option");
+        option.value = modelID;
+        option.textContent = `${modelID} (${model.size})`;
+        modelEl.appendChild(option);
+      });
+      const whisperWorker = new Worker("./whisperWorker.js", {
+        type: "module",
+      });
+
+      async function classifyAudio(
+        weightsURL, // URL to the weights file
+        modelID, // model ID
+        tokenizerURL, // URL to the tokenizer file
+        configURL, // model config URL
+        mel_filtersURL, // URL to the mel filters file
+        audioURL, // URL to the audio file
+        updateStatus // function to update the status
+      ) {
+        return new Promise((resolve, reject) => {
+          whisperWorker.postMessage({
+            weightsURL,
+            modelID,
+            tokenizerURL,
+            configURL,
+            mel_filtersURL,
+            audioURL,
+          });
+          function messageHandler(event) {
+            console.log(event.data);
+            if ("status" in event.data) {
+              updateStatus(event.data);
+            }
+            if ("error" in event.data) {
+              whisperWorker.removeEventListener("message", messageHandler);
+              reject(new Error(event.data.error));
+            }
+            if (event.data.status === "complete") {
+              whisperWorker.removeEventListener("message", messageHandler);
+              resolve(event.data);
+            }
+          }
+          whisperWorker.addEventListener("message", messageHandler);
+        });
+      }
+
+      // keep track of the audio URL
+      let audioURL = null;
+      function setAudio(src) {
+        const audio = document.querySelector("#audio");
+        audio.src = src;
+        audio.controls = true;
+        audio.hidden = false;
+        document.querySelector("#detect").disabled = false;
+        audioURL = src;
+      }
+      // add event listener to audio buttons
+      document.querySelectorAll("#audios-select > button").forEach((target) => {
+        target.addEventListener("click", (e) => {
+          const value = target.dataset.value;
+          const href = AUDIO_BASE_URL + value;
+          setAudio(href);
+        });
+      });
+      //add event listener to file input
+      document.querySelector("#file-upload").addEventListener("change", (e) => {
+        const target = e.target;
+        if (target.files.length > 0) {
+          const href = URL.createObjectURL(target.files[0]);
+          setAudio(href);
+        }
+      });
+      // add event listener to drop-area
+      const dropArea = document.querySelector("#drop-area");
+      dropArea.addEventListener("dragenter", (e) => {
+        e.preventDefault();
+        dropArea.classList.add("border-blue-700");
+      });
+      dropArea.addEventListener("dragleave", (e) => {
+        e.preventDefault();
+        dropArea.classList.remove("border-blue-700");
+      });
+      dropArea.addEventListener("dragover", (e) => {
+        e.preventDefault();
+        dropArea.classList.add("border-blue-700");
+      });
+      dropArea.addEventListener("drop", (e) => {
+        e.preventDefault();
+        dropArea.classList.remove("border-blue-700");
+        const url = e.dataTransfer.getData("text/uri-list");
+        const files = e.dataTransfer.files;
+        if (files.length > 0) {
+          const href = URL.createObjectURL(files[0]);
+          setAudio(href);
+        } else if (url) {
+          setAudio(url);
+        }
+      });
+
+      // add event listener to detect button
+      document.querySelector("#detect").addEventListener("click", async () => {
+        if (audioURL === null) {
+          return;
+        }
+        const modelID = modelEl.value;
+        const model = MODELS[modelID];
+        const modelURL = model.base_url + model.model;
+        const tokenizerURL = model.base_url + model.tokenizer;
+        const configURL = model.base_url + model.config;
+
+        classifyAudio(
+          modelURL,
+          modelID,
+          tokenizerURL,
+          configURL,
+          "mel_filters.safetensors",
+          audioURL,
+          updateStatus
+        )
+          .then((result) => {
+            console.log("RESULT", result);
+            const { output } = result;
+            const text = output.map((segment) => segment.dr.text).join(" ");
+            console.log(text);
+            document.querySelector("#output-status").hidden = true;
+            document.querySelector("#output-generation").hidden = false;
+            document.querySelector("#output-generation").textContent = text;
+          })
+          .catch((error) => {
+            console.error(error);
+          });
+      });
+
+      function updateStatus(data) {
+        const { status, message } = data;
+        const button = document.querySelector("#detect");
+        if (status === "decoding" || status === "loading") {
+          button.disabled = true;
+          button.textContent = message;
+        } else if (status === "complete") {
+          button.disabled = false;
+          button.textContent = "Transcribe Audio";
+        }
+      }
+    </script>
+  </head>
+  <body class="container max-w-4xl mx-auto p-4">
+    <main class="grid grid-cols-1 gap-8 relative">
+      <span class="absolute text-5xl -ml-[1em]"> 🕯️ </span>
+      <div>
+        <h1 class="text-5xl font-bold">Candle Whisper</h1>
+        <h2 class="text-2xl font-bold">Rust/WASM Demo</h2>
+        <p class="max-w-lg">
+          Transcribe audio in the browser using rust/wasm with an audio file.
+          This demo uses the
+          <a
+            href="https://huggingface.co/openai/"
+            target="_blank"
+            class="underline hover:text-blue-500 hover:no-underline">
+            OpenAI Whisper models
+          </a>
+          and WASM runtime built with
+          <a
+            href="https://github.com/huggingface/candle/"
+            target="_blank"
+            class="underline hover:text-blue-500 hover:no-underline"
+            >Candle
+          </a>
+        </p>
+      </div>
+
+      <div>
+        <label for="model" class="font-medium">Models Options: </label>
+        <select
+          id="model"
+          class="border-2 border-gray-500 rounded-md font-light">
+        </select>
+      </div>
+      <!-- drag and drop area -->
+      <div class="relative">
+        <div
+          id="drop-area"
+          class="flex flex-col items-center justify-center border-2 border-gray-300 border-dashed rounded-xl relative h-48 w-full overflow-hidden">
+          <div
+            class="flex flex-col items-center justify-center space-y-1 text-center">
+            <svg
+              width="25"
+              height="25"
+              viewBox="0 0 25 25"
+              fill="none"
+              xmlns="http://www.w3.org/2000/svg">
+              <path
+                d="M3.5 24.3a3 3 0 0 1-1.9-.8c-.5-.5-.8-1.2-.8-1.9V2.9c0-.7.3-1.3.8-1.9.6-.5 1.2-.7 2-.7h18.6c.7 0 1.3.2 1.9.7.5.6.7 1.2.7 2v18.6c0 .7-.2 1.4-.7 1.9a3 3 0 0 1-2 .8H3.6Zm0-2.7h18.7V2.9H3.5v18.7Zm2.7-2.7h13.3c.3 0 .5 0 .6-.3v-.7l-3.7-5a.6.6 0 0 0-.6-.2c-.2 0-.4 0-.5.3l-3.5 4.6-2.4-3.3a.6.6 0 0 0-.6-.3c-.2 0-.4.1-.5.3l-2.7 3.6c-.1.2-.2.4 0 .7.1.2.3.3.6.3Z"
+                fill="#000" />
+            </svg>
+            <div class="flex text-sm text-gray-600">
+              <label
+                for="file-upload"
+                class="relative cursor-pointer bg-white rounded-md font-medium text-blue-950 hover:text-blue-700">
+                <span>Drag and drop your audio here</span>
+                <span class="block text-xs">or</span>
+                <span class="block text-xs">Click to upload</span>
+              </label>
+            </div>
+            <input
+              id="file-upload"
+              name="file-upload"
+              type="file"
+              accept="audio/*"
+              class="sr-only" />
+          </div>
+          <audio
+            id="audio"
+            hidden
+            controls
+            class="w-full p-2 select-none"></audio>
+        </div>
+      </div>
+      <div>
+        <div class="flex flex-wrap gap-3 items-center" id="audios-select">
+          <h3 class="font-medium">Examples:</h3>
+          <button
+            data-value="samples_jfk.wav"
+            class="text-gray-500 border border-gray-500 rounded-md p-2 underline hover:no-underline">
+            <span>jfk.wav</span>
+            <span class="text-xs block"> (352 kB)</span>
+          </button>
+          <button
+            data-value="samples_a13.wav"
+            class="text-gray-500 border border-gray-500 rounded-md p-2 underline hover:no-underline">
+            <span>a13.wav</span>
+            <span class="text-xs block"> (960 kB)</span>
+          </button>
+          <button
+            data-value="samples_mm0.wav"
+            class="text-gray-500 border border-gray-500 rounded-md p-2 underline hover:no-underline">
+            <span>mm0.wav</span>
+            <span class="text-xs block new"> (957 kB)</span>
+          </button>
+          <button
+            data-value="samples_gb0.wav"
+            class="text-gray-500 border border-gray-500 rounded-md p-2 underline hover:no-underline">
+            <span>gb0.wav </span>
+            <span class="text-xs block">(4.08 MB)</span>
+          </button>
+          <button
+            data-value="samples_gb1.wav"
+            class="text-gray-500 border border-gray-500 rounded-md p-2 underline hover:no-underline">
+            <span>gb1.wav </span>
+            <span class="text-xs block">(6.36 MB)</span>
+          </button>
+          <button
+            data-value="samples_hp0.wav"
+            class="text-gray-500 border border-gray-500 rounded-md p-2 underline hover:no-underline">
+            <span>hp0.wav </span>
+            <span class="text-xs block">(8.75 MB)</span>
+          </button>
+        </div>
+      </div>
+
+      <div>
+        <button
+          id="detect"
+          disabled
+          class="bg-gray-700 hover:bg-gray-800 text-white font-normal py-2 px-4 rounded disabled:bg-gray-300 disabled:cursor-not-allowed">
+          Transcribe Audio
+        </button>
+      </div>
+      <div>
+        <h3 class="font-medium">Transcription:</h3>
+        <div
+          class="min-h-[250px] bg-slate-100 text-gray-500 p-4 rounded-md flex flex-col gap-2">
+          <p hidden id="output-generation" class="grid-rows-2"></p>
+          <span id="output-status" class="m-auto font-light"
+            >No transcription results yet</span
+          >
+        </div>
+      </div>
+    </main>
+  </body>
+</html>

candle-wasm-examples-whisper/main.js

@@ -0,0 +1,6 @@
+import init, { run_app } from './pkg/candle_wasm_example_whisper.js';
+async function main() {
+   await init('/pkg/candle_wasm_example_whisper_bg.wasm');
+   run_app();
+}
+main()

candle-wasm-examples-whisper/src/app.rs

@@ -0,0 +1,283 @@
+use crate::console_log;
+use crate::worker::{ModelData, Segment, Worker, WorkerInput, WorkerOutput};
+use js_sys::Date;
+use wasm_bindgen::prelude::*;
+use wasm_bindgen_futures::JsFuture;
+use yew::{html, Component, Context, Html};
+use yew_agent::{Bridge, Bridged};
+
+const SAMPLE_NAMES: [&str; 6] = [
+    "audios/samples_jfk.wav",
+    "audios/samples_a13.wav",
+    "audios/samples_gb0.wav",
+    "audios/samples_gb1.wav",
+    "audios/samples_hp0.wav",
+    "audios/samples_mm0.wav",
+];
+
+async fn fetch_url(url: &str) -> Result<Vec<u8>, JsValue> {
+    use web_sys::{Request, RequestCache, RequestInit, RequestMode, Response};
+    let window = web_sys::window().ok_or("window")?;
+    let mut opts = RequestInit::new();
+    let opts = opts
+        .method("GET")
+        .mode(RequestMode::Cors)
+        .cache(RequestCache::NoCache);
+
+    let request = Request::new_with_str_and_init(url, opts)?;
+
+    let resp_value = JsFuture::from(window.fetch_with_request(&request)).await?;
+
+    // `resp_value` is a `Response` object.
+    assert!(resp_value.is_instance_of::<Response>());
+    let resp: Response = resp_value.dyn_into()?;
+    let data = JsFuture::from(resp.blob()?).await?;
+    let blob = web_sys::Blob::from(data);
+    let array_buffer = JsFuture::from(blob.array_buffer()).await?;
+    let data = js_sys::Uint8Array::new(&array_buffer).to_vec();
+    Ok(data)
+}
+
+pub enum Msg {
+    Run(usize),
+    UpdateStatus(String),
+    SetDecoder(ModelData),
+    WorkerInMsg(WorkerInput),
+    WorkerOutMsg(Result<WorkerOutput, String>),
+}
+
+pub struct CurrentDecode {
+    start_time: Option<f64>,
+}
+
+pub struct App {
+    status: String,
+    loaded: bool,
+    segments: Vec<Segment>,
+    current_decode: Option<CurrentDecode>,
+    worker: Box<dyn Bridge<Worker>>,
+}
+
+async fn model_data_load() -> Result<ModelData, JsValue> {
+    let quantized = false;
+    let is_multilingual = false;
+
+    let (tokenizer, mel_filters, weights, config) = if quantized {
+        console_log!("loading quantized weights");
+        let tokenizer = fetch_url("quantized/tokenizer-tiny-en.json").await?;
+        let mel_filters = fetch_url("mel_filters.safetensors").await?;
+        let weights = fetch_url("quantized/model-tiny-en-q80.gguf").await?;
+        let config = fetch_url("quantized/config-tiny-en.json").await?;
+        (tokenizer, mel_filters, weights, config)
+    } else {
+        console_log!("loading float weights");
+        if is_multilingual {
+            let mel_filters = fetch_url("mel_filters.safetensors").await?;
+            let tokenizer = fetch_url("whisper-tiny/tokenizer.json").await?;
+            let weights = fetch_url("whisper-tiny/model.safetensors").await?;
+            let config = fetch_url("whisper-tiny/config.json").await?;
+            (tokenizer, mel_filters, weights, config)
+        } else {
+            let mel_filters = fetch_url("mel_filters.safetensors").await?;
+            let tokenizer = fetch_url("whisper-tiny.en/tokenizer.json").await?;
+            let weights = fetch_url("whisper-tiny.en/model.safetensors").await?;
+            let config = fetch_url("whisper-tiny.en/config.json").await?;
+            (tokenizer, mel_filters, weights, config)
+        }
+    };
+
+    let timestamps = true;
+    let _task = Some("transcribe".to_string());
+    console_log!("{}", weights.len());
+    Ok(ModelData {
+        tokenizer,
+        mel_filters,
+        weights,
+        config,
+        quantized,
+        timestamps,
+        task: None,
+        is_multilingual,
+        language: None,
+    })
+}
+
+fn performance_now() -> Option<f64> {
+    let window = web_sys::window()?;
+    let performance = window.performance()?;
+    Some(performance.now() / 1000.)
+}
+
+impl Component for App {
+    type Message = Msg;
+    type Properties = ();
+
+    fn create(ctx: &Context<Self>) -> Self {
+        let status = "loading weights".to_string();
+        let cb = {
+            let link = ctx.link().clone();
+            move |e| link.send_message(Self::Message::WorkerOutMsg(e))
+        };
+        let worker = Worker::bridge(std::rc::Rc::new(cb));
+        Self {
+            status,
+            segments: vec![],
+            current_decode: None,
+            worker,
+            loaded: false,
+        }
+    }
+
+    fn rendered(&mut self, ctx: &Context<Self>, first_render: bool) {
+        if first_render {
+            ctx.link().send_future(async {
+                match model_data_load().await {
+                    Err(err) => {
+                        let status = format!("{err:?}");
+                        Msg::UpdateStatus(status)
+                    }
+                    Ok(model_data) => Msg::SetDecoder(model_data),
+                }
+            });
+        }
+    }
+
+    fn update(&mut self, ctx: &Context<Self>, msg: Self::Message) -> bool {
+        match msg {
+            Msg::SetDecoder(md) => {
+                self.status = "weights loaded successfully!".to_string();
+                self.loaded = true;
+                console_log!("loaded weights");
+                self.worker.send(WorkerInput::ModelData(md));
+                true
+            }
+            Msg::Run(sample_index) => {
+                let sample = SAMPLE_NAMES[sample_index];
+                if self.current_decode.is_some() {
+                    self.status = "already decoding some sample at the moment".to_string()
+                } else {
+                    let start_time = performance_now();
+                    self.current_decode = Some(CurrentDecode { start_time });
+                    self.status = format!("decoding {sample}");
+                    self.segments.clear();
+                    ctx.link().send_future(async move {
+                        match fetch_url(sample).await {
+                            Err(err) => {
+                                let output = Err(format!("decoding error: {err:?}"));
+                                // Mimic a worker output to so as to release current_decode
+                                Msg::WorkerOutMsg(output)
+                            }
+                            Ok(wav_bytes) => {
+                                Msg::WorkerInMsg(WorkerInput::DecodeTask { wav_bytes })
+                            }
+                        }
+                    })
+                }
+                //
+                true
+            }
+            Msg::WorkerOutMsg(output) => {
+                let dt = self.current_decode.as_ref().and_then(|current_decode| {
+                    current_decode.start_time.and_then(|start_time| {
+                        performance_now().map(|stop_time| stop_time - start_time)
+                    })
+                });
+                self.current_decode = None;
+                match output {
+                    Ok(WorkerOutput::WeightsLoaded) => self.status = "weights loaded!".to_string(),
+                    Ok(WorkerOutput::Decoded(segments)) => {
+                        self.status = match dt {
+                            None => "decoding succeeded!".to_string(),
+                            Some(dt) => format!("decoding succeeded in {:.2}s", dt),
+                        };
+                        self.segments = segments;
+                    }
+                    Err(err) => {
+                        self.status = format!("decoding error {err:?}");
+                    }
+                }
+                true
+            }
+            Msg::WorkerInMsg(inp) => {
+                self.worker.send(inp);
+                true
+            }
+            Msg::UpdateStatus(status) => {
+                self.status = status;
+                true
+            }
+        }
+    }
+
+    fn view(&self, ctx: &Context<Self>) -> Html {
+        html! {
+            <div>
+                <table>
+                <thead>
+                <tr>
+                  <th>{"Sample"}</th>
+                  <th></th>
+                  <th></th>
+                </tr>
+                </thead>
+                <tbody>
+                {
+                    SAMPLE_NAMES.iter().enumerate().map(|(i, name)| { html! {
+                <tr>
+                  <th>{name}</th>
+                  <th><audio controls=true src={format!("./{name}")}></audio></th>
+                  { if self.loaded {
+                      html!(<th><button class="button" onclick={ctx.link().callback(move |_| Msg::Run(i))}> { "run" }</button></th>)
+                       }else{html!()}
+                  }
+                </tr>
+                    }
+                    }).collect::<Html>()
+                }
+                </tbody>
+                </table>
+                <h2>
+                  {&self.status}
+                </h2>
+                {
+                    if !self.loaded{
+                        html! { <progress id="progress-bar" aria-label="loading weights…"></progress> }
+                    } else if self.current_decode.is_some() {
+                        html! { <progress id="progress-bar" aria-label="decoding…"></progress> }
+                    } else { html!{
+                <blockquote>
+                <p>
+                  {
+                      self.segments.iter().map(|segment| { html! {
+                          <>
+                          <i>
+                          {
+                              format!("{:.2}s-{:.2}s: (avg-logprob: {:.4}, no-speech-prob: {:.4})",
+                                  segment.start,
+                                  segment.start + segment.duration,
+                                  segment.dr.avg_logprob,
+                                  segment.dr.no_speech_prob,
+                              )
+                          }
+                          </i>
+                          <br/ >
+                          {&segment.dr.text}
+                          <br/ >
+                          </>
+                      } }).collect::<Html>()
+                  }
+                </p>
+                </blockquote>
+                }
+                }
+                }
+
+                // Display the current date and time the page was rendered
+                <p class="footer">
+                    { "Rendered: " }
+                    { String::from(Date::new_0().to_string()) }
+                </p>
+            </div>
+        }
+    }
+}

candle-wasm-examples-whisper/src/audio.rs

@@ -0,0 +1,215 @@
+// Audio processing code, adapted from whisper.cpp
+// https://github.com/ggerganov/whisper.cpp
+use super::worker;
+
+pub trait Float: num_traits::Float + num_traits::FloatConst + num_traits::NumAssign {}
+
+impl Float for f32 {}
+impl Float for f64 {}
+
+// https://github.com/ggerganov/whisper.cpp/blob/4774d2feb01a772a15de81ffc34b34a1f294f020/whisper.cpp#L2357
+fn fft<T: Float>(inp: &[T]) -> Vec<T> {
+    let n = inp.len();
+    let zero = T::zero();
+    if n == 1 {
+        return vec![inp[0], zero];
+    }
+    if n % 2 == 1 {
+        return dft(inp);
+    }
+    let mut out = vec![zero; n * 2];
+
+    let mut even = Vec::with_capacity(n / 2);
+    let mut odd = Vec::with_capacity(n / 2);
+
+    for (i, &inp) in inp.iter().enumerate() {
+        if i % 2 == 0 {
+            even.push(inp)
+        } else {
+            odd.push(inp);
+        }
+    }
+
+    let even_fft = fft(&even);
+    let odd_fft = fft(&odd);
+
+    let two_pi = T::PI() + T::PI();
+    let n_t = T::from(n).unwrap();
+    for k in 0..n / 2 {
+        let k_t = T::from(k).unwrap();
+        let theta = two_pi * k_t / n_t;
+        let re = theta.cos();
+        let im = -theta.sin();
+
+        let re_odd = odd_fft[2 * k];
+        let im_odd = odd_fft[2 * k + 1];
+
+        out[2 * k] = even_fft[2 * k] + re * re_odd - im * im_odd;
+        out[2 * k + 1] = even_fft[2 * k + 1] + re * im_odd + im * re_odd;
+
+        out[2 * (k + n / 2)] = even_fft[2 * k] - re * re_odd + im * im_odd;
+        out[2 * (k + n / 2) + 1] = even_fft[2 * k + 1] - re * im_odd - im * re_odd;
+    }
+    out
+}
+
+// https://github.com/ggerganov/whisper.cpp/blob/4774d2feb01a772a15de81ffc34b34a1f294f020/whisper.cpp#L2337
+fn dft<T: Float>(inp: &[T]) -> Vec<T> {
+    let zero = T::zero();
+    let n = inp.len();
+    let two_pi = T::PI() + T::PI();
+
+    let mut out = Vec::with_capacity(2 * n);
+    let n_t = T::from(n).unwrap();
+    for k in 0..n {
+        let k_t = T::from(k).unwrap();
+        let mut re = zero;
+        let mut im = zero;
+
+        for (j, &inp) in inp.iter().enumerate() {
+            let j_t = T::from(j).unwrap();
+            let angle = two_pi * k_t * j_t / n_t;
+            re += inp * angle.cos();
+            im -= inp * angle.sin();
+        }
+
+        out.push(re);
+        out.push(im);
+    }
+    out
+}
+
+#[allow(clippy::too_many_arguments)]
+// https://github.com/ggerganov/whisper.cpp/blob/4774d2feb01a772a15de81ffc34b34a1f294f020/whisper.cpp#L2414
+fn log_mel_spectrogram_w<T: Float>(
+    ith: usize,
+    hann: &[T],
+    samples: &[T],
+    filters: &[T],
+    fft_size: usize,
+    fft_step: usize,
+    speed_up: bool,
+    n_len: usize,
+    n_mel: usize,
+    n_threads: usize,
+) -> Vec<T> {
+    let n_fft = if speed_up {
+        1 + fft_size / 4
+    } else {
+        1 + fft_size / 2
+    };
+
+    let zero = T::zero();
+    let half = T::from(0.5).unwrap();
+    let mut fft_in = vec![zero; fft_size];
+    let mut mel = vec![zero; n_len * n_mel];
+
+    for i in (ith..n_len).step_by(n_threads) {
+        let offset = i * fft_step;
+
+        // apply Hanning window
+        for j in 0..fft_size {
+            fft_in[j] = if offset + j < samples.len() {
+                hann[j] * samples[offset + j]
+            } else {
+                zero
+            }
+        }
+
+        // FFT -> mag^2
+        let mut fft_out: Vec<T> = fft(&fft_in);
+
+        for j in 0..fft_size {
+            fft_out[j] = fft_out[2 * j] * fft_out[2 * j] + fft_out[2 * j + 1] * fft_out[2 * j + 1];
+        }
+        for j in 1..fft_size / 2 {
+            let v = fft_out[fft_size - j];
+            fft_out[j] += v;
+        }
+
+        if speed_up {
+            // scale down in the frequency domain results in a speed up in the time domain
+            for j in 0..n_fft {
+                fft_out[j] = half * (fft_out[2 * j] + fft_out[2 * j + 1]);
+            }
+        }
+
+        // mel spectrogram
+        for j in 0..n_mel {
+            let mut sum = zero;
+            for k in 0..n_fft {
+                sum += fft_out[k] * filters[j * n_fft + k];
+            }
+            mel[j * n_len + i] = T::max(sum, T::from(1e-10).unwrap()).log10();
+        }
+    }
+    mel
+}
+
+fn log_mel_spectrogram_<T: Float + std::fmt::Display>(
+    samples: &[T],
+    filters: &[T],
+    fft_size: usize,
+    fft_step: usize,
+    n_mel: usize,
+    speed_up: bool,
+) -> Vec<T> {
+    let zero = T::zero();
+    let two_pi = T::PI() + T::PI();
+    let half = T::from(0.5).unwrap();
+    let one = T::from(1.0).unwrap();
+    let four = T::from(4.0).unwrap();
+    let fft_size_t = T::from(fft_size).unwrap();
+
+    let hann: Vec<T> = (0..fft_size)
+        .map(|i| half * (one - ((two_pi * T::from(i).unwrap()) / fft_size_t).cos()))
+        .collect();
+    let n_len = samples.len() / fft_step;
+
+    // pad audio with at least one extra chunk of zeros
+    let pad = 100 * worker::m::CHUNK_LENGTH / 2;
+    let n_len = if n_len % pad != 0 {
+        (n_len / pad + 1) * pad
+    } else {
+        n_len
+    };
+    let n_len = n_len + pad;
+    let samples = {
+        let mut samples_padded = samples.to_vec();
+        let to_add = n_len * fft_step - samples.len();
+        samples_padded.extend(std::iter::repeat(zero).take(to_add));
+        samples_padded
+    };
+
+    // Use a single thread for now.
+    let mut mel = log_mel_spectrogram_w(
+        0, &hann, &samples, filters, fft_size, fft_step, speed_up, n_len, n_mel, 1,
+    );
+    let mmax = mel
+        .iter()
+        .max_by(|&u, &v| u.partial_cmp(v).unwrap_or(std::cmp::Ordering::Greater))
+        .copied()
+        .unwrap_or(zero)
+        - T::from(8).unwrap();
+    for m in mel.iter_mut() {
+        let v = T::max(*m, mmax);
+        *m = v / four + one
+    }
+    mel
+}
+
+pub fn pcm_to_mel<T: Float + std::fmt::Display>(
+    cfg: &worker::m::Config,
+    samples: &[T],
+    filters: &[T],
+) -> anyhow::Result<Vec<T>> {
+    let mel = log_mel_spectrogram_(
+        samples,
+        filters,
+        worker::m::N_FFT,
+        worker::m::HOP_LENGTH,
+        cfg.num_mel_bins,
+        false,
+    );
+    Ok(mel)
+}

candle-wasm-examples-whisper/src/bin/app.rs

@@ -0,0 +1,4 @@
+fn main() {
+    wasm_logger::init(wasm_logger::Config::new(log::Level::Trace));
+    yew::Renderer::<candle_wasm_example_whisper::App>::new().render();
+}

candle-wasm-examples-whisper/src/bin/m.rs

@@ -0,0 +1,53 @@
+use candle_wasm_example_whisper::worker::{Decoder as D, ModelData};
+use wasm_bindgen::prelude::*;
+
+#[wasm_bindgen]
+pub struct Decoder {
+    decoder: D,
+}
+
+#[wasm_bindgen]
+impl Decoder {
+    #[wasm_bindgen(constructor)]
+    #[allow(clippy::too_many_arguments)]
+    pub fn new(
+        weights: Vec<u8>,
+        tokenizer: Vec<u8>,
+        mel_filters: Vec<u8>,
+        config: Vec<u8>,
+        quantized: bool,
+        is_multilingual: bool,
+        timestamps: bool,
+        task: Option<String>,
+        language: Option<String>,
+    ) -> Result<Decoder, JsError> {
+        let decoder = D::load(ModelData {
+            tokenizer,
+            mel_filters,
+            config,
+            quantized,
+            weights,
+            is_multilingual,
+            timestamps,
+            task,
+            language,
+        });
+
+        match decoder {
+            Ok(decoder) => Ok(Self { decoder }),
+            Err(e) => Err(JsError::new(&e.to_string())),
+        }
+    }
+
+    #[wasm_bindgen]
+    pub fn decode(&mut self, wav_input: Vec<u8>) -> Result<String, JsError> {
+        let segments = self
+            .decoder
+            .convert_and_run(&wav_input)
+            .map_err(|e| JsError::new(&e.to_string()))?;
+        let json = serde_json::to_string(&segments)?;
+        Ok(json)
+    }
+}
+
+fn main() {}

candle-wasm-examples-whisper/src/bin/worker.rs

@@ -0,0 +1,4 @@
+use yew_agent::PublicWorker;
+fn main() {
+    candle_wasm_example_whisper::Worker::register();
+}

candle-wasm-examples-whisper/src/languages.rs

@@ -0,0 +1,101 @@
+pub const LANGUAGES: [(&str, &str); 99] = [
+    ("en", "english"),
+    ("zh", "chinese"),
+    ("de", "german"),
+    ("es", "spanish"),
+    ("ru", "russian"),
+    ("ko", "korean"),
+    ("fr", "french"),
+    ("ja", "japanese"),
+    ("pt", "portuguese"),
+    ("tr", "turkish"),
+    ("pl", "polish"),
+    ("ca", "catalan"),
+    ("nl", "dutch"),
+    ("ar", "arabic"),
+    ("sv", "swedish"),
+    ("it", "italian"),
+    ("id", "indonesian"),
+    ("hi", "hindi"),
+    ("fi", "finnish"),
+    ("vi", "vietnamese"),
+    ("he", "hebrew"),
+    ("uk", "ukrainian"),
+    ("el", "greek"),
+    ("ms", "malay"),
+    ("cs", "czech"),
+    ("ro", "romanian"),
+    ("da", "danish"),
+    ("hu", "hungarian"),
+    ("ta", "tamil"),
+    ("no", "norwegian"),
+    ("th", "thai"),
+    ("ur", "urdu"),
+    ("hr", "croatian"),
+    ("bg", "bulgarian"),
+    ("lt", "lithuanian"),
+    ("la", "latin"),
+    ("mi", "maori"),
+    ("ml", "malayalam"),
+    ("cy", "welsh"),
+    ("sk", "slovak"),
+    ("te", "telugu"),
+    ("fa", "persian"),
+    ("lv", "latvian"),
+    ("bn", "bengali"),
+    ("sr", "serbian"),
+    ("az", "azerbaijani"),
+    ("sl", "slovenian"),
+    ("kn", "kannada"),
+    ("et", "estonian"),
+    ("mk", "macedonian"),
+    ("br", "breton"),
+    ("eu", "basque"),
+    ("is", "icelandic"),
+    ("hy", "armenian"),
+    ("ne", "nepali"),
+    ("mn", "mongolian"),
+    ("bs", "bosnian"),
+    ("kk", "kazakh"),
+    ("sq", "albanian"),
+    ("sw", "swahili"),
+    ("gl", "galician"),
+    ("mr", "marathi"),
+    ("pa", "punjabi"),
+    ("si", "sinhala"),
+    ("km", "khmer"),
+    ("sn", "shona"),
+    ("yo", "yoruba"),
+    ("so", "somali"),
+    ("af", "afrikaans"),
+    ("oc", "occitan"),
+    ("ka", "georgian"),
+    ("be", "belarusian"),
+    ("tg", "tajik"),
+    ("sd", "sindhi"),
+    ("gu", "gujarati"),
+    ("am", "amharic"),
+    ("yi", "yiddish"),
+    ("lo", "lao"),
+    ("uz", "uzbek"),
+    ("fo", "faroese"),
+    ("ht", "haitian creole"),
+    ("ps", "pashto"),
+    ("tk", "turkmen"),
+    ("nn", "nynorsk"),
+    ("mt", "maltese"),
+    ("sa", "sanskrit"),
+    ("lb", "luxembourgish"),
+    ("my", "myanmar"),
+    ("bo", "tibetan"),
+    ("tl", "tagalog"),
+    ("mg", "malagasy"),
+    ("as", "assamese"),
+    ("tt", "tatar"),
+    ("haw", "hawaiian"),
+    ("ln", "lingala"),
+    ("ha", "hausa"),
+    ("ba", "bashkir"),
+    ("jw", "javanese"),
+    ("su", "sundanese"),
+];

candle-wasm-examples-whisper/src/lib.rs

@@ -0,0 +1,29 @@
+pub const WITH_TIMER: bool = true;
+
+struct Timer {
+    label: &'static str,
+}
+
+// impl Timer {
+//     fn new(label: &'static str) -> Self {
+//         if WITH_TIMER {
+//             web_sys::console::time_with_label(label);
+//         }
+//         Self { label }
+//     }
+// }
+
+impl Drop for Timer {
+    fn drop(&mut self) {
+        if WITH_TIMER {
+            web_sys::console::time_end_with_label(self.label)
+        }
+    }
+}
+
+mod app;
+mod audio;
+pub mod languages;
+pub mod worker;
+pub use app::App;
+pub use worker::Worker;

candle-wasm-examples-whisper/src/worker.rs

@@ -0,0 +1,492 @@
+use crate::languages::LANGUAGES;
+use anyhow::Error as E;
+use candle::{safetensors::Load, DType, Device, IndexOp, Tensor, D};
+use candle_nn::{ops::softmax, VarBuilder};
+pub use candle_transformers::models::whisper::{self as m, Config};
+use rand::{distributions::Distribution, rngs::StdRng, SeedableRng};
+use serde::{Deserialize, Serialize};
+use tokenizers::Tokenizer;
+use wasm_bindgen::prelude::*;
+use yew_agent::{HandlerId, Public, WorkerLink};
+
+#[wasm_bindgen]
+extern "C" {
+    // Use `js_namespace` here to bind `console.log(..)` instead of just
+    // `log(..)`
+    #[wasm_bindgen(js_namespace = console)]
+    pub fn log(s: &str);
+}
+
+#[macro_export]
+macro_rules! console_log {
+    // Note that this is using the `log` function imported above during
+    // `bare_bones`
+    ($($t:tt)*) => ($crate::worker::log(&format_args!($($t)*).to_string()))
+}
+
+pub const DTYPE: DType = DType::F32;
+
+pub enum Model {
+    Normal(m::model::Whisper),
+    Quantized(m::quantized_model::Whisper),
+}
+
+// Maybe we should use some traits rather than doing the dispatch for all these.
+impl Model {
+    pub fn config(&self) -> &Config {
+        match self {
+            Self::Normal(m) => &m.config,
+            Self::Quantized(m) => &m.config,
+        }
+    }
+
+    pub fn encoder_forward(&mut self, x: &Tensor, flush: bool) -> candle::Result<Tensor> {
+        match self {
+            Self::Normal(m) => m.encoder.forward(x, flush),
+            Self::Quantized(m) => m.encoder.forward(x, flush),
+        }
+    }
+
+    pub fn decoder_forward(
+        &mut self,
+        x: &Tensor,
+        xa: &Tensor,
+        flush: bool,
+    ) -> candle::Result<Tensor> {
+        match self {
+            Self::Normal(m) => m.decoder.forward(x, xa, flush),
+            Self::Quantized(m) => m.decoder.forward(x, xa, flush),
+        }
+    }
+
+    pub fn decoder_final_linear(&self, x: &Tensor) -> candle::Result<Tensor> {
+        match self {
+            Self::Normal(m) => m.decoder.final_linear(x),
+            Self::Quantized(m) => m.decoder.final_linear(x),
+        }
+    }
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct DecodingResult {
+    pub tokens: Vec<u32>,
+    pub text: String,
+    pub avg_logprob: f64,
+    pub no_speech_prob: f64,
+    temperature: f64,
+    compression_ratio: f64,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Segment {
+    pub start: f64,
+    pub duration: f64,
+    pub dr: DecodingResult,
+}
+
+pub struct Decoder {
+    model: Model,
+    rng: rand::rngs::StdRng,
+    task: Option<Task>,
+    language: Option<String>,
+    is_multilingual: bool,
+    mel_filters: Vec<f32>,
+    timestamps: bool,
+    tokenizer: Tokenizer,
+    suppress_tokens: Tensor,
+    sot_token: u32,
+    transcribe_token: u32,
+    translate_token: u32,
+    eot_token: u32,
+    no_speech_token: u32,
+    no_timestamps_token: u32,
+}
+
+impl Decoder {
+    #[allow(clippy::too_many_arguments)]
+    fn new(
+        model: Model,
+        tokenizer: Tokenizer,
+        mel_filters: Vec<f32>,
+        device: &Device,
+        task: Option<Task>,
+        language: Option<String>,
+        is_multilingual: bool,
+        timestamps: bool,
+    ) -> anyhow::Result<Self> {
+        let suppress_tokens: Vec<f32> = (0..model.config().vocab_size as u32)
+            .map(|i| {
+                if model.config().suppress_tokens.contains(&i) {
+                    f32::NEG_INFINITY
+                } else {
+                    0f32
+                }
+            })
+            .collect();
+        let no_timestamps_token = token_id(&tokenizer, m::NO_TIMESTAMPS_TOKEN)?;
+        let suppress_tokens = Tensor::new(suppress_tokens.as_slice(), device)?;
+        let sot_token = token_id(&tokenizer, m::SOT_TOKEN)?;
+        let transcribe_token = token_id(&tokenizer, m::TRANSCRIBE_TOKEN)?;
+        let translate_token = token_id(&tokenizer, m::TRANSLATE_TOKEN)?;
+        let eot_token = token_id(&tokenizer, m::EOT_TOKEN)?;
+        let no_speech_token = m::NO_SPEECH_TOKENS
+            .iter()
+            .find_map(|token| token_id(&tokenizer, token).ok());
+        let no_speech_token = match no_speech_token {
+            None => anyhow::bail!("unable to find any non-speech token"),
+            Some(n) => n,
+        };
+        let seed = 299792458;
+        Ok(Self {
+            model,
+            rng: StdRng::seed_from_u64(seed),
+            tokenizer,
+            mel_filters,
+            task,
+            timestamps,
+            language,
+            is_multilingual,
+            suppress_tokens,
+            sot_token,
+            transcribe_token,
+            translate_token,
+            eot_token,
+            no_speech_token,
+            no_timestamps_token,
+        })
+    }
+
+    fn decode(&mut self, mel: &Tensor, t: f64) -> anyhow::Result<DecodingResult> {
+        let model = &mut self.model;
+        let language_token = match (self.is_multilingual, &self.language) {
+            (true, None) => Some(detect_language(model, &self.tokenizer, mel)?),
+            (false, None) => None,
+            (true, Some(language)) => {
+                match token_id(&self.tokenizer, &format!("<|{:?}|>", self.language)) {
+                    Ok(token_id) => Some(token_id),
+                    Err(_) => anyhow::bail!("language {language} is not supported"),
+                }
+            }
+            (false, Some(_)) => {
+                anyhow::bail!("a language cannot be set for non-multilingual models")
+            }
+        };
+
+        let audio_features = model.encoder_forward(mel, true)?;
+        println!("audio features: {:?}", audio_features.dims());
+        let sample_len = model.config().max_target_positions / 2;
+        let mut sum_logprob = 0f64;
+        let mut no_speech_prob = f64::NAN;
+        let mut tokens = vec![self.sot_token];
+        if let Some(language_token) = language_token {
+            tokens.push(language_token);
+        }
+        match self.task {
+            None | Some(Task::Transcribe) => tokens.push(self.transcribe_token),
+            Some(Task::Translate) => tokens.push(self.translate_token),
+        }
+        if !self.timestamps {
+            tokens.push(self.no_timestamps_token);
+        }
+        for i in 0..sample_len {
+            let tokens_t = Tensor::new(tokens.as_slice(), mel.device())?;
+
+            // The model expects a batch dim but this inference loop does not handle
+            // it so we add it at this point.
+            let tokens_t = tokens_t.unsqueeze(0)?;
+            let ys = model.decoder_forward(&tokens_t, &audio_features, i == 0)?;
+
+            // Extract the no speech probability on the first iteration by looking at the first
+            // token logits and the probability for the according token.
+            if i == 0 {
+                let logits = model.decoder_final_linear(&ys.i(..1)?)?.i(0)?.i(0)?;
+                no_speech_prob = softmax(&logits, 0)?
+                    .i(self.no_speech_token as usize)?
+                    .to_scalar::<f32>()? as f64;
+            }
+
+            let (_, seq_len, _) = ys.dims3()?;
+            let logits = model
+                .decoder_final_linear(&ys.i((..1, seq_len - 1..))?)?
+                .i(0)?
+                .i(0)?;
+            // TODO: Besides suppress tokens, we should apply the heuristics from
+            // ApplyTimestampRules, i.e.:
+            // - Timestamps come in pairs, except before EOT.
+            // - Timestamps should be non-decreasing.
+            // - If the sum of the probabilities of timestamps is higher than any other tokens,
+            //   only consider timestamps when sampling.
+            // https://github.com/openai/whisper/blob/e8622f9afc4eba139bf796c210f5c01081000472/whisper/decoding.py#L439
+            let logits = logits.broadcast_add(&self.suppress_tokens)?;
+            let next_token = if t > 0f64 {
+                let prs = softmax(&(&logits / t)?, 0)?;
+                let logits_v: Vec<f32> = prs.to_vec1()?;
+                let distr = rand::distributions::WeightedIndex::new(&logits_v)?;
+                distr.sample(&mut self.rng) as u32
+            } else {
+                let logits_v: Vec<f32> = logits.to_vec1()?;
+                logits_v
+                    .iter()
+                    .enumerate()
+                    .max_by(|(_, u), (_, v)| u.total_cmp(v))
+                    .map(|(i, _)| i as u32)
+                    .unwrap()
+            };
+            tokens.push(next_token);
+            let prob = softmax(&logits, candle::D::Minus1)?
+                .i(next_token as usize)?
+                .to_scalar::<f32>()? as f64;
+            if next_token == self.eot_token || tokens.len() > model.config().max_target_positions {
+                break;
+            }
+            sum_logprob += prob.ln();
+        }
+        let text = self.tokenizer.decode(&tokens, true).map_err(E::msg)?;
+        let avg_logprob = sum_logprob / tokens.len() as f64;
+
+        Ok(DecodingResult {
+            tokens,
+            text,
+            avg_logprob,
+            no_speech_prob,
+            temperature: t,
+            compression_ratio: f64::NAN,
+        })
+    }
+
+    fn decode_with_fallback(&mut self, segment: &Tensor) -> anyhow::Result<DecodingResult> {
+        for (i, &t) in m::TEMPERATURES.iter().enumerate() {
+            let dr: Result<DecodingResult, _> = self.decode(segment, t);
+            if i == m::TEMPERATURES.len() - 1 {
+                return dr;
+            }
+            // On errors, we try again with a different temperature.
+            match dr {
+                Ok(dr) => {
+                    let needs_fallback = dr.compression_ratio > m::COMPRESSION_RATIO_THRESHOLD
+                        || dr.avg_logprob < m::LOGPROB_THRESHOLD;
+                    if !needs_fallback || dr.no_speech_prob > m::NO_SPEECH_THRESHOLD {
+                        return Ok(dr);
+                    }
+                }
+                Err(err) => {
+                    console_log!("Error running at {t}: {err}")
+                }
+            }
+        }
+        unreachable!()
+    }
+
+    fn run(&mut self, mel: &Tensor) -> anyhow::Result<Vec<Segment>> {
+        let (_, _, content_frames) = mel.dims3()?;
+        let mut seek = 0;
+        let mut segments = vec![];
+        while seek < content_frames {
+            let time_offset = (seek * m::HOP_LENGTH) as f64 / m::SAMPLE_RATE as f64;
+            let segment_size = usize::min(content_frames - seek, m::N_FRAMES);
+            let mel_segment = mel.narrow(2, seek, segment_size)?;
+            let segment_duration = (segment_size * m::HOP_LENGTH) as f64 / m::SAMPLE_RATE as f64;
+            let dr = self.decode_with_fallback(&mel_segment)?;
+            seek += segment_size;
+            if dr.no_speech_prob > m::NO_SPEECH_THRESHOLD && dr.avg_logprob < m::LOGPROB_THRESHOLD {
+                console_log!("no speech detected, skipping {seek} {dr:?}");
+                continue;
+            }
+            let segment = Segment {
+                start: time_offset,
+                duration: segment_duration,
+                dr,
+            };
+            console_log!("{seek}: {segment:?}");
+            segments.push(segment)
+        }
+        Ok(segments)
+    }
+
+    pub fn load(md: ModelData) -> anyhow::Result<Self> {
+        let device = Device::Cpu;
+        let tokenizer = Tokenizer::from_bytes(&md.tokenizer).map_err(E::msg)?;
+
+        let mel_filters = safetensors::tensor::SafeTensors::deserialize(&md.mel_filters)?;
+        let mel_filters = mel_filters.tensor("mel_80")?.load(&device)?;
+        console_log!("loaded mel filters {:?}", mel_filters.shape());
+        let mel_filters = mel_filters.flatten_all()?.to_vec1::<f32>()?;
+        let config: Config = serde_json::from_slice(&md.config)?;
+        let model = if md.quantized {
+            let vb = candle_transformers::quantized_var_builder::VarBuilder::from_gguf_buffer(
+                &md.weights,
+            )?;
+            Model::Quantized(m::quantized_model::Whisper::load(&vb, config)?)
+        } else {
+            let vb = VarBuilder::from_buffered_safetensors(md.weights, m::DTYPE, &device)?;
+            Model::Normal(m::model::Whisper::load(&vb, config)?)
+        };
+        console_log!("done loading model");
+
+        let task = match md.task.as_deref() {
+            Some("translate") => Some(Task::Translate),
+            _ => Some(Task::Transcribe),
+        };
+
+        let decoder = Self::new(
+            model,
+            tokenizer,
+            mel_filters,
+            &device,
+            task,
+            md.language,
+            md.is_multilingual,
+            md.timestamps,
+        )?;
+        Ok(decoder)
+    }
+
+    pub fn convert_and_run(&mut self, wav_input: &[u8]) -> anyhow::Result<Vec<Segment>> {
+        let device = Device::Cpu;
+        let mut wav_input = std::io::Cursor::new(wav_input);
+        let (header, data) = wav::read(&mut wav_input)?;
+        console_log!("loaded wav data: {header:?}");
+        if header.sampling_rate != m::SAMPLE_RATE as u32 {
+            anyhow::bail!("wav file must have a {} sampling rate", m::SAMPLE_RATE);
+        }
+        let data = data.as_sixteen().expect("expected 16 bit wav file");
+        let pcm_data: Vec<_> = data[..data.len() / header.channel_count as usize]
+            .iter()
+            .map(|v| *v as f32 / 32768.)
+            .collect();
+        console_log!("pcm data loaded {}", pcm_data.len());
+        let mel = crate::audio::pcm_to_mel(self.model.config(), &pcm_data, &self.mel_filters)?;
+        let mel_len = mel.len();
+        let n_mels = self.model.config().num_mel_bins;
+        let mel = Tensor::from_vec(mel, (1, n_mels, mel_len / n_mels), &device)?;
+        console_log!("loaded mel: {:?}", mel.dims());
+        let segments = self.run(&mel)?;
+        Ok(segments)
+    }
+}
+
+/// Returns the token id for the selected language.
+pub fn detect_language(model: &mut Model, tokenizer: &Tokenizer, mel: &Tensor) -> Result<u32, E> {
+    console_log!("detecting language");
+    let (_bsize, _, seq_len) = mel.dims3()?;
+    let mel = mel.narrow(
+        2,
+        0,
+        usize::min(seq_len, model.config().max_source_positions),
+    )?;
+    let device = mel.device();
+
+    let language_token_ids = LANGUAGES
+        .iter()
+        .map(|(t, _)| token_id(tokenizer, &format!("<|{t}|>")))
+        .map(|e| e.map_err(E::msg))
+        .collect::<Result<Vec<_>, E>>()?;
+
+    let sot_token = token_id(tokenizer, m::SOT_TOKEN)?;
+    let audio_features = model.encoder_forward(&mel, true)?;
+    let tokens = Tensor::new(&[[sot_token]], device)?;
+    let language_token_ids = Tensor::new(language_token_ids.as_slice(), device)?;
+    let ys = model.decoder_forward(&tokens, &audio_features, true)?;
+    let logits = model.decoder_final_linear(&ys.i(..1)?)?.i(0)?.i(0)?;
+    let logits = logits.index_select(&language_token_ids, 0)?;
+    let probs = candle_nn::ops::softmax(&logits, D::Minus1)?;
+    let probs = probs.to_vec1::<f32>()?;
+    let mut probs = LANGUAGES.iter().zip(probs.iter()).collect::<Vec<_>>();
+    probs.sort_by(|(_, p1), (_, p2)| p2.total_cmp(p1));
+    for ((_, language), p) in probs.iter().take(5) {
+        println!("{language}: {p}")
+    }
+    let token = &format!("<|{}|>", probs[0].0 .0);
+    let language = token_id(tokenizer, token)?;
+    console_log!("detected language: {language} {token}");
+    Ok(language)
+}
+pub fn token_id(tokenizer: &Tokenizer, token: &str) -> candle::Result<u32> {
+    match tokenizer.token_to_id(token) {
+        None => candle::bail!("no token-id for {token}"),
+        Some(id) => Ok(id),
+    }
+}
+#[derive(Serialize, Deserialize, Clone, Copy, Debug)]
+pub enum Task {
+    Transcribe,
+    Translate,
+}
+
+// Communication to the worker happens through bincode, the model weights and configs are fetched
+// on the main thread and transferred via the following structure.
+#[derive(Serialize, Deserialize)]
+pub struct ModelData {
+    pub weights: Vec<u8>,
+    pub tokenizer: Vec<u8>,
+    pub mel_filters: Vec<u8>,
+    pub config: Vec<u8>,
+    pub quantized: bool,
+    pub timestamps: bool,
+    pub is_multilingual: bool,
+    pub language: Option<String>,
+    pub task: Option<String>,
+}
+
+pub struct Worker {
+    link: WorkerLink<Self>,
+    decoder: Option<Decoder>,
+}
+
+#[derive(Serialize, Deserialize)]
+pub enum WorkerInput {
+    ModelData(ModelData),
+    DecodeTask { wav_bytes: Vec<u8> },
+}
+
+#[derive(Serialize, Deserialize)]
+pub enum WorkerOutput {
+    Decoded(Vec<Segment>),
+    WeightsLoaded,
+}
+
+impl yew_agent::Worker for Worker {
+    type Input = WorkerInput;
+    type Message = ();
+    type Output = Result<WorkerOutput, String>;
+    type Reach = Public<Self>;
+
+    fn create(link: WorkerLink<Self>) -> Self {
+        Self {
+            link,
+            decoder: None,
+        }
+    }
+
+    fn update(&mut self, _msg: Self::Message) {
+        // no messaging
+    }
+
+    fn handle_input(&mut self, msg: Self::Input, id: HandlerId) {
+        let output = match msg {
+            WorkerInput::ModelData(md) => match Decoder::load(md) {
+                Ok(decoder) => {
+                    self.decoder = Some(decoder);
+                    Ok(WorkerOutput::WeightsLoaded)
+                }
+                Err(err) => Err(format!("model creation error {err:?}")),
+            },
+            WorkerInput::DecodeTask { wav_bytes } => match &mut self.decoder {
+                None => Err("model has not been set".to_string()),
+                Some(decoder) => decoder
+                    .convert_and_run(&wav_bytes)
+                    .map(WorkerOutput::Decoded)
+                    .map_err(|e| e.to_string()),
+            },
+        };
+        self.link.respond(id, output);
+    }
+
+    fn name_of_resource() -> &'static str {
+        "worker.js"
+    }
+
+    fn resource_path_is_relative() -> bool {
+        true
+    }
+}

candle-wasm-examples-whisper/whisperWorker.js

@@ -0,0 +1,116 @@
+//load the candle Whisper decoder wasm module
+import init, { Decoder } from "./build/m.js";
+
+async function fetchArrayBuffer(url) {
+  const cacheName = "whisper-candle-cache";
+  const cache = await caches.open(cacheName);
+  const cachedResponse = await cache.match(url);
+  if (cachedResponse) {
+    const data = await cachedResponse.arrayBuffer();
+    return new Uint8Array(data);
+  }
+  const res = await fetch(url, { cache: "force-cache" });
+  cache.put(url, res.clone());
+  return new Uint8Array(await res.arrayBuffer());
+}
+class Whisper {
+  static instance = {};
+  // Retrieve the Whisper model. When called for the first time,
+  // this will load the model and save it for future use.
+  static async getInstance(params) {
+    const {
+      weightsURL,
+      modelID,
+      tokenizerURL,
+      mel_filtersURL,
+      configURL,
+      quantized,
+      is_multilingual,
+      timestamps,
+      task,
+      language,
+    } = params;
+    // load individual modelID only once
+    if (!this.instance[modelID]) {
+      await init();
+
+      self.postMessage({ status: "loading", message: "Loading Model" });
+      const [
+        weightsArrayU8,
+        tokenizerArrayU8,
+        mel_filtersArrayU8,
+        configArrayU8,
+      ] = await Promise.all([
+        fetchArrayBuffer(weightsURL),
+        fetchArrayBuffer(tokenizerURL),
+        fetchArrayBuffer(mel_filtersURL),
+        fetchArrayBuffer(configURL),
+      ]);
+
+      this.instance[modelID] = new Decoder(
+        weightsArrayU8,
+        tokenizerArrayU8,
+        mel_filtersArrayU8,
+        configArrayU8,
+        quantized,
+        is_multilingual,
+        timestamps,
+        task,
+        language
+      );
+    } else {
+      self.postMessage({ status: "loading", message: "Model Already Loaded" });
+    }
+    return this.instance[modelID];
+  }
+}
+
+self.addEventListener("message", async (event) => {
+  const {
+    weightsURL,
+    modelID,
+    tokenizerURL,
+    configURL,
+    mel_filtersURL,
+    audioURL,
+  } = event.data;
+  try {
+    self.postMessage({ status: "decoding", message: "Starting Decoder" });
+    let quantized = false;
+    if (modelID.includes("quantized")) {
+      quantized = true;
+    }
+    let is_multilingual = false;
+    if (modelID.includes("multilingual")) {
+      is_multilingual = true;
+    }
+    let timestamps = true;
+    const decoder = await Whisper.getInstance({
+      weightsURL,
+      modelID,
+      tokenizerURL,
+      mel_filtersURL,
+      configURL,
+      quantized,
+      is_multilingual,
+      timestamps,
+      task: null,
+      language: null,
+    });
+
+    self.postMessage({ status: "decoding", message: "Loading Audio" });
+    const audioArrayU8 = await fetchArrayBuffer(audioURL);
+
+    self.postMessage({ status: "decoding", message: "Running Decoder..." });
+    const segments = decoder.decode(audioArrayU8);
+
+    // Send the segment back to the main thread as JSON
+    self.postMessage({
+      status: "complete",
+      message: "complete",
+      output: JSON.parse(segments),
+    });
+  } catch (e) {
+    self.postMessage({ error: e });
+  }
+});

hotwords/hotwords.py

@@ -47,37 +47,60 @@ class Reader(threading.Thread):
       self.name = os.environ.get("MIC_NAME", "pulse_monitor")
       if not self.name:
          for index, name in enumerate(sr.Microphone.list_microphone_names()):
-             print("[{0}] Microphone with name \"{1}\" found for `Microphone(device_index={0})`".format(index, name))
+             log("[{0}] Microphone with name \"{1}\" found for `Microphone(device_index={0})`".format(index, name))
          exit()
       self.inq = inq
       self.outq = outq
 
-   def run(self):
-      log("Reader.run: start")
-      try:
-         idx = [
+   def mic_idx(self):
+      mics = []
+      while not mics and not self.should_stop():
+         log(f'searching for one of {self.name.split(",")} in {sr.Microphone.list_microphone_names()}...')
+         time.sleep(1)
+         mics = [
             idx for idx,v in enumerate(
                sr.Microphone.list_microphone_names(),
             ) if v in self.name.split(",")
-         ][0]
-         with sr.Microphone(device_index=idx) as mic:
-            while not self.should_stop():
-               try:
-                  self.outq.put(self._run(mic))
-               except Exception as e:
-                  if not "timed out" in str(e):
-                     log("Reader.run: error:", e)
+         ]
+      log("mic#", mics[0])
+      return mics[0]
+
+   def run(self):
+      try:
+         log("Reader.run: start")
+         self._run()
       except Exception as e:
-         log("Reader.run panic:", e)
-         log("microphones:", sr.Microphone.list_microphone_names())
+         log("Reader:run: exit:", e)
       finally:
          self.outq.put(None)
-      log("Reader.run: stop")
+         log("Reader.run: stop")
+
+   def _run(self):
+      while not self.should_stop():
+         time.sleep(3)
+         mic = sr.Microphone(device_index=self.mic_idx())
+         try:
+            mic.__enter__()
+            while not self.should_stop():
+               try:
+                  self.outq.put(self.read(mic))
+               except Exception as e:
+                  if not "timed out" in str(e):
+                     raise e
+         except Exception as e:
+            import traceback
+            traceback.print_exception(e)
+            log("Reader.run: error:", e)
+         finally:
+            try:
+               mic.__exit__(None, None, None)
+            except Exception as e:
+               log("Reader.run.catch: error:", e)
 
    def should_stop(self):
       return not self.inq.empty()
 
-   def _run(self, mic):
+   def read(self, mic):
       mic_timeout = int(os.environ.get("MIC_TIMEOUT", 5))
       r = sr.Recognizer()
       return r.listen(
@@ -113,13 +136,19 @@ class Parser(threading.Thread):
       p = "/tmp/whisper-cpp.wav"
       with open("/tmp/whisper-cpp.wav", "wb") as f:
          f.write(wav)
-      proc = subprocess.run(f"MODEL=./models/ggml-{os.environ.get('MODEL','tiny.en')}.bin WAV={p} P=2 rust-whisper", capture_output=True, shell=True)
+      proc = subprocess.run(f"MODEL=./models/ggml-{os.environ.get('MODEL','tiny.en')}.bin WAV={p} P={os.environ.get('P', '2')} rust-whisper", capture_output=True, shell=True)
       result = proc.stdout.decode().strip()
       if os.environ.get("DEBUG", None):
+         log("stderr:", proc.stderr.decode().strip())
          log("raw transcript:", result)
       result = result.replace(">>", "")
-      result = "".join([i.split("]")[-1] for i in result.split("[")[0]])
-      result = "".join([i.split(")")[-1] for i in result.split("(")[0]])
+      for pair in [
+         ("[", "]"),
+         ("(", ")"),
+         ("<", ">"),
+         ("*", "*"),
+      ]:
+         result = "".join([i.split(pair[1])[-1] for i in result.split(pair[0])[0]])
       if os.environ.get("DEBUG", None):
          log("annotation-free transcript:", result)
       return result
@@ -156,6 +185,8 @@ def _load_dot_notation(v, items):
                else:
                   result.append(subresult)
             return result
+      elif k == "KEYS":
+         v = [k for k in v]
       else:
          if isinstance(v, list):
             v = v[int(k)]
@@ -200,8 +231,15 @@ class Reactor(threading.Thread):
             def load_hotwords_in_yaml_file():
                with open(p.split("@")[0], "r") as f:
                   v = yaml.safe_load(f)
-                  v = load_dot_notation(v, p.split("@")[-1])
-                  return ["".join(i.strip().lower().split()) for i in v if i]
+                  if os.environ.get("DEBUG", None):
+                     log(f'opened {p.split("@")[0]} and got {v}')
+                  result = []
+                  for to_find in [i for i in p.split("@")[-1].split(",") if i]:
+                     if os.environ.get("DEBUG", None):
+                        log(f'finding {to_find} in {v}')
+                     v2 = load_dot_notation(v, to_find)
+                     result.extend(["".join(i.strip().lower().split()) for i in v2 if i])
+                  return result
             load_hotwords_in_yaml_file()
             return load_hotwords_in_yaml_file
          else:
@@ -230,13 +268,23 @@ class Reactor(threading.Thread):
       log("Reactor.run: stop")
 
    def handle(self, text):
+      try:
+         self._handle(text)
+      except Exception:
+         pass
+
+   def _handle(self, text):
       hotwords = self.load_hotwords()
       if os.environ.get("DEBUG", None):
-         log(f"seeking {hotwords} in {text}")
+         log(f"seeking {hotwords} in {text}. $HOTWORDS={os.environ.get('HOTWORDS', None)}")
       if not hotwords:
          if not os.environ.get("HOTWORDS", None):
-            print(text)
+            if os.environ.get("DEBUG", None):
+               log(f"HOTWORDS is False; {text}")
+            print(text, flush=True)
          else:
+            if os.environ.get("DEBUG", None):
+               log(f"HOTWORDS is True; {text}")
             log(text)
          return
       cleantext = "".join([i for i in "".join(text.lower().split()) if i.isalpha()])
@@ -256,7 +304,7 @@ class Actor(threading.Thread):
          self.pid = int(environ["SIGUSR2"])
          self.handle = self.handle_signal
       elif os.environ.get("URL", ""):
-         self.url = environ["URL"]
+         self.url = os.environ["URL"]
          self.handle = self.handle_url
          self.headers = [i.split("=")[:2] for i in os.environ.get("HEADERS", "").split("//") if i]
          self.body = os.environ.get("BODY", '{"hotword":"{{hotword}}","context":"{{context}}"}')
@@ -276,7 +324,7 @@ class Actor(threading.Thread):
 
    def handle_stdout(self, hotword, context):
       log(context)
-      print(hotword)
+      print(hotword, flush=True)
 
    def handle_signal(self, hotword, context):
       self.handle_stderr(hotword, context)
@@ -295,9 +343,12 @@ class Actor(threading.Thread):
          body = self.body
          body = body.replace("{{hotword}}", hotword)
          body = body.replace("{{context}}", context)
+         url = self.url
+         url = url.replace("{{hotword}}", hotword)
+         url = url.replace("{{context}}", context)
          if os.environ.get("DEBUG", "") :
-            log("POST", self.url, headers, body)
-         requests.post(self.url, headers=headers, data=body)
+            log("POST", url, headers, body)
+         requests.post(url, headers=headers, data=body)
       except Exception as e:
          log("Actor.handle_url:", e)
 

hotwords/requirements.txt

@@ -2,3 +2,4 @@ git+https://github.com/openai/whisper.git
 soundfile
 PyAudio
 SpeechRecognition
+PyYAML

hotwords/transcript.sh

@@ -0,0 +1,19 @@
+#! /bin/bash
+
+set -euo pipefail
+
+cd "$(dirname "$(realpath "$BASH_SOURCE")")"
+
+if ! which rust-whisper-baked; then
+   (
+      set -euo pipefail
+      cd ../rust-whisper-baked
+      cargo install --path .
+   )
+fi >&2
+
+cat <<EOF
+rust-whisper-baked --stream-device 'BlackHole 2ch' --stream-step 30 --stream-retain 1 --stream-{head,tail}=0.25 --threads 9 2> /dev/null
+   | tee -a "$HOME/Sync/drawful/DnD/bdoob/__log.d/$(date +%Y.%m.%d).transcript.txt"
+   | tee -a "$HOME/Sync/drawful/DnD/nessira.d/_log.d/$(date +%Y.%m.%d).transcript.txt"
+EOF

listen-lib/Cargo.lock

@@ -0,0 +1,939 @@
+# This file is automatically @generated by Cargo.
+# It is not intended for manual editing.
+version = 3
+
+[[package]]
+name = "aho-corasick"
+version = "1.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b2969dcb958b36655471fc61f7e416fa76033bdd4bfed0678d8fee1e2d07a1f0"
+dependencies = [
+ "memchr",
+]
+
+[[package]]
+name = "alsa"
+version = "0.7.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e2562ad8dcf0f789f65c6fdaad8a8a9708ed6b488e649da28c01656ad66b8b47"
+dependencies = [
+ "alsa-sys",
+ "bitflags 1.3.2",
+ "libc",
+ "nix",
+]
+
+[[package]]
+name = "alsa-sys"
+version = "0.3.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "db8fee663d06c4e303404ef5f40488a53e062f89ba8bfed81f42325aafad1527"
+dependencies = [
+ "libc",
+ "pkg-config",
+]
+
+[[package]]
+name = "autocfg"
+version = "1.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d468802bab17cbc0cc575e9b053f41e72aa36bfa6b7f55e3529ffa43161b97fa"
+
+[[package]]
+name = "bindgen"
+version = "0.69.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9ffcebc3849946a7170a05992aac39da343a90676ab392c51a4280981d6379c2"
+dependencies = [
+ "bitflags 2.4.1",
+ "cexpr",
+ "clang-sys",
+ "lazy_static",
+ "lazycell",
+ "peeking_take_while",
+ "proc-macro2",
+ "quote",
+ "regex",
+ "rustc-hash",
+ "shlex",
+ "syn 2.0.41",
+]
+
+[[package]]
+name = "bitflags"
+version = "1.3.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a"
+
+[[package]]
+name = "bitflags"
+version = "2.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "327762f6e5a765692301e5bb513e0d9fef63be86bbc14528052b1cd3e6f03e07"
+
+[[package]]
+name = "bumpalo"
+version = "3.14.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7f30e7476521f6f8af1a1c4c0b8cc94f0bee37d91763d0ca2665f299b6cd8aec"
+
+[[package]]
+name = "bytes"
+version = "1.5.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a2bd12c1caf447e69cd4528f47f94d203fd2582878ecb9e9465484c4148a8223"
+
+[[package]]
+name = "cc"
+version = "1.0.83"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f1174fb0b6ec23863f8b971027804a42614e347eafb0a95bf0b12cdae21fc4d0"
+dependencies = [
+ "jobserver",
+ "libc",
+]
+
+[[package]]
+name = "cesu8"
+version = "1.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6d43a04d8753f35258c91f8ec639f792891f748a1edbd759cf1dcea3382ad83c"
+
+[[package]]
+name = "cexpr"
+version = "0.6.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6fac387a98bb7c37292057cffc56d62ecb629900026402633ae9160df93a8766"
+dependencies = [
+ "nom",
+]
+
+[[package]]
+name = "cfg-if"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
+
+[[package]]
+name = "clang-sys"
+version = "1.6.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c688fc74432808e3eb684cae8830a86be1d66a2bd58e1f248ed0960a590baf6f"
+dependencies = [
+ "glob",
+ "libc",
+ "libloading",
+]
+
+[[package]]
+name = "combine"
+version = "4.6.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "35ed6e9d84f0b51a7f52daf1c7d71dd136fd7a3f41a8462b8cdb8c78d920fad4"
+dependencies = [
+ "bytes",
+ "memchr",
+]
+
+[[package]]
+name = "core-foundation-sys"
+version = "0.8.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "06ea2b9bc92be3c2baa9334a323ebca2d6f074ff852cd1d7b11064035cd3868f"
+
+[[package]]
+name = "coreaudio-rs"
+version = "0.11.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "321077172d79c662f64f5071a03120748d5bb652f5231570141be24cfcd2bace"
+dependencies = [
+ "bitflags 1.3.2",
+ "core-foundation-sys",
+ "coreaudio-sys",
+]
+
+[[package]]
+name = "coreaudio-sys"
+version = "0.2.14"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f3120ebb80a9de008e638ad833d4127d50ea3d3a960ea23ea69bc66d9358a028"
+dependencies = [
+ "bindgen",
+]
+
+[[package]]
+name = "cpal"
+version = "0.15.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6d959d90e938c5493000514b446987c07aed46c668faaa7d34d6c7a67b1a578c"
+dependencies = [
+ "alsa",
+ "core-foundation-sys",
+ "coreaudio-rs",
+ "dasp_sample",
+ "jni 0.19.0",
+ "js-sys",
+ "libc",
+ "mach2",
+ "ndk",
+ "ndk-context",
+ "oboe",
+ "once_cell",
+ "parking_lot",
+ "wasm-bindgen",
+ "wasm-bindgen-futures",
+ "web-sys",
+ "windows",
+]
+
+[[package]]
+name = "dasp_sample"
+version = "0.11.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0c87e182de0887fd5361989c677c4e8f5000cd9491d6d563161a8f3a5519fc7f"
+
+[[package]]
+name = "equivalent"
+version = "1.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5443807d6dff69373d433ab9ef5378ad8df50ca6298caf15de6e52e24aaf54d5"
+
+[[package]]
+name = "glob"
+version = "0.3.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d2fabcfbdc87f4758337ca535fb41a6d701b65693ce38287d856d1674551ec9b"
+
+[[package]]
+name = "hashbrown"
+version = "0.14.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "290f1a1d9242c78d09ce40a5e87e7554ee637af1351968159f4952f028f75604"
+
+[[package]]
+name = "indexmap"
+version = "2.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d530e1a18b1cb4c484e6e34556a0d948706958449fca0cab753d649f2bce3d1f"
+dependencies = [
+ "equivalent",
+ "hashbrown",
+]
+
+[[package]]
+name = "jni"
+version = "0.19.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c6df18c2e3db7e453d3c6ac5b3e9d5182664d28788126d39b91f2d1e22b017ec"
+dependencies = [
+ "cesu8",
+ "combine",
+ "jni-sys",
+ "log",
+ "thiserror",
+ "walkdir",
+]
+
+[[package]]
+name = "jni"
+version = "0.20.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "039022cdf4d7b1cf548d31f60ae783138e5fd42013f6271049d7df7afadef96c"
+dependencies = [
+ "cesu8",
+ "combine",
+ "jni-sys",
+ "log",
+ "thiserror",
+ "walkdir",
+]
+
+[[package]]
+name = "jni-sys"
+version = "0.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8eaf4bc02d17cbdd7ff4c7438cafcdf7fb9a4613313ad11b4f8fefe7d3fa0130"
+
+[[package]]
+name = "jobserver"
+version = "0.1.27"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8c37f63953c4c63420ed5fd3d6d398c719489b9f872b9fa683262f8edd363c7d"
+dependencies = [
+ "libc",
+]
+
+[[package]]
+name = "js-sys"
+version = "0.3.66"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cee9c64da59eae3b50095c18d3e74f8b73c0b86d2792824ff01bbce68ba229ca"
+dependencies = [
+ "wasm-bindgen",
+]
+
+[[package]]
+name = "lazy_static"
+version = "1.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
+
+[[package]]
+name = "lazycell"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55"
+
+[[package]]
+name = "libc"
+version = "0.2.151"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "302d7ab3130588088d277783b1e2d2e10c9e9e4a16dd9050e6ec93fb3e7048f4"
+
+[[package]]
+name = "libloading"
+version = "0.7.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b67380fd3b2fbe7527a606e18729d21c6f3951633d0500574c4dc22d2d638b9f"
+dependencies = [
+ "cfg-if",
+ "winapi",
+]
+
+[[package]]
+name = "listen-lib"
+version = "0.1.0"
+dependencies = [
+ "cpal",
+ "signal-hook",
+]
+
+[[package]]
+name = "lock_api"
+version = "0.4.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3c168f8615b12bc01f9c17e2eb0cc07dcae1940121185446edc3744920e8ef45"
+dependencies = [
+ "autocfg",
+ "scopeguard",
+]
+
+[[package]]
+name = "log"
+version = "0.4.20"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b5e6163cb8c49088c2c36f57875e58ccd8c87c7427f7fbd50ea6710b2f3f2e8f"
+
+[[package]]
+name = "mach2"
+version = "0.4.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "19b955cdeb2a02b9117f121ce63aa52d08ade45de53e48fe6a38b39c10f6f709"
+dependencies = [
+ "libc",
+]
+
+[[package]]
+name = "memchr"
+version = "2.6.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f665ee40bc4a3c5590afb1e9677db74a508659dfd71e126420da8274909a0167"
+
+[[package]]
+name = "minimal-lexical"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a"
+
+[[package]]
+name = "ndk"
+version = "0.7.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "451422b7e4718271c8b5b3aadf5adedba43dc76312454b387e98fae0fc951aa0"
+dependencies = [
+ "bitflags 1.3.2",
+ "jni-sys",
+ "ndk-sys",
+ "num_enum",
+ "raw-window-handle",
+ "thiserror",
+]
+
+[[package]]
+name = "ndk-context"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "27b02d87554356db9e9a873add8782d4ea6e3e58ea071a9adb9a2e8ddb884a8b"
+
+[[package]]
+name = "ndk-sys"
+version = "0.4.1+23.1.7779620"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3cf2aae958bd232cac5069850591667ad422d263686d75b52a065f9badeee5a3"
+dependencies = [
+ "jni-sys",
+]
+
+[[package]]
+name = "nix"
+version = "0.24.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "fa52e972a9a719cecb6864fb88568781eb706bac2cd1d4f04a648542dbf78069"
+dependencies = [
+ "bitflags 1.3.2",
+ "cfg-if",
+ "libc",
+]
+
+[[package]]
+name = "nom"
+version = "7.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a"
+dependencies = [
+ "memchr",
+ "minimal-lexical",
+]
+
+[[package]]
+name = "num-derive"
+version = "0.3.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "876a53fff98e03a936a674b29568b0e605f06b29372c2489ff4de23f1949743d"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn 1.0.109",
+]
+
+[[package]]
+name = "num-traits"
+version = "0.2.17"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "39e3200413f237f41ab11ad6d161bc7239c84dcb631773ccd7de3dfe4b5c267c"
+dependencies = [
+ "autocfg",
+]
+
+[[package]]
+name = "num_enum"
+version = "0.5.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1f646caf906c20226733ed5b1374287eb97e3c2a5c227ce668c1f2ce20ae57c9"
+dependencies = [
+ "num_enum_derive",
+]
+
+[[package]]
+name = "num_enum_derive"
+version = "0.5.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "dcbff9bc912032c62bf65ef1d5aea88983b420f4f839db1e9b0c281a25c9c799"
+dependencies = [
+ "proc-macro-crate",
+ "proc-macro2",
+ "quote",
+ "syn 1.0.109",
+]
+
+[[package]]
+name = "oboe"
+version = "0.5.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8868cc237ee02e2d9618539a23a8d228b9bb3fc2e7a5b11eed3831de77c395d0"
+dependencies = [
+ "jni 0.20.0",
+ "ndk",
+ "ndk-context",
+ "num-derive",
+ "num-traits",
+ "oboe-sys",
+]
+
+[[package]]
+name = "oboe-sys"
+version = "0.5.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7f44155e7fb718d3cfddcf70690b2b51ac4412f347cd9e4fbe511abe9cd7b5f2"
+dependencies = [
+ "cc",
+]
+
+[[package]]
+name = "once_cell"
+version = "1.19.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3fdb12b2476b595f9358c5161aa467c2438859caa136dec86c26fdd2efe17b92"
+
+[[package]]
+name = "parking_lot"
+version = "0.12.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3742b2c103b9f06bc9fff0a37ff4912935851bee6d36f3c02bcc755bcfec228f"
+dependencies = [
+ "lock_api",
+ "parking_lot_core",
+]
+
+[[package]]
+name = "parking_lot_core"
+version = "0.9.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4c42a9226546d68acdd9c0a280d17ce19bfe27a46bf68784e4066115788d008e"
+dependencies = [
+ "cfg-if",
+ "libc",
+ "redox_syscall",
+ "smallvec",
+ "windows-targets 0.48.5",
+]
+
+[[package]]
+name = "peeking_take_while"
+version = "0.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "19b17cddbe7ec3f8bc800887bab5e717348c95ea2ca0b1bf0837fb964dc67099"
+
+[[package]]
+name = "pkg-config"
+version = "0.3.27"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "26072860ba924cbfa98ea39c8c19b4dd6a4a25423dbdf219c1eca91aa0cf6964"
+
+[[package]]
+name = "proc-macro-crate"
+version = "1.3.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7f4c021e1093a56626774e81216a4ce732a735e5bad4868a03f3ed65ca0c3919"
+dependencies = [
+ "once_cell",
+ "toml_edit",
+]
+
+[[package]]
+name = "proc-macro2"
+version = "1.0.70"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "39278fbbf5fb4f646ce651690877f89d1c5811a3d4acb27700c1cb3cdb78fd3b"
+dependencies = [
+ "unicode-ident",
+]
+
+[[package]]
+name = "quote"
+version = "1.0.33"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5267fca4496028628a95160fc423a33e8b2e6af8a5302579e322e4b520293cae"
+dependencies = [
+ "proc-macro2",
+]
+
+[[package]]
+name = "raw-window-handle"
+version = "0.5.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f2ff9a1f06a88b01621b7ae906ef0211290d1c8a168a15542486a8f61c0833b9"
+
+[[package]]
+name = "redox_syscall"
+version = "0.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4722d768eff46b75989dd134e5c353f0d6296e5aaa3132e776cbdb56be7731aa"
+dependencies = [
+ "bitflags 1.3.2",
+]
+
+[[package]]
+name = "regex"
+version = "1.10.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "380b951a9c5e80ddfd6136919eef32310721aa4aacd4889a8d39124b026ab343"
+dependencies = [
+ "aho-corasick",
+ "memchr",
+ "regex-automata",
+ "regex-syntax",
+]
+
+[[package]]
+name = "regex-automata"
+version = "0.4.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5f804c7828047e88b2d32e2d7fe5a105da8ee3264f01902f796c8e067dc2483f"
+dependencies = [
+ "aho-corasick",
+ "memchr",
+ "regex-syntax",
+]
+
+[[package]]
+name = "regex-syntax"
+version = "0.8.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c08c74e62047bb2de4ff487b251e4a92e24f48745648451635cec7d591162d9f"
+
+[[package]]
+name = "rustc-hash"
+version = "1.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2"
+
+[[package]]
+name = "same-file"
+version = "1.0.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "93fc1dc3aaa9bfed95e02e6eadabb4baf7e3078b0bd1b4d7b6b0b68378900502"
+dependencies = [
+ "winapi-util",
+]
+
+[[package]]
+name = "scopeguard"
+version = "1.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "94143f37725109f92c262ed2cf5e59bce7498c01bcc1502d7b9afe439a4e9f49"
+
+[[package]]
+name = "shlex"
+version = "1.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a7cee0529a6d40f580e7a5e6c495c8fbfe21b7b52795ed4bb5e62cdf92bc6380"
+
+[[package]]
+name = "signal-hook"
+version = "0.3.17"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8621587d4798caf8eb44879d42e56b9a93ea5dcd315a6487c357130095b62801"
+dependencies = [
+ "libc",
+ "signal-hook-registry",
+]
+
+[[package]]
+name = "signal-hook-registry"
+version = "1.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d8229b473baa5980ac72ef434c4415e70c4b5e71b423043adb4ba059f89c99a1"
+dependencies = [
+ "libc",
+]
+
+[[package]]
+name = "smallvec"
+version = "1.11.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4dccd0940a2dcdf68d092b8cbab7dc0ad8fa938bf95787e1b916b0e3d0e8e970"
+
+[[package]]
+name = "syn"
+version = "1.0.109"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "72b64191b275b66ffe2469e8af2c1cfe3bafa67b529ead792a6d0160888b4237"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "unicode-ident",
+]
+
+[[package]]
+name = "syn"
+version = "2.0.41"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "44c8b28c477cc3bf0e7966561e3460130e1255f7a1cf71931075f1c5e7a7e269"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "unicode-ident",
+]
+
+[[package]]
+name = "thiserror"
+version = "1.0.51"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f11c217e1416d6f036b870f14e0413d480dbf28edbee1f877abaf0206af43bb7"
+dependencies = [
+ "thiserror-impl",
+]
+
+[[package]]
+name = "thiserror-impl"
+version = "1.0.51"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "01742297787513b79cf8e29d1056ede1313e2420b7b3b15d0a768b4921f549df"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn 2.0.41",
+]
+
+[[package]]
+name = "toml_datetime"
+version = "0.6.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3550f4e9685620ac18a50ed434eb3aec30db8ba93b0287467bca5826ea25baf1"
+
+[[package]]
+name = "toml_edit"
+version = "0.19.15"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1b5bb770da30e5cbfde35a2d7b9b8a2c4b8ef89548a7a6aeab5c9a576e3e7421"
+dependencies = [
+ "indexmap",
+ "toml_datetime",
+ "winnow",
+]
+
+[[package]]
+name = "unicode-ident"
+version = "1.0.12"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3354b9ac3fae1ff6755cb6db53683adb661634f67557942dea4facebec0fee4b"
+
+[[package]]
+name = "walkdir"
+version = "2.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d71d857dc86794ca4c280d616f7da00d2dbfd8cd788846559a6813e6aa4b54ee"
+dependencies = [
+ "same-file",
+ "winapi-util",
+]
+
+[[package]]
+name = "wasm-bindgen"
+version = "0.2.89"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0ed0d4f68a3015cc185aff4db9506a015f4b96f95303897bfa23f846db54064e"
+dependencies = [
+ "cfg-if",
+ "wasm-bindgen-macro",
+]
+
+[[package]]
+name = "wasm-bindgen-backend"
+version = "0.2.89"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1b56f625e64f3a1084ded111c4d5f477df9f8c92df113852fa5a374dbda78826"
+dependencies = [
+ "bumpalo",
+ "log",
+ "once_cell",
+ "proc-macro2",
+ "quote",
+ "syn 2.0.41",
+ "wasm-bindgen-shared",
+]
+
+[[package]]
+name = "wasm-bindgen-futures"
+version = "0.4.39"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ac36a15a220124ac510204aec1c3e5db8a22ab06fd6706d881dc6149f8ed9a12"
+dependencies = [
+ "cfg-if",
+ "js-sys",
+ "wasm-bindgen",
+ "web-sys",
+]
+
+[[package]]
+name = "wasm-bindgen-macro"
+version = "0.2.89"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0162dbf37223cd2afce98f3d0785506dcb8d266223983e4b5b525859e6e182b2"
+dependencies = [
+ "quote",
+ "wasm-bindgen-macro-support",
+]
+
+[[package]]
+name = "wasm-bindgen-macro-support"
+version = "0.2.89"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f0eb82fcb7930ae6219a7ecfd55b217f5f0893484b7a13022ebb2b2bf20b5283"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn 2.0.41",
+ "wasm-bindgen-backend",
+ "wasm-bindgen-shared",
+]
+
+[[package]]
+name = "wasm-bindgen-shared"
+version = "0.2.89"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7ab9b36309365056cd639da3134bf87fa8f3d86008abf99e612384a6eecd459f"
+
+[[package]]
+name = "web-sys"
+version = "0.3.66"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "50c24a44ec86bb68fbecd1b3efed7e85ea5621b39b35ef2766b66cd984f8010f"
+dependencies = [
+ "js-sys",
+ "wasm-bindgen",
+]
+
+[[package]]
+name = "winapi"
+version = "0.3.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5c839a674fcd7a98952e593242ea400abe93992746761e38641405d28b00f419"
+dependencies = [
+ "winapi-i686-pc-windows-gnu",
+ "winapi-x86_64-pc-windows-gnu",
+]
+
+[[package]]
+name = "winapi-i686-pc-windows-gnu"
+version = "0.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ac3b87c63620426dd9b991e5ce0329eff545bccbbb34f3be09ff6fb6ab51b7b6"
+
+[[package]]
+name = "winapi-util"
+version = "0.1.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f29e6f9198ba0d26b4c9f07dbe6f9ed633e1f3d5b8b414090084349e46a52596"
+dependencies = [
+ "winapi",
+]
+
+[[package]]
+name = "winapi-x86_64-pc-windows-gnu"
+version = "0.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f"
+
+[[package]]
+name = "windows"
+version = "0.46.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cdacb41e6a96a052c6cb63a144f24900236121c6f63f4f8219fef5977ecb0c25"
+dependencies = [
+ "windows-targets 0.42.2",
+]
+
+[[package]]
+name = "windows-targets"
+version = "0.42.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8e5180c00cd44c9b1c88adb3693291f1cd93605ded80c250a75d472756b4d071"
+dependencies = [
+ "windows_aarch64_gnullvm 0.42.2",
+ "windows_aarch64_msvc 0.42.2",
+ "windows_i686_gnu 0.42.2",
+ "windows_i686_msvc 0.42.2",
+ "windows_x86_64_gnu 0.42.2",
+ "windows_x86_64_gnullvm 0.42.2",
+ "windows_x86_64_msvc 0.42.2",
+]
+
+[[package]]
+name = "windows-targets"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9a2fa6e2155d7247be68c096456083145c183cbbbc2764150dda45a87197940c"
+dependencies = [
+ "windows_aarch64_gnullvm 0.48.5",
+ "windows_aarch64_msvc 0.48.5",
+ "windows_i686_gnu 0.48.5",
+ "windows_i686_msvc 0.48.5",
+ "windows_x86_64_gnu 0.48.5",
+ "windows_x86_64_gnullvm 0.48.5",
+ "windows_x86_64_msvc 0.48.5",
+]
+
+[[package]]
+name = "windows_aarch64_gnullvm"
+version = "0.42.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "597a5118570b68bc08d8d59125332c54f1ba9d9adeedeef5b99b02ba2b0698f8"
+
+[[package]]
+name = "windows_aarch64_gnullvm"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2b38e32f0abccf9987a4e3079dfb67dcd799fb61361e53e2882c3cbaf0d905d8"
+
+[[package]]
+name = "windows_aarch64_msvc"
+version = "0.42.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e08e8864a60f06ef0d0ff4ba04124db8b0fb3be5776a5cd47641e942e58c4d43"
+
+[[package]]
+name = "windows_aarch64_msvc"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "dc35310971f3b2dbbf3f0690a219f40e2d9afcf64f9ab7cc1be722937c26b4bc"
+
+[[package]]
+name = "windows_i686_gnu"
+version = "0.42.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c61d927d8da41da96a81f029489353e68739737d3beca43145c8afec9a31a84f"
+
+[[package]]
+name = "windows_i686_gnu"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a75915e7def60c94dcef72200b9a8e58e5091744960da64ec734a6c6e9b3743e"
+
+[[package]]
+name = "windows_i686_msvc"
+version = "0.42.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "44d840b6ec649f480a41c8d80f9c65108b92d89345dd94027bfe06ac444d1060"
+
+[[package]]
+name = "windows_i686_msvc"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8f55c233f70c4b27f66c523580f78f1004e8b5a8b659e05a4eb49d4166cca406"
+
+[[package]]
+name = "windows_x86_64_gnu"
+version = "0.42.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8de912b8b8feb55c064867cf047dda097f92d51efad5b491dfb98f6bbb70cb36"
+
+[[package]]
+name = "windows_x86_64_gnu"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "53d40abd2583d23e4718fddf1ebec84dbff8381c07cae67ff7768bbf19c6718e"
+
+[[package]]
+name = "windows_x86_64_gnullvm"
+version = "0.42.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "26d41b46a36d453748aedef1486d5c7a85db22e56aff34643984ea85514e94a3"
+
+[[package]]
+name = "windows_x86_64_gnullvm"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0b7b52767868a23d5bab768e390dc5f5c55825b6d30b86c844ff2dc7414044cc"
+
+[[package]]
+name = "windows_x86_64_msvc"
+version = "0.42.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9aec5da331524158c6d1a4ac0ab1541149c0b9505fde06423b02f5ef0106b9f0"
+
+[[package]]
+name = "windows_x86_64_msvc"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ed94fce61571a4006852b7389a063ab983c02eb1bb37b47f8272ce92d06d9538"
+
+[[package]]
+name = "winnow"
+version = "0.5.30"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9b5c3db89721d50d0e2a673f5043fc4722f76dcc352d7b1ab8b8288bed4ed2c5"
+dependencies = [
+ "memchr",
+]

listen-lib/Cargo.toml

@@ -1,11 +1,10 @@
 [package]
-name = "rust-whisper"
+name = "listen-lib"
 version = "0.1.0"
 edition = "2021"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-whisper-rs = "0.5"
-wav = "1"
-tokio = "1.27"
+signal-hook = "0.3.17"
+cpal = "0.15.2"

listen-lib/src/lib.rs

@@ -0,0 +1,135 @@
+use cpal::traits::{HostTrait, DeviceTrait, StreamTrait};
+use signal_hook::{iterator::Signals, consts::signal::SIGINT};
+use std::thread;
+
+pub fn main(callback: impl FnMut(Vec<f32>)) {
+   new_listener(None).unwrap().listen(callback);
+}
+
+pub fn main_with(callback: impl FnMut(Vec<f32>), device_name: String) {
+   new_listener(Some(device_name)).unwrap().listen(callback);
+}
+
+pub fn devices() -> Vec<String> {
+   match _devices() {
+      Ok(devices) => devices.
+         iter().
+         map(|device| device.name()).
+         filter(|device_name| device_name.is_ok()).
+         map(|device_name| device_name.unwrap()).
+         collect(),
+      Err(_) => vec![],
+   }
+}
+
+fn _devices() -> Result<Vec<cpal::Device>, String> {
+   match cpal::default_host().devices() {
+      Ok(devices) => Ok(devices.filter(|device| {
+         device.supported_input_configs().unwrap().count() > 0
+      }).collect()),
+      Err(msg) => Err(format!("failed to get devices: {}", msg)),
+   }
+}
+
+struct Listener {
+   device_name: String,
+}
+
+fn new_listener(device_name: Option<String>) -> Result<Listener, String> {
+   let device_name = match device_name {
+      Some(device_name) => {
+         let devices: Vec<_> = _devices()?.
+            iter().
+            map(|device| device.name()).
+            filter(|device_name| device_name.is_ok()).
+            map(|device_name| device_name.unwrap()).
+            filter(|a_device_name| *a_device_name == device_name).
+            collect();
+         match devices.first() {
+            Some(_) => Ok(device_name),
+            None => Err(format!("device like {} not found", device_name)),
+         }
+      },
+      None => {
+         let host = cpal::default_host();
+         match host.default_input_device() {
+            Some(device) => {
+               match device.name() {
+                  Ok(device_name) => Ok(device_name),
+                  Err(msg) => Err(format!("failed to get default input device name: {}", msg)),
+               }
+            },
+            None => Err(format!("failed to get any input device")),
+         }
+      },
+   };
+   match device_name {
+      Ok(device_name) => Ok(Listener{device_name: device_name}),
+      Err(msg) => Err(msg),
+   }
+}
+
+impl Listener {
+   fn listen(self, mut cb: impl FnMut(Vec<f32>)) {
+      let (send, recv) = std::sync::mpsc::sync_channel(100);
+      thread::spawn(move || { self._listen(send); });
+      let mut signals = Signals::new(&[SIGINT]).unwrap();
+      loop {
+         match recv.recv_timeout(std::time::Duration::new(1, 0)) {
+            Ok(msg) => cb(msg),
+            Err(_) => {
+               if signals.pending().count() > 0 {
+                  break;
+               }
+            },
+         };
+      }
+   }
+
+   fn _listen(self, send: std::sync::mpsc::SyncSender<Vec<f32>>) {
+      let devices = _devices().unwrap();
+      let devices = devices.iter().
+         filter(|device| device.name().unwrap() == self.device_name).
+         collect::<Vec<_>>();
+      let device = devices.first().unwrap();
+      let cfg = device.supported_input_configs()
+         .unwrap()
+         .filter(|x| x.sample_format() == cpal::SampleFormat::F32)
+         .filter(|x| x.min_sample_rate() >= cpal::SampleRate(15_500))
+         .nth(0)
+         .unwrap()
+         .with_max_sample_rate();
+
+      let downsample_ratio = cfg.channels() as f32 * (cfg.sample_rate().0 as f32 / 16_000.0);
+      let stream = device.build_input_stream(
+         &cfg.clone().into(),
+         move |data: &[f32], _: &cpal::InputCallbackInfo| {
+            let mut downsampled_data = vec![];
+            for i in 0..(data.len() as f32 / downsample_ratio) as usize {
+               let mut upsampled = i as f32 * downsample_ratio;
+               if upsampled > (data.len()-1) as f32 {
+                  upsampled = (data.len()-1) as f32
+               }
+               downsampled_data.push(data[upsampled as usize]);
+            }
+            match send.try_send(downsampled_data) {
+               Ok(_) => (),
+               Err(msg) => eprintln!("failed to ingest audio: {}", msg),
+            };
+         },
+         move |err| {
+            eprintln!("input error: {}", err)
+         },
+         None,
+      ).unwrap();
+      stream.play().unwrap();
+
+      eprintln!("listening on {}", device.name().unwrap());
+      let mut signals = Signals::new(&[SIGINT]).unwrap();
+      for sig in signals.forever() {
+         eprintln!("sig {}", sig);
+         break;
+      }
+      let _ = stream.pause();
+   }
+}

models/download_models.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+src="https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml"
+cd "$(dirname "$(realpath "$BASH_SOURCE")")"
+
+# Whisper models
+for model in "tiny.en" "base.en" "small.en" "medium.en" "large-v2"; do
+    	test -f ./ggml-$model.bin || wget --quiet --show-progress -O ./ggml-$model.bin "$src-$model.bin"
+done
+
+test -f ./ggml-distil-medium.en.bin || wget https://huggingface.co/distil-whisper/distil-medium.en/resolve/main/ggml-medium-32-2.en.bin?download=true -O ./ggml-distil-medium.en.bin
+
+test -f ./ggml-distil-large-v2.bin || wget https://huggingface.co/distil-whisper/distil-large-v2/resolve/main/ggml-large-32-2.en.bin?download=true -O ./ggml-distil-large-v2.bin

models/testme/Cargo.lock

@@ -0,0 +1,7 @@
+# This file is automatically @generated by Cargo.
+# It is not intended for manual editing.
+version = 3
+
+[[package]]
+name = "testme"
+version = "0.1.0"

models/testme/Cargo.toml

@@ -0,0 +1,8 @@
+[package]
+name = "testme"
+version = "0.1.0"
+edition = "2021"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]

models/testme/src/main.rs

@@ -0,0 +1,4 @@
+fn main() {
+   let bytes = include_bytes!("./test.txt");
+   println!("{}", String::from_utf8_lossy(bytes));
+}

models/testme/src/test.txt

@@ -0,0 +1 @@
+../../ggml-tiny.en.bin

rust-whisper-baked-lib-wasm/Cargo.lock

@@ -0,0 +1,857 @@
+# This file is automatically @generated by Cargo.
+# It is not intended for manual editing.
+version = 3
+
+[[package]]
+name = "addr2line"
+version = "0.21.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8a30b2e23b9e17a9f90641c7ab1549cd9b44f296d3ccbf309d2863cfe398a0cb"
+dependencies = [
+ "gimli",
+]
+
+[[package]]
+name = "adler"
+version = "1.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f26201604c87b1e01bd3d98f8d5d9a8fcbb815e8cedb41ffccbeb4bf593a35fe"
+
+[[package]]
+name = "aho-corasick"
+version = "1.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b2969dcb958b36655471fc61f7e416fa76033bdd4bfed0678d8fee1e2d07a1f0"
+dependencies = [
+ "memchr",
+]
+
+[[package]]
+name = "android-tzdata"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e999941b234f3131b00bc13c22d06e8c5ff726d1b6318ac7eb276997bbb4fef0"
+
+[[package]]
+name = "android_system_properties"
+version = "0.1.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "819e7219dbd41043ac279b19830f2efc897156490d7fd6ea916720117ee66311"
+dependencies = [
+ "libc",
+]
+
+[[package]]
+name = "anstream"
+version = "0.6.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d664a92ecae85fd0a7392615844904654d1d5f5514837f471ddef4a057aba1b6"
+dependencies = [
+ "anstyle",
+ "anstyle-parse",
+ "anstyle-query",
+ "anstyle-wincon",
+ "colorchoice",
+ "utf8parse",
+]
+
+[[package]]
+name = "anstyle"
+version = "1.0.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7079075b41f533b8c61d2a4d073c4676e1f8b249ff94a393b0595db304e0dd87"
+
+[[package]]
+name = "anstyle-parse"
+version = "0.2.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c75ac65da39e5fe5ab759307499ddad880d724eed2f6ce5b5e8a26f4f387928c"
+dependencies = [
+ "utf8parse",
+]
+
+[[package]]
+name = "anstyle-query"
+version = "1.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e28923312444cdd728e4738b3f9c9cac739500909bb3d3c94b43551b16517648"
+dependencies = [
+ "windows-sys",
+]
+
+[[package]]
+name = "anstyle-wincon"
+version = "3.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1cd54b81ec8d6180e24654d0b371ad22fc3dd083b6ff8ba325b72e00c87660a7"
+dependencies = [
+ "anstyle",
+ "windows-sys",
+]
+
+[[package]]
+name = "autocfg"
+version = "1.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d468802bab17cbc0cc575e9b053f41e72aa36bfa6b7f55e3529ffa43161b97fa"
+
+[[package]]
+name = "backtrace"
+version = "0.3.69"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2089b7e3f35b9dd2d0ed921ead4f6d318c27680d4a5bd167b3ee120edb105837"
+dependencies = [
+ "addr2line",
+ "cc",
+ "cfg-if",
+ "libc",
+ "miniz_oxide",
+ "object",
+ "rustc-demangle",
+]
+
+[[package]]
+name = "bindgen"
+version = "0.68.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "726e4313eb6ec35d2730258ad4e15b547ee75d6afaa1361a922e78e59b7d8078"
+dependencies = [
+ "bitflags",
+ "cexpr",
+ "clang-sys",
+ "lazy_static",
+ "lazycell",
+ "log",
+ "peeking_take_while",
+ "prettyplease",
+ "proc-macro2",
+ "quote",
+ "regex",
+ "rustc-hash",
+ "shlex",
+ "syn",
+ "which",
+]
+
+[[package]]
+name = "bitflags"
+version = "2.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "327762f6e5a765692301e5bb513e0d9fef63be86bbc14528052b1cd3e6f03e07"
+
+[[package]]
+name = "bumpalo"
+version = "3.14.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7f30e7476521f6f8af1a1c4c0b8cc94f0bee37d91763d0ca2665f299b6cd8aec"
+
+[[package]]
+name = "byteorder"
+version = "1.5.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1fd0f2584146f6f2ef48085050886acf353beff7305ebd1ae69500e27c67f64b"
+
+[[package]]
+name = "cc"
+version = "1.0.83"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f1174fb0b6ec23863f8b971027804a42614e347eafb0a95bf0b12cdae21fc4d0"
+dependencies = [
+ "libc",
+]
+
+[[package]]
+name = "cexpr"
+version = "0.6.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6fac387a98bb7c37292057cffc56d62ecb629900026402633ae9160df93a8766"
+dependencies = [
+ "nom",
+]
+
+[[package]]
+name = "cfg-if"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
+
+[[package]]
+name = "chrono"
+version = "0.4.31"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7f2c685bad3eb3d45a01354cedb7d5faa66194d1d58ba6e267a8de788f79db38"
+dependencies = [
+ "android-tzdata",
+ "iana-time-zone",
+ "js-sys",
+ "num-traits",
+ "wasm-bindgen",
+ "windows-targets 0.48.5",
+]
+
+[[package]]
+name = "clang-sys"
+version = "1.6.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c688fc74432808e3eb684cae8830a86be1d66a2bd58e1f248ed0960a590baf6f"
+dependencies = [
+ "glob",
+ "libc",
+ "libloading",
+]
+
+[[package]]
+name = "clap"
+version = "4.4.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bfaff671f6b22ca62406885ece523383b9b64022e341e53e009a62ebc47a45f2"
+dependencies = [
+ "clap_builder",
+ "clap_derive",
+]
+
+[[package]]
+name = "clap_builder"
+version = "4.4.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a216b506622bb1d316cd51328dce24e07bdff4a6128a47c7e7fad11878d5adbb"
+dependencies = [
+ "anstream",
+ "anstyle",
+ "clap_lex",
+ "strsim",
+]
+
+[[package]]
+name = "clap_derive"
+version = "4.4.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cf9804afaaf59a91e75b022a30fb7229a7901f60c755489cc61c9b423b836442"
+dependencies = [
+ "heck",
+ "proc-macro2",
+ "quote",
+ "syn",
+]
+
+[[package]]
+name = "clap_lex"
+version = "0.6.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "702fc72eb24e5a1e48ce58027a675bc24edd52096d5397d4aea7c6dd9eca0bd1"
+
+[[package]]
+name = "cmake"
+version = "0.1.50"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a31c789563b815f77f4250caee12365734369f942439b7defd71e18a48197130"
+dependencies = [
+ "cc",
+]
+
+[[package]]
+name = "colorchoice"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "acbf1af155f9b9ef647e42cdc158db4b64a1b61f743629225fde6f3e0be2a7c7"
+
+[[package]]
+name = "core-foundation-sys"
+version = "0.8.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "06ea2b9bc92be3c2baa9334a323ebca2d6f074ff852cd1d7b11064035cd3868f"
+
+[[package]]
+name = "either"
+version = "1.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a26ae43d7bcc3b814de94796a5e736d4029efb0ee900c12e2d54c993ad1a1e07"
+
+[[package]]
+name = "errno"
+version = "0.3.8"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a258e46cdc063eb8519c00b9fc845fc47bcfca4130e2f08e88665ceda8474245"
+dependencies = [
+ "libc",
+ "windows-sys",
+]
+
+[[package]]
+name = "fs_extra"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "42703706b716c37f96a77aea830392ad231f44c9e9a67872fa5548707e11b11c"
+
+[[package]]
+name = "gimli"
+version = "0.28.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4271d37baee1b8c7e4b708028c57d816cf9d2434acb33a549475f78c181f6253"
+
+[[package]]
+name = "glob"
+version = "0.3.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d2fabcfbdc87f4758337ca535fb41a6d701b65693ce38287d856d1674551ec9b"
+
+[[package]]
+name = "heck"
+version = "0.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "95505c38b4572b2d910cecb0281560f54b440a19336cbbcb27bf6ce6adc6f5a8"
+
+[[package]]
+name = "home"
+version = "0.5.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e3d1354bf6b7235cb4a0576c2619fd4ed18183f689b12b006a0ee7329eeff9a5"
+dependencies = [
+ "windows-sys",
+]
+
+[[package]]
+name = "iana-time-zone"
+version = "0.1.58"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8326b86b6cff230b97d0d312a6c40a60726df3332e721f72a1b035f451663b20"
+dependencies = [
+ "android_system_properties",
+ "core-foundation-sys",
+ "iana-time-zone-haiku",
+ "js-sys",
+ "wasm-bindgen",
+ "windows-core",
+]
+
+[[package]]
+name = "iana-time-zone-haiku"
+version = "0.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f31827a206f56af32e590ba56d5d2d085f558508192593743f16b2306495269f"
+dependencies = [
+ "cc",
+]
+
+[[package]]
+name = "js-sys"
+version = "0.3.66"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cee9c64da59eae3b50095c18d3e74f8b73c0b86d2792824ff01bbce68ba229ca"
+dependencies = [
+ "wasm-bindgen",
+]
+
+[[package]]
+name = "lazy_static"
+version = "1.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
+
+[[package]]
+name = "lazycell"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55"
+
+[[package]]
+name = "libc"
+version = "0.2.151"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "302d7ab3130588088d277783b1e2d2e10c9e9e4a16dd9050e6ec93fb3e7048f4"
+
+[[package]]
+name = "libloading"
+version = "0.7.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b67380fd3b2fbe7527a606e18729d21c6f3951633d0500574c4dc22d2d638b9f"
+dependencies = [
+ "cfg-if",
+ "winapi",
+]
+
+[[package]]
+name = "linux-raw-sys"
+version = "0.4.12"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c4cd1a83af159aa67994778be9070f0ae1bd732942279cabb14f86f986a21456"
+
+[[package]]
+name = "log"
+version = "0.4.20"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b5e6163cb8c49088c2c36f57875e58ccd8c87c7427f7fbd50ea6710b2f3f2e8f"
+
+[[package]]
+name = "memchr"
+version = "2.6.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f665ee40bc4a3c5590afb1e9677db74a508659dfd71e126420da8274909a0167"
+
+[[package]]
+name = "minimal-lexical"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a"
+
+[[package]]
+name = "miniz_oxide"
+version = "0.7.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e7810e0be55b428ada41041c41f32c9f1a42817901b4ccf45fa3d4b6561e74c7"
+dependencies = [
+ "adler",
+]
+
+[[package]]
+name = "nom"
+version = "7.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a"
+dependencies = [
+ "memchr",
+ "minimal-lexical",
+]
+
+[[package]]
+name = "num-traits"
+version = "0.2.17"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "39e3200413f237f41ab11ad6d161bc7239c84dcb631773ccd7de3dfe4b5c267c"
+dependencies = [
+ "autocfg",
+]
+
+[[package]]
+name = "object"
+version = "0.32.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9cf5f9dd3933bd50a9e1f149ec995f39ae2c496d31fd772c1fd45ebc27e902b0"
+dependencies = [
+ "memchr",
+]
+
+[[package]]
+name = "once_cell"
+version = "1.19.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3fdb12b2476b595f9358c5161aa467c2438859caa136dec86c26fdd2efe17b92"
+
+[[package]]
+name = "peeking_take_while"
+version = "0.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "19b17cddbe7ec3f8bc800887bab5e717348c95ea2ca0b1bf0837fb964dc67099"
+
+[[package]]
+name = "pin-project-lite"
+version = "0.2.13"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8afb450f006bf6385ca15ef45d71d2288452bc3683ce2e2cacc0d18e4be60b58"
+
+[[package]]
+name = "prettyplease"
+version = "0.2.15"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ae005bd773ab59b4725093fd7df83fd7892f7d8eafb48dbd7de6e024e4215f9d"
+dependencies = [
+ "proc-macro2",
+ "syn",
+]
+
+[[package]]
+name = "proc-macro2"
+version = "1.0.70"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "39278fbbf5fb4f646ce651690877f89d1c5811a3d4acb27700c1cb3cdb78fd3b"
+dependencies = [
+ "unicode-ident",
+]
+
+[[package]]
+name = "quote"
+version = "1.0.33"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5267fca4496028628a95160fc423a33e8b2e6af8a5302579e322e4b520293cae"
+dependencies = [
+ "proc-macro2",
+]
+
+[[package]]
+name = "regex"
+version = "1.10.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "380b951a9c5e80ddfd6136919eef32310721aa4aacd4889a8d39124b026ab343"
+dependencies = [
+ "aho-corasick",
+ "memchr",
+ "regex-automata",
+ "regex-syntax",
+]
+
+[[package]]
+name = "regex-automata"
+version = "0.4.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5f804c7828047e88b2d32e2d7fe5a105da8ee3264f01902f796c8e067dc2483f"
+dependencies = [
+ "aho-corasick",
+ "memchr",
+ "regex-syntax",
+]
+
+[[package]]
+name = "regex-syntax"
+version = "0.8.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c08c74e62047bb2de4ff487b251e4a92e24f48745648451635cec7d591162d9f"
+
+[[package]]
+name = "riff"
+version = "1.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b9b1a3d5f46d53f4a3478e2be4a5a5ce5108ea58b100dcd139830eae7f79a3a1"
+
+[[package]]
+name = "rust-whisper-baked-lib"
+version = "0.1.0"
+dependencies = [
+ "rust-whisper-lib",
+]
+
+[[package]]
+name = "rust-whisper-baked-lib-wasm"
+version = "0.1.0"
+dependencies = [
+ "rust-whisper-baked-lib",
+ "rust-whisper-lib",
+ "wasm-bindgen",
+]
+
+[[package]]
+name = "rust-whisper-lib"
+version = "0.1.0"
+dependencies = [
+ "byteorder",
+ "chrono",
+ "clap",
+ "tokio",
+ "wav",
+ "whisper-rs",
+]
+
+[[package]]
+name = "rustc-demangle"
+version = "0.1.23"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d626bb9dae77e28219937af045c257c28bfd3f69333c512553507f5f9798cb76"
+
+[[package]]
+name = "rustc-hash"
+version = "1.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2"
+
+[[package]]
+name = "rustix"
+version = "0.38.28"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "72e572a5e8ca657d7366229cdde4bd14c4eb5499a9573d4d366fe1b599daa316"
+dependencies = [
+ "bitflags",
+ "errno",
+ "libc",
+ "linux-raw-sys",
+ "windows-sys",
+]
+
+[[package]]
+name = "shlex"
+version = "1.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a7cee0529a6d40f580e7a5e6c495c8fbfe21b7b52795ed4bb5e62cdf92bc6380"
+
+[[package]]
+name = "strsim"
+version = "0.10.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "73473c0e59e6d5812c5dfe2a064a6444949f089e20eec9a2e5506596494e4623"
+
+[[package]]
+name = "syn"
+version = "2.0.41"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "44c8b28c477cc3bf0e7966561e3460130e1255f7a1cf71931075f1c5e7a7e269"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "unicode-ident",
+]
+
+[[package]]
+name = "tokio"
+version = "1.35.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c89b4efa943be685f629b149f53829423f8f5531ea21249408e8e2f8671ec104"
+dependencies = [
+ "backtrace",
+ "pin-project-lite",
+]
+
+[[package]]
+name = "unicode-ident"
+version = "1.0.12"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3354b9ac3fae1ff6755cb6db53683adb661634f67557942dea4facebec0fee4b"
+
+[[package]]
+name = "utf8parse"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "711b9620af191e0cdc7468a8d14e709c3dcdb115b36f838e601583af800a370a"
+
+[[package]]
+name = "wasm-bindgen"
+version = "0.2.89"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0ed0d4f68a3015cc185aff4db9506a015f4b96f95303897bfa23f846db54064e"
+dependencies = [
+ "cfg-if",
+ "wasm-bindgen-macro",
+]
+
+[[package]]
+name = "wasm-bindgen-backend"
+version = "0.2.89"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1b56f625e64f3a1084ded111c4d5f477df9f8c92df113852fa5a374dbda78826"
+dependencies = [
+ "bumpalo",
+ "log",
+ "once_cell",
+ "proc-macro2",
+ "quote",
+ "syn",
+ "wasm-bindgen-shared",
+]
+
+[[package]]
+name = "wasm-bindgen-macro"
+version = "0.2.89"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0162dbf37223cd2afce98f3d0785506dcb8d266223983e4b5b525859e6e182b2"
+dependencies = [
+ "quote",
+ "wasm-bindgen-macro-support",
+]
+
+[[package]]
+name = "wasm-bindgen-macro-support"
+version = "0.2.89"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f0eb82fcb7930ae6219a7ecfd55b217f5f0893484b7a13022ebb2b2bf20b5283"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn",
+ "wasm-bindgen-backend",
+ "wasm-bindgen-shared",
+]
+
+[[package]]
+name = "wasm-bindgen-shared"
+version = "0.2.89"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7ab9b36309365056cd639da3134bf87fa8f3d86008abf99e612384a6eecd459f"
+
+[[package]]
+name = "wav"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a65e199c799848b4f997072aa4d673c034f80f40191f97fe2f0a23f410be1609"
+dependencies = [
+ "riff",
+]
+
+[[package]]
+name = "which"
+version = "4.4.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "87ba24419a2078cd2b0f2ede2691b6c66d8e47836da3b6db8265ebad47afbfc7"
+dependencies = [
+ "either",
+ "home",
+ "once_cell",
+ "rustix",
+]
+
+[[package]]
+name = "whisper-rs"
+version = "0.8.0"
+dependencies = [
+ "whisper-rs-sys",
+]
+
+[[package]]
+name = "whisper-rs-sys"
+version = "0.7.3"
+dependencies = [
+ "bindgen",
+ "cfg-if",
+ "cmake",
+ "fs_extra",
+]
+
+[[package]]
+name = "winapi"
+version = "0.3.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5c839a674fcd7a98952e593242ea400abe93992746761e38641405d28b00f419"
+dependencies = [
+ "winapi-i686-pc-windows-gnu",
+ "winapi-x86_64-pc-windows-gnu",
+]
+
+[[package]]
+name = "winapi-i686-pc-windows-gnu"
+version = "0.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ac3b87c63620426dd9b991e5ce0329eff545bccbbb34f3be09ff6fb6ab51b7b6"
+
+[[package]]
+name = "winapi-x86_64-pc-windows-gnu"
+version = "0.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f"
+
+[[package]]
+name = "windows-core"
+version = "0.51.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f1f8cf84f35d2db49a46868f947758c7a1138116f7fac3bc844f43ade1292e64"
+dependencies = [
+ "windows-targets 0.48.5",
+]
+
+[[package]]
+name = "windows-sys"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "282be5f36a8ce781fad8c8ae18fa3f9beff57ec1b52cb3de0789201425d9a33d"
+dependencies = [
+ "windows-targets 0.52.0",
+]
+
+[[package]]
+name = "windows-targets"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9a2fa6e2155d7247be68c096456083145c183cbbbc2764150dda45a87197940c"
+dependencies = [
+ "windows_aarch64_gnullvm 0.48.5",
+ "windows_aarch64_msvc 0.48.5",
+ "windows_i686_gnu 0.48.5",
+ "windows_i686_msvc 0.48.5",
+ "windows_x86_64_gnu 0.48.5",
+ "windows_x86_64_gnullvm 0.48.5",
+ "windows_x86_64_msvc 0.48.5",
+]
+
+[[package]]
+name = "windows-targets"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8a18201040b24831fbb9e4eb208f8892e1f50a37feb53cc7ff887feb8f50e7cd"
+dependencies = [
+ "windows_aarch64_gnullvm 0.52.0",
+ "windows_aarch64_msvc 0.52.0",
+ "windows_i686_gnu 0.52.0",
+ "windows_i686_msvc 0.52.0",
+ "windows_x86_64_gnu 0.52.0",
+ "windows_x86_64_gnullvm 0.52.0",
+ "windows_x86_64_msvc 0.52.0",
+]
+
+[[package]]
+name = "windows_aarch64_gnullvm"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2b38e32f0abccf9987a4e3079dfb67dcd799fb61361e53e2882c3cbaf0d905d8"
+
+[[package]]
+name = "windows_aarch64_gnullvm"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cb7764e35d4db8a7921e09562a0304bf2f93e0a51bfccee0bd0bb0b666b015ea"
+
+[[package]]
+name = "windows_aarch64_msvc"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "dc35310971f3b2dbbf3f0690a219f40e2d9afcf64f9ab7cc1be722937c26b4bc"
+
+[[package]]
+name = "windows_aarch64_msvc"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bbaa0368d4f1d2aaefc55b6fcfee13f41544ddf36801e793edbbfd7d7df075ef"
+
+[[package]]
+name = "windows_i686_gnu"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a75915e7def60c94dcef72200b9a8e58e5091744960da64ec734a6c6e9b3743e"
+
+[[package]]
+name = "windows_i686_gnu"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a28637cb1fa3560a16915793afb20081aba2c92ee8af57b4d5f28e4b3e7df313"
+
+[[package]]
+name = "windows_i686_msvc"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8f55c233f70c4b27f66c523580f78f1004e8b5a8b659e05a4eb49d4166cca406"
+
+[[package]]
+name = "windows_i686_msvc"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ffe5e8e31046ce6230cc7215707b816e339ff4d4d67c65dffa206fd0f7aa7b9a"
+
+[[package]]
+name = "windows_x86_64_gnu"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "53d40abd2583d23e4718fddf1ebec84dbff8381c07cae67ff7768bbf19c6718e"
+
+[[package]]
+name = "windows_x86_64_gnu"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3d6fa32db2bc4a2f5abeacf2b69f7992cd09dca97498da74a151a3132c26befd"
+
+[[package]]
+name = "windows_x86_64_gnullvm"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0b7b52767868a23d5bab768e390dc5f5c55825b6d30b86c844ff2dc7414044cc"
+
+[[package]]
+name = "windows_x86_64_gnullvm"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1a657e1e9d3f514745a572a6846d3c7aa7dbe1658c056ed9c3344c4109a6949e"
+
+[[package]]
+name = "windows_x86_64_msvc"
+version = "0.48.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ed94fce61571a4006852b7389a063ab983c02eb1bb37b47f8272ce92d06d9538"
+
+[[package]]
+name = "windows_x86_64_msvc"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "dff9641d1cd4be8d1a070daf9e3773c5f67e78b4d9d42263020c057706765c04"

rust-whisper-baked-lib-wasm/Cargo.toml

@@ -0,0 +1,12 @@
+[package]
+name = "rust-whisper-baked-lib-wasm"
+version = "0.1.0"
+edition = "2021"
+
+[lib]
+crate-type = ["cdylib"]
+
+[dependencies]
+rust-whisper-lib = { path = "../rust-whisper-lib" }
+rust-whisper-baked-lib = { path = "../rust-whisper-baked-lib" }
+wasm-bindgen = "0.2.89"

rust-whisper-baked-lib-wasm/build.sh

@@ -0,0 +1,5 @@
+#! /bin/bash
+
+echo https://developer.mozilla.org/en-US/docs/WebAssembly/Rust_to_Wasm
+cargo install wasm-pack
+wasm-pack build --target web

rust-whisper-baked-lib-wasm/src/lib.rs

@@ -0,0 +1,43 @@
+use wasm_bindgen::prelude::*;
+use rust_whisper_lib;
+use rust_whisper_baked_lib;
+
+#[wasm_bindgen]
+extern {
+    pub fn whisper_data() -> Vec<f32>;
+    pub fn whisper_on_success(s: String);
+    pub fn whisper_on_error(s: String);
+}
+
+#[wasm_bindgen]
+pub fn new_transcriber(
+   stream_step: Option<u64>,
+   stream_retain: Option<f32>,
+   stream_head: Option<f32>,
+   stream_tail: Option<f32>,
+) {
+   let flags = rust_whisper_lib::Flags {
+      model_path: None,
+      model_buffer: None,
+      threads: 4,
+      stream_step: stream_step.unwrap_or(5),
+      stream_retain: stream_retain.unwrap_or(1.0),
+      stream_head: stream_head.unwrap_or(0.5),
+      stream_tail: stream_tail.unwrap_or(0.5),
+      wav: None,
+      debug: false,
+      stream_device: None,
+   };
+   let (send, recv) = std::sync::mpsc::sync_channel(100);
+   rust_whisper_baked_lib::channel(flags, |result| {
+      match result {
+         Ok(msg) => whisper_on_success(msg.to_string()),
+         Err(msg) => whisper_on_error(msg),
+      };
+   }, recv);
+   loop {
+      let data = whisper_data();
+      send.send(data).unwrap();
+   }
+}
+

rust-whisper-baked-lib/Cargo.lock

@@ -0,0 +1,555 @@
+# This file is automatically @generated by Cargo.
+# It is not intended for manual editing.
+version = 3
+
+[[package]]
+name = "aho-corasick"
+version = "1.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b2969dcb958b36655471fc61f7e416fa76033bdd4bfed0678d8fee1e2d07a1f0"
+dependencies = [
+ "memchr",
+]
+
+[[package]]
+name = "anstream"
+version = "0.6.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d664a92ecae85fd0a7392615844904654d1d5f5514837f471ddef4a057aba1b6"
+dependencies = [
+ "anstyle",
+ "anstyle-parse",
+ "anstyle-query",
+ "anstyle-wincon",
+ "colorchoice",
+ "utf8parse",
+]
+
+[[package]]
+name = "anstyle"
+version = "1.0.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7079075b41f533b8c61d2a4d073c4676e1f8b249ff94a393b0595db304e0dd87"
+
+[[package]]
+name = "anstyle-parse"
+version = "0.2.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c75ac65da39e5fe5ab759307499ddad880d724eed2f6ce5b5e8a26f4f387928c"
+dependencies = [
+ "utf8parse",
+]
+
+[[package]]
+name = "anstyle-query"
+version = "1.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e28923312444cdd728e4738b3f9c9cac739500909bb3d3c94b43551b16517648"
+dependencies = [
+ "windows-sys",
+]
+
+[[package]]
+name = "anstyle-wincon"
+version = "3.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1cd54b81ec8d6180e24654d0b371ad22fc3dd083b6ff8ba325b72e00c87660a7"
+dependencies = [
+ "anstyle",
+ "windows-sys",
+]
+
+[[package]]
+name = "bindgen"
+version = "0.68.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "726e4313eb6ec35d2730258ad4e15b547ee75d6afaa1361a922e78e59b7d8078"
+dependencies = [
+ "bitflags",
+ "cexpr",
+ "clang-sys",
+ "lazy_static",
+ "lazycell",
+ "log",
+ "peeking_take_while",
+ "prettyplease",
+ "proc-macro2",
+ "quote",
+ "regex",
+ "rustc-hash",
+ "shlex",
+ "syn",
+ "which",
+]
+
+[[package]]
+name = "bitflags"
+version = "2.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "327762f6e5a765692301e5bb513e0d9fef63be86bbc14528052b1cd3e6f03e07"
+
+[[package]]
+name = "cc"
+version = "1.0.83"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f1174fb0b6ec23863f8b971027804a42614e347eafb0a95bf0b12cdae21fc4d0"
+dependencies = [
+ "libc",
+]
+
+[[package]]
+name = "cexpr"
+version = "0.6.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6fac387a98bb7c37292057cffc56d62ecb629900026402633ae9160df93a8766"
+dependencies = [
+ "nom",
+]
+
+[[package]]
+name = "cfg-if"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
+
+[[package]]
+name = "clang-sys"
+version = "1.6.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c688fc74432808e3eb684cae8830a86be1d66a2bd58e1f248ed0960a590baf6f"
+dependencies = [
+ "glob",
+ "libc",
+ "libloading",
+]
+
+[[package]]
+name = "clap"
+version = "4.4.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bfaff671f6b22ca62406885ece523383b9b64022e341e53e009a62ebc47a45f2"
+dependencies = [
+ "clap_builder",
+ "clap_derive",
+]
+
+[[package]]
+name = "clap_builder"
+version = "4.4.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a216b506622bb1d316cd51328dce24e07bdff4a6128a47c7e7fad11878d5adbb"
+dependencies = [
+ "anstream",
+ "anstyle",
+ "clap_lex",
+ "strsim",
+]
+
+[[package]]
+name = "clap_derive"
+version = "4.4.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cf9804afaaf59a91e75b022a30fb7229a7901f60c755489cc61c9b423b836442"
+dependencies = [
+ "heck",
+ "proc-macro2",
+ "quote",
+ "syn",
+]
+
+[[package]]
+name = "clap_lex"
+version = "0.6.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "702fc72eb24e5a1e48ce58027a675bc24edd52096d5397d4aea7c6dd9eca0bd1"
+
+[[package]]
+name = "cmake"
+version = "0.1.50"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a31c789563b815f77f4250caee12365734369f942439b7defd71e18a48197130"
+dependencies = [
+ "cc",
+]
+
+[[package]]
+name = "colorchoice"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "acbf1af155f9b9ef647e42cdc158db4b64a1b61f743629225fde6f3e0be2a7c7"
+
+[[package]]
+name = "either"
+version = "1.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a26ae43d7bcc3b814de94796a5e736d4029efb0ee900c12e2d54c993ad1a1e07"
+
+[[package]]
+name = "errno"
+version = "0.3.8"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a258e46cdc063eb8519c00b9fc845fc47bcfca4130e2f08e88665ceda8474245"
+dependencies = [
+ "libc",
+ "windows-sys",
+]
+
+[[package]]
+name = "fs_extra"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "42703706b716c37f96a77aea830392ad231f44c9e9a67872fa5548707e11b11c"
+
+[[package]]
+name = "glob"
+version = "0.3.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d2fabcfbdc87f4758337ca535fb41a6d701b65693ce38287d856d1674551ec9b"
+
+[[package]]
+name = "heck"
+version = "0.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "95505c38b4572b2d910cecb0281560f54b440a19336cbbcb27bf6ce6adc6f5a8"
+
+[[package]]
+name = "home"
+version = "0.5.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e3d1354bf6b7235cb4a0576c2619fd4ed18183f689b12b006a0ee7329eeff9a5"
+dependencies = [
+ "windows-sys",
+]
+
+[[package]]
+name = "lazy_static"
+version = "1.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
+
+[[package]]
+name = "lazycell"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55"
+
+[[package]]
+name = "libc"
+version = "0.2.151"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "302d7ab3130588088d277783b1e2d2e10c9e9e4a16dd9050e6ec93fb3e7048f4"
+
+[[package]]
+name = "libloading"
+version = "0.7.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b67380fd3b2fbe7527a606e18729d21c6f3951633d0500574c4dc22d2d638b9f"
+dependencies = [
+ "cfg-if",
+ "winapi",
+]
+
+[[package]]
+name = "linux-raw-sys"
+version = "0.4.12"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c4cd1a83af159aa67994778be9070f0ae1bd732942279cabb14f86f986a21456"
+
+[[package]]
+name = "log"
+version = "0.4.20"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b5e6163cb8c49088c2c36f57875e58ccd8c87c7427f7fbd50ea6710b2f3f2e8f"
+
+[[package]]
+name = "memchr"
+version = "2.6.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f665ee40bc4a3c5590afb1e9677db74a508659dfd71e126420da8274909a0167"
+
+[[package]]
+name = "minimal-lexical"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a"
+
+[[package]]
+name = "nom"
+version = "7.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a"
+dependencies = [
+ "memchr",
+ "minimal-lexical",
+]
+
+[[package]]
+name = "once_cell"
+version = "1.19.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3fdb12b2476b595f9358c5161aa467c2438859caa136dec86c26fdd2efe17b92"
+
+[[package]]
+name = "peeking_take_while"
+version = "0.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "19b17cddbe7ec3f8bc800887bab5e717348c95ea2ca0b1bf0837fb964dc67099"
+
+[[package]]
+name = "prettyplease"
+version = "0.2.15"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ae005bd773ab59b4725093fd7df83fd7892f7d8eafb48dbd7de6e024e4215f9d"
+dependencies = [
+ "proc-macro2",
+ "syn",
+]
+
+[[package]]
+name = "proc-macro2"
+version = "1.0.70"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "39278fbbf5fb4f646ce651690877f89d1c5811a3d4acb27700c1cb3cdb78fd3b"
+dependencies = [
+ "unicode-ident",
+]
+
+[[package]]
+name = "quote"
+version = "1.0.33"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5267fca4496028628a95160fc423a33e8b2e6af8a5302579e322e4b520293cae"
+dependencies = [
+ "proc-macro2",
+]
+
+[[package]]
+name = "regex"
+version = "1.10.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "380b951a9c5e80ddfd6136919eef32310721aa4aacd4889a8d39124b026ab343"
+dependencies = [
+ "aho-corasick",
+ "memchr",
+ "regex-automata",
+ "regex-syntax",
+]
+
+[[package]]
+name = "regex-automata"
+version = "0.4.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5f804c7828047e88b2d32e2d7fe5a105da8ee3264f01902f796c8e067dc2483f"
+dependencies = [
+ "aho-corasick",
+ "memchr",
+ "regex-syntax",
+]
+
+[[package]]
+name = "regex-syntax"
+version = "0.8.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c08c74e62047bb2de4ff487b251e4a92e24f48745648451635cec7d591162d9f"
+
+[[package]]
+name = "riff"
+version = "1.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b9b1a3d5f46d53f4a3478e2be4a5a5ce5108ea58b100dcd139830eae7f79a3a1"
+
+[[package]]
+name = "rust-whisper-baked-lib"
+version = "0.1.0"
+dependencies = [
+ "rust-whisper-lib",
+]
+
+[[package]]
+name = "rust-whisper-lib"
+version = "0.1.0"
+dependencies = [
+ "clap",
+ "wav",
+ "whisper-rs",
+]
+
+[[package]]
+name = "rustc-hash"
+version = "1.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2"
+
+[[package]]
+name = "rustix"
+version = "0.38.28"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "72e572a5e8ca657d7366229cdde4bd14c4eb5499a9573d4d366fe1b599daa316"
+dependencies = [
+ "bitflags",
+ "errno",
+ "libc",
+ "linux-raw-sys",
+ "windows-sys",
+]
+
+[[package]]
+name = "shlex"
+version = "1.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a7cee0529a6d40f580e7a5e6c495c8fbfe21b7b52795ed4bb5e62cdf92bc6380"
+
+[[package]]
+name = "strsim"
+version = "0.10.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "73473c0e59e6d5812c5dfe2a064a6444949f089e20eec9a2e5506596494e4623"
+
+[[package]]
+name = "syn"
+version = "2.0.41"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "44c8b28c477cc3bf0e7966561e3460130e1255f7a1cf71931075f1c5e7a7e269"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "unicode-ident",
+]
+
+[[package]]
+name = "unicode-ident"
+version = "1.0.12"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3354b9ac3fae1ff6755cb6db53683adb661634f67557942dea4facebec0fee4b"
+
+[[package]]
+name = "utf8parse"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "711b9620af191e0cdc7468a8d14e709c3dcdb115b36f838e601583af800a370a"
+
+[[package]]
+name = "wav"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a65e199c799848b4f997072aa4d673c034f80f40191f97fe2f0a23f410be1609"
+dependencies = [
+ "riff",
+]
+
+[[package]]
+name = "which"
+version = "4.4.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "87ba24419a2078cd2b0f2ede2691b6c66d8e47836da3b6db8265ebad47afbfc7"
+dependencies = [
+ "either",
+ "home",
+ "once_cell",
+ "rustix",
+]
+
+[[package]]
+name = "whisper-rs"
+version = "0.8.0"
+dependencies = [
+ "whisper-rs-sys",
+]
+
+[[package]]
+name = "whisper-rs-sys"
+version = "0.7.3"
+dependencies = [
+ "bindgen",
+ "cfg-if",
+ "cmake",
+ "fs_extra",
+]
+
+[[package]]
+name = "winapi"
+version = "0.3.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5c839a674fcd7a98952e593242ea400abe93992746761e38641405d28b00f419"
+dependencies = [
+ "winapi-i686-pc-windows-gnu",
+ "winapi-x86_64-pc-windows-gnu",
+]
+
+[[package]]
+name = "winapi-i686-pc-windows-gnu"
+version = "0.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ac3b87c63620426dd9b991e5ce0329eff545bccbbb34f3be09ff6fb6ab51b7b6"
+
+[[package]]
+name = "winapi-x86_64-pc-windows-gnu"
+version = "0.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f"
+
+[[package]]
+name = "windows-sys"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "282be5f36a8ce781fad8c8ae18fa3f9beff57ec1b52cb3de0789201425d9a33d"
+dependencies = [
+ "windows-targets",
+]
+
+[[package]]
+name = "windows-targets"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8a18201040b24831fbb9e4eb208f8892e1f50a37feb53cc7ff887feb8f50e7cd"
+dependencies = [
+ "windows_aarch64_gnullvm",
+ "windows_aarch64_msvc",
+ "windows_i686_gnu",
+ "windows_i686_msvc",
+ "windows_x86_64_gnu",
+ "windows_x86_64_gnullvm",
+ "windows_x86_64_msvc",
+]
+
+[[package]]
+name = "windows_aarch64_gnullvm"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cb7764e35d4db8a7921e09562a0304bf2f93e0a51bfccee0bd0bb0b666b015ea"
+
+[[package]]
+name = "windows_aarch64_msvc"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bbaa0368d4f1d2aaefc55b6fcfee13f41544ddf36801e793edbbfd7d7df075ef"
+
+[[package]]
+name = "windows_i686_gnu"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a28637cb1fa3560a16915793afb20081aba2c92ee8af57b4d5f28e4b3e7df313"
+
+[[package]]
+name = "windows_i686_msvc"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ffe5e8e31046ce6230cc7215707b816e339ff4d4d67c65dffa206fd0f7aa7b9a"
+
+[[package]]
+name = "windows_x86_64_gnu"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3d6fa32db2bc4a2f5abeacf2b69f7992cd09dca97498da74a151a3132c26befd"
+
+[[package]]
+name = "windows_x86_64_gnullvm"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1a657e1e9d3f514745a572a6846d3c7aa7dbe1658c056ed9c3344c4109a6949e"
+
+[[package]]
+name = "windows_x86_64_msvc"
+version = "0.52.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "dff9641d1cd4be8d1a070daf9e3773c5f67e78b4d9d42263020c057706765c04"

rust-whisper-baked-lib/Cargo.toml

@@ -0,0 +1,9 @@
+[package]
+name = "rust-whisper-baked-lib"
+version = "0.1.0"
+edition = "2021"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+rust-whisper-lib = { path = "../rust-whisper-lib" }

rust-whisper-baked-lib/src/lib.rs

@@ -0,0 +1,33 @@
+use rust_whisper_lib;
+
+pub fn channel<F>(
+   mut flags: rust_whisper_lib::Flags,
+   handler_fn: F,
+   stream: std::sync::mpsc::Receiver<Vec<f32>>,
+) where F: FnMut(Result<rust_whisper_lib::Transcribed, String>) + Send + 'static {
+   flags.model_path = None;
+   flags.model_buffer = Some(include_bytes!("../../models/ggml-tiny.en.bin").to_vec());
+   rust_whisper_lib::channel(flags.clone(), handler_fn, stream);
+}
+
+pub fn wav<F>(
+   mut flags: rust_whisper_lib::Flags,
+   handler_fn: F
+) where F: FnMut(Result<rust_whisper_lib::Transcribed, String>) + Send + 'static {
+   flags.model_path = None;
+   flags.model_buffer = Some(include_bytes!("../../models/ggml-distil-medium.en.bin").to_vec());
+   rust_whisper_lib::wav(flags.clone(), handler_fn, flags.wav.unwrap());
+}
+
+pub fn wav_channel<F>(
+   mut flags: rust_whisper_lib::Flags,
+   handler_fn: F
+) where F: FnMut(Result<rust_whisper_lib::Transcribed, String>) + Send + 'static {
+   flags.model_path = None;
+   flags.model_buffer = Some(include_bytes!("../../models/ggml-base.en.bin").to_vec());
+   rust_whisper_lib::wav_channel(flags, handler_fn);
+}
+
+pub fn f32_from_wav_file(path: &String) -> Result<Vec<f32>, String> {
+   rust_whisper_lib::f32_from_wav_file(path)
+}

rust-whisper-baked/Cargo.toml

@@ -0,0 +1,12 @@
+[package]
+name = "rust-whisper-baked"
+version = "0.1.0"
+edition = "2021"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+rust-whisper-lib = { path = "../rust-whisper-lib" }
+rust-whisper-baked-lib = { path = "../rust-whisper-baked-lib" }
+listen-lib = { path = "../listen-lib" }
+clap = { version = "4.4.10", features = ["derive"] }

rust-whisper-baked/cargo.sh

@@ -0,0 +1,15 @@
+#! /bin/bash
+
+#RUSTFLAGS="-Clink-args=-lstd++" \
+
+echo 0. brew install blackhole-2ch / https://github.com/ExistentialAudio/BlackHole
+echo 1. Audio Midi Setup / https://existential.audio/howto/StreamFromGarageBandToZoom.php
+echo 2. Multi-Output Device / primary External Headphones, Drift Correction for BlackHole 2ch
+echo 3. Aggregate Device / clock MacBook Pro Microphone, Drift Correction BlackHole 2ch
+echo 4. Discord outputs to Multi-Output Device
+echo 5. apps read Aggregate Device
+
+export C_INCLUDE_PATH="$C_INCLUDE_PATH:$PWD/.."
+export LIBRARY_PATH="$LIBRARY_PATH:$PWD/.."
+
+cargo "$@"

rust-whisper-baked/src/main.rs

@@ -0,0 +1,147 @@
+use rust_whisper_lib;
+use rust_whisper_baked_lib;
+use clap::Parser;
+use listen_lib;
+use std::thread;
+
+fn main() {
+   let flags = rust_whisper_lib::Flags::parse();
+   match flags.wav.clone() {
+      Some(_) => wav_channel(flags),
+      None => channel(flags),
+   };
+}
+
+fn wav_channel(flags: rust_whisper_lib::Flags) {
+   let mut w = new_destutterer();
+   rust_whisper_baked_lib::wav_channel(
+      flags.clone(),
+      move |result: Result<rust_whisper_lib::Transcribed, String>| {
+         match result {
+            Ok(transcribed) => {
+               let s = w.step(transcribed.to_string());
+               println!("{}", s);
+            },
+            Err(msg) => { eprintln!("error: {}", msg); },
+         };
+      },
+   );
+}
+
+fn wav(flags: rust_whisper_lib::Flags, _path: String) {
+   let mut w = new_destutterer();
+   rust_whisper_baked_lib::wav(flags,
+      move |result: Result<rust_whisper_lib::Transcribed, String>| {
+         match result {
+            Ok(transcribed) => {
+               let s = w.step(transcribed.to_string());
+               println!("{}", s);
+            },
+            Err(msg) => { eprintln!("error: {}", msg); },
+         };
+      },
+   );
+}
+
+fn channel(flags: rust_whisper_lib::Flags) {
+   let (send, recv) = std::sync::mpsc::sync_channel(100);
+
+   eprintln!("rust whisper baked lib channel...");
+   thread::spawn(move || {
+      rust_whisper_baked_lib::channel(
+         flags.clone(),
+         |result: Result<rust_whisper_lib::Transcribed, String>| {
+            match result {
+               Ok(transcribed) => { println!("{}", transcribed.to_string()); },
+               Err(msg) => { eprintln!("error: {}", msg); },
+            };
+         },
+         recv,
+      );
+   });
+
+   eprintln!("listen lib main...");
+   let flags = rust_whisper_lib::Flags::parse();
+   match flags.stream_device {
+      Some(device_name) => {
+         if device_name == "" {
+            for device in listen_lib::devices() {
+               eprintln!("{}", device);
+            }
+         } else {
+            listen_lib::main_with(|data| {
+               send.send(data).unwrap();
+            }, device_name);
+         }
+      },
+      None => {
+         listen_lib::main(|data| {
+            send.send(data).unwrap();
+         });
+      }
+   }
+   eprintln!("/listen lib main...");
+}
+
+struct Destutterer {
+   prev: Option<String>,
+}
+
+fn new_destutterer() -> Destutterer {
+   Destutterer{prev: None}
+}
+
+impl Destutterer {
+   fn step(&mut self, next: String) -> String {
+      let next = next.trim().to_string();
+      if next.len() == 0 {
+         return next;
+      }
+      match &self.prev {
+         None => {
+            self.prev = Some(next.clone());
+            next
+         },
+         Some(prev) => {
+            let without_trailing_punctuation = {
+               let mut next = next.clone();
+               while next.ends_with("?") || next.ends_with(".") {
+                  next = next[..next.len()-1].to_string();
+               }
+               next
+            };
+            let trailing_punctuation = next[without_trailing_punctuation.len() ..].to_string();
+            let next = without_trailing_punctuation;
+            let next = {
+               let mut n = prev.len().clamp(0, next.len());
+               while n > 0 {
+                  if prev[prev.len() - n..] == next[..n] {
+                     break;
+                  }
+                  n -= 1;
+               }
+               next[n..].to_string()
+            };
+            if next.len() == 0 {
+               return "".to_string();
+            }
+            self.prev = Some(next.clone());
+            next + &trailing_punctuation
+         },
+      }
+   }
+}
+
+#[cfg(test)]
+mod tests {
+   use super::*;
+
+   #[test]
+   fn test_destutterer() {
+      let mut w = new_destutterer();
+      assert_eq!("abcde".to_string(), w.step("abcde".to_string()));
+      assert_eq!("fg".to_string(), w.step("cdefg".to_string()));
+      assert_eq!("hij".to_string(), w.step("fghij".to_string()));
+      assert_eq!("fghij".to_string(), w.step("fghij".to_string()));
+   }
+}

rust-whisper-lib/Cargo.lock

@@ -3,16 +3,67 @@
 version = 3
 
 [[package]]
-name = "autocfg"
-version = "1.1.0"
+name = "aho-corasick"
+version = "1.0.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d468802bab17cbc0cc575e9b053f41e72aa36bfa6b7f55e3529ffa43161b97fa"
+checksum = "43f6cb1bf222025340178f382c426f13757b2960e89779dfcb319c32542a5a41"
+dependencies = [
+ "memchr",
+]
+
+[[package]]
+name = "anstream"
+version = "0.6.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d664a92ecae85fd0a7392615844904654d1d5f5514837f471ddef4a057aba1b6"
+dependencies = [
+ "anstyle",
+ "anstyle-parse",
+ "anstyle-query",
+ "anstyle-wincon",
+ "colorchoice",
+ "utf8parse",
+]
+
+[[package]]
+name = "anstyle"
+version = "1.0.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7079075b41f533b8c61d2a4d073c4676e1f8b249ff94a393b0595db304e0dd87"
+
+[[package]]
+name = "anstyle-parse"
+version = "0.2.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c75ac65da39e5fe5ab759307499ddad880d724eed2f6ce5b5e8a26f4f387928c"
+dependencies = [
+ "utf8parse",
+]
+
+[[package]]
+name = "anstyle-query"
+version = "1.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e28923312444cdd728e4738b3f9c9cac739500909bb3d3c94b43551b16517648"
+dependencies = [
+ "windows-sys",
+]
+
+[[package]]
+name = "anstyle-wincon"
+version = "3.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1cd54b81ec8d6180e24654d0b371ad22fc3dd083b6ff8ba325b72e00c87660a7"
+dependencies = [
+ "anstyle",
+ "windows-sys",
+]
 
 [[package]]
 name = "bindgen"
-version = "0.64.0"
+version = "0.68.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c4243e6031260db77ede97ad86c27e501d646a27ab57b59a574f725d98ab1fb4"
+checksum = "726e4313eb6ec35d2730258ad4e15b547ee75d6afaa1361a922e78e59b7d8078"
 dependencies = [
  "bitflags",
  "cexpr",
@@ -21,6 +72,7 @@ dependencies = [
  "lazycell",
  "log",
  "peeking_take_while",
+ "prettyplease",
  "proc-macro2",
  "quote",
  "regex",
@@ -32,9 +84,15 @@ dependencies = [
 
 [[package]]
 name = "bitflags"
-version = "1.3.2"
+version = "2.3.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a"
+checksum = "630be753d4e58660abd17930c71b647fe46c27ea6b63cc59e1e3851406972e42"
+
+[[package]]
+name = "cc"
+version = "1.0.79"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "50d30906286121d95be3d479533b458f87493b30a4b5f79a607db8f5d11aa91f"
 
 [[package]]
 name = "cexpr"
@@ -63,10 +121,71 @@ dependencies = [
 ]
 
 [[package]]
-name = "either"
-version = "1.8.1"
+name = "clap"
+version = "4.4.11"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7fcaabb2fef8c910e7f4c7ce9f67a1283a1715879a7c230ca9d6d1ae31f16d91"
+checksum = "bfaff671f6b22ca62406885ece523383b9b64022e341e53e009a62ebc47a45f2"
+dependencies = [
+ "clap_builder",
+ "clap_derive",
+]
+
+[[package]]
+name = "clap_builder"
+version = "4.4.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a216b506622bb1d316cd51328dce24e07bdff4a6128a47c7e7fad11878d5adbb"
+dependencies = [
+ "anstream",
+ "anstyle",
+ "clap_lex",
+ "strsim",
+]
+
+[[package]]
+name = "clap_derive"
+version = "4.4.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cf9804afaaf59a91e75b022a30fb7229a7901f60c755489cc61c9b423b836442"
+dependencies = [
+ "heck",
+ "proc-macro2",
+ "quote",
+ "syn",
+]
+
+[[package]]
+name = "clap_lex"
+version = "0.6.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "702fc72eb24e5a1e48ce58027a675bc24edd52096d5397d4aea7c6dd9eca0bd1"
+
+[[package]]
+name = "cmake"
+version = "0.1.50"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a31c789563b815f77f4250caee12365734369f942439b7defd71e18a48197130"
+dependencies = [
+ "cc",
+]
+
+[[package]]
+name = "colorchoice"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "acbf1af155f9b9ef647e42cdc158db4b64a1b61f743629225fde6f3e0be2a7c7"
+
+[[package]]
+name = "either"
+version = "1.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a26ae43d7bcc3b814de94796a5e736d4029efb0ee900c12e2d54c993ad1a1e07"
+
+[[package]]
+name = "fs_extra"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "42703706b716c37f96a77aea830392ad231f44c9e9a67872fa5548707e11b11c"
 
 [[package]]
 name = "glob"
@@ -74,6 +193,12 @@ version = "0.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d2fabcfbdc87f4758337ca535fb41a6d701b65693ce38287d856d1674551ec9b"
 
+[[package]]
+name = "heck"
+version = "0.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "95505c38b4572b2d910cecb0281560f54b440a19336cbbcb27bf6ce6adc6f5a8"
+
 [[package]]
 name = "lazy_static"
 version = "1.4.0"
@@ -88,9 +213,9 @@ checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55"
 
 [[package]]
 name = "libc"
-version = "0.2.140"
+version = "0.2.147"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "99227334921fae1a979cf0bfdfcc6b3e5ce376ef57e16fb6fb3ea2ed6095f80c"
+checksum = "b4668fb0ea861c1df094127ac5f1da3409a82116a4ba74fca2e58ef927159bb3"
 
 [[package]]
 name = "libloading"
@@ -104,12 +229,9 @@ dependencies = [
 
 [[package]]
 name = "log"
-version = "0.4.17"
+version = "0.4.19"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "abb12e687cfb44aa40f41fc3978ef76448f9b6038cad6aef4259d3c095a2382e"
-dependencies = [
- "cfg-if",
-]
+checksum = "b06a4cde4c0f271a446782e3eff8de789548ce57dbc8eca9292c27f4a42004b4"
 
 [[package]]
 name = "memchr"
@@ -135,9 +257,9 @@ dependencies = [
 
 [[package]]
 name = "once_cell"
-version = "1.17.1"
+version = "1.18.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b7e5500299e16ebb147ae15a00a942af264cf3688f47923b8fc2cd5858f23ad3"
+checksum = "dd8b5dd2ae5ed71462c540258bedcb51965123ad7e7ccf4b9a8cafaa4a63576d"
 
 [[package]]
 name = "peeking_take_while"
@@ -146,43 +268,61 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "19b17cddbe7ec3f8bc800887bab5e717348c95ea2ca0b1bf0837fb964dc67099"
 
 [[package]]
-name = "pin-project-lite"
-version = "0.2.9"
+name = "prettyplease"
+version = "0.2.15"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e0a7ae3ac2f1173085d398531c705756c94a4c56843785df85a60c1a0afac116"
+checksum = "ae005bd773ab59b4725093fd7df83fd7892f7d8eafb48dbd7de6e024e4215f9d"
+dependencies = [
+ "proc-macro2",
+ "syn",
+]
 
 [[package]]
 name = "proc-macro2"
-version = "1.0.54"
+version = "1.0.66"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e472a104799c74b514a57226160104aa483546de37e839ec50e3c2e41dd87534"
+checksum = "18fb31db3f9bddb2ea821cde30a9f70117e3f119938b5ee630b7403aa6e2ead9"
 dependencies = [
  "unicode-ident",
 ]
 
 [[package]]
 name = "quote"
-version = "1.0.26"
+version = "1.0.32"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4424af4bf778aae2051a77b60283332f386554255d722233d09fbfc7e30da2fc"
+checksum = "50f3b39ccfb720540debaa0164757101c08ecb8d326b15358ce76a62c7e85965"
 dependencies = [
  "proc-macro2",
 ]
 
 [[package]]
 name = "regex"
-version = "1.7.3"
+version = "1.9.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8b1f693b24f6ac912f4893ef08244d70b6067480d2f1a46e950c9691e6749d1d"
+checksum = "b2eae68fc220f7cf2532e4494aded17545fce192d59cd996e0fe7887f4ceb575"
 dependencies = [
+ "aho-corasick",
+ "memchr",
+ "regex-automata",
+ "regex-syntax",
+]
+
+[[package]]
+name = "regex-automata"
+version = "0.3.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "39354c10dd07468c2e73926b23bb9c2caca74c5501e38a35da70406f1d923310"
+dependencies = [
+ "aho-corasick",
+ "memchr",
  "regex-syntax",
 ]
 
 [[package]]
 name = "regex-syntax"
-version = "0.6.29"
+version = "0.7.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f162c6dd7b008981e4d40210aca20b4bd0f9b60ca9271061b07f78537722f2e1"
+checksum = "e5ea92a5b6195c6ef2a0295ea818b312502c6fc94dde986c5553242e18fd4ce2"
 
 [[package]]
 name = "riff"
@@ -191,10 +331,10 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b9b1a3d5f46d53f4a3478e2be4a5a5ce5108ea58b100dcd139830eae7f79a3a1"
 
 [[package]]
-name = "rust-whisper"
+name = "rust-whisper-lib"
 version = "0.1.0"
 dependencies = [
- "tokio",
+ "clap",
  "wav",
  "whisper-rs",
 ]
@@ -207,15 +347,21 @@ checksum = "08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2"
 
 [[package]]
 name = "shlex"
-version = "1.1.0"
+version = "1.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "43b2853a4d09f215c24cc5489c992ce46052d359b5109343cbafbf26bc62f8a3"
+checksum = "a7cee0529a6d40f580e7a5e6c495c8fbfe21b7b52795ed4bb5e62cdf92bc6380"
+
+[[package]]
+name = "strsim"
+version = "0.10.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "73473c0e59e6d5812c5dfe2a064a6444949f089e20eec9a2e5506596494e4623"
 
 [[package]]
 name = "syn"
-version = "1.0.109"
+version = "2.0.32"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "72b64191b275b66ffe2469e8af2c1cfe3bafa67b529ead792a6d0160888b4237"
+checksum = "239814284fd6f1a4ffe4ca893952cdd93c224b6a1571c9a9eadd670295c0c9e2"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -223,21 +369,16 @@ dependencies = [
 ]
 
 [[package]]
-name = "tokio"
-version = "1.27.0"
+name = "unicode-ident"
+version = "1.0.11"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d0de47a4eecbe11f498978a9b29d792f0d2692d1dd003650c24c76510e3bc001"
-dependencies = [
- "autocfg",
- "pin-project-lite",
- "windows-sys",
-]
+checksum = "301abaae475aa91687eb82514b328ab47a211a533026cb25fc3e519b86adfc3c"
 
 [[package]]
-name = "unicode-ident"
-version = "1.0.8"
+name = "utf8parse"
+version = "0.2.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e5464a87b239f13a63a501f2701565754bae92d243d4bb7eb12f6d57d2269bf4"
+checksum = "711b9620af191e0cdc7468a8d14e709c3dcdb115b36f838e601583af800a370a"
 
 [[package]]
 name = "wav"
@@ -261,20 +402,19 @@ dependencies = [
 
 [[package]]
 name = "whisper-rs"
-version = "0.5.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fa7e1b9b003aa3285a0e4469219566266aa1d51ced1be38587251a4f713a1677"
+version = "0.8.0"
 dependencies = [
  "whisper-rs-sys",
 ]
 
 [[package]]
 name = "whisper-rs-sys"
-version = "0.3.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "97a389dc665c7354ba6b1982850d4ba05b862907e535708ebdec92cbd9c599e8"
+version = "0.7.3"
 dependencies = [
  "bindgen",
+ "cfg-if",
+ "cmake",
+ "fs_extra",
 ]
 
 [[package]]
@@ -301,18 +441,18 @@ checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f"
 
 [[package]]
 name = "windows-sys"
-version = "0.45.0"
+version = "0.52.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "75283be5efb2831d37ea142365f009c02ec203cd29a3ebecbc093d52315b66d0"
+checksum = "282be5f36a8ce781fad8c8ae18fa3f9beff57ec1b52cb3de0789201425d9a33d"
 dependencies = [
  "windows-targets",
 ]
 
 [[package]]
 name = "windows-targets"
-version = "0.42.2"
+version = "0.52.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8e5180c00cd44c9b1c88adb3693291f1cd93605ded80c250a75d472756b4d071"
+checksum = "8a18201040b24831fbb9e4eb208f8892e1f50a37feb53cc7ff887feb8f50e7cd"
 dependencies = [
  "windows_aarch64_gnullvm",
  "windows_aarch64_msvc",
@@ -325,42 +465,42 @@ dependencies = [
 
 [[package]]
 name = "windows_aarch64_gnullvm"
-version = "0.42.2"
+version = "0.52.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "597a5118570b68bc08d8d59125332c54f1ba9d9adeedeef5b99b02ba2b0698f8"
+checksum = "cb7764e35d4db8a7921e09562a0304bf2f93e0a51bfccee0bd0bb0b666b015ea"
 
 [[package]]
 name = "windows_aarch64_msvc"
-version = "0.42.2"
+version = "0.52.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e08e8864a60f06ef0d0ff4ba04124db8b0fb3be5776a5cd47641e942e58c4d43"
+checksum = "bbaa0368d4f1d2aaefc55b6fcfee13f41544ddf36801e793edbbfd7d7df075ef"
 
 [[package]]
 name = "windows_i686_gnu"
-version = "0.42.2"
+version = "0.52.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c61d927d8da41da96a81f029489353e68739737d3beca43145c8afec9a31a84f"
+checksum = "a28637cb1fa3560a16915793afb20081aba2c92ee8af57b4d5f28e4b3e7df313"
 
 [[package]]
 name = "windows_i686_msvc"
-version = "0.42.2"
+version = "0.52.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "44d840b6ec649f480a41c8d80f9c65108b92d89345dd94027bfe06ac444d1060"
+checksum = "ffe5e8e31046ce6230cc7215707b816e339ff4d4d67c65dffa206fd0f7aa7b9a"
 
 [[package]]
 name = "windows_x86_64_gnu"
-version = "0.42.2"
+version = "0.52.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8de912b8b8feb55c064867cf047dda097f92d51efad5b491dfb98f6bbb70cb36"
+checksum = "3d6fa32db2bc4a2f5abeacf2b69f7992cd09dca97498da74a151a3132c26befd"
 
 [[package]]
 name = "windows_x86_64_gnullvm"
-version = "0.42.2"
+version = "0.52.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "26d41b46a36d453748aedef1486d5c7a85db22e56aff34643984ea85514e94a3"
+checksum = "1a657e1e9d3f514745a572a6846d3c7aa7dbe1658c056ed9c3344c4109a6949e"
 
 [[package]]
 name = "windows_x86_64_msvc"
-version = "0.42.2"
+version = "0.52.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9aec5da331524158c6d1a4ac0ab1541149c0b9505fde06423b02f5ef0106b9f0"
+checksum = "dff9641d1cd4be8d1a070daf9e3773c5f67e78b4d9d42263020c057706765c04"

rust-whisper-lib/Cargo.toml

@@ -0,0 +1,11 @@
+[package]
+name = "rust-whisper-lib"
+version = "0.1.0"
+edition = "2021"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+whisper-rs = { path = "../gitea-whisper-rs", version = "0.8.0" }
+clap = { version = "4.4.10", features = ["derive"] }
+wav = "1"

rust-whisper-lib/src/lib.rs

@@ -0,0 +1,415 @@
+use whisper_rs::{WhisperContext, FullParams, SamplingStrategy, WhisperError};
+use clap::Parser;
+use std::thread;
+use std::fs::File;
+use std::io::Write;
+
+#[derive(Parser, Debug, Clone)]
+pub struct Flags {
+   #[arg(long, default_value = "../models/ggml-tiny.en.bin")]
+   pub model_path: Option<String>,
+   #[arg(long, default_value = None)]
+   pub model_buffer: Option<Vec<u8>>,
+
+   #[arg(long, default_value = "8")]
+   pub threads: i32,
+
+   #[arg(long, default_value = "5")]
+   pub stream_step: u64,
+   #[arg(long, default_value = "0.6")]
+   pub stream_retain: f32,
+   #[arg(long, default_value = "0.3")]
+   pub stream_head: f32,
+   #[arg(long, default_value = "0.3")]
+   pub stream_tail: f32,
+
+   #[arg(long, default_value = "false")]
+   pub debug: bool,
+
+   #[arg(long, default_value = None)]
+   pub wav: Option<String>,
+   #[arg(long, default_value = None)]
+   pub stream_device: Option<String>,
+}
+
+pub fn wav<F>(flags: Flags, handler_fn: F, wav_path: String) where F: FnMut(Result<Transcribed, String>) + Send + 'static {
+   let w = new_service(
+      flags.model_path,
+      flags.model_buffer,
+      flags.threads,
+      flags.stream_head,
+      flags.stream_tail,
+      handler_fn,
+   ).unwrap();
+   w.transcribe(&f32_from_wav_file(&wav_path).unwrap())
+}
+
+pub fn f32_from_wav_file(path: &String) -> Result<Vec<f32>, String> {
+   let f = std::fs::File::open(path);
+   if let Some(err) = f.as_ref().err() {
+      return Err(format!("failed to open wav file: {}", err));
+   }
+   let wav_read = wav::read(&mut f.unwrap());
+   if let Some(err) = wav_read.as_ref().err() {
+      return Err(format!("failed to parse wav file: {}", err));
+   }
+   let (header, data) = wav_read.unwrap();
+   if header.channel_count != 1 {
+      return Err("!= 1 channel".to_string());
+   }
+   if header.sampling_rate != 16_000 {
+      return Err("!= 16_000 hz".to_string());
+   }
+   match data.as_sixteen() {
+      Some(data16) => {
+         let mut floats = Vec::with_capacity(data16.len());
+         for sample in data16 {
+             floats.push(*sample as f32 / 32768.0);
+         }
+         Ok(floats)
+      },
+      None => Err(format!("couldnt translate wav to 16s")),
+   }
+}
+
+pub fn wav_channel<F>(flags: Flags, handler_fn: F) where F: FnMut(Result<Transcribed, String>) + Send + 'static {
+   let path = flags.wav.as_ref().unwrap();
+   let mut audio = f32_from_wav_file(&path).unwrap();
+   let mut iter = vec![];
+   let n = audio.len() / match audio.len() % 100 {
+      0 => 100,
+      _ => 99,
+   };
+   for _ in 0..100 {
+      iter.push(audio.drain(0..n.clamp(0, audio.len())).collect());
+   }
+   let (fin_send, fin_recv) = std::sync::mpsc::sync_channel::<Option<i32>>(1);
+   channel_and_close(flags.clone(), handler_fn, iter, move || { fin_send.send(None).unwrap(); });
+   match fin_recv.recv() {
+      Ok(_) => {},
+      Err(x) => panic!("failed to receive: {}", x),
+   };
+}
+
+pub fn channel<F, I>(flags: Flags, handler_fn: F, stream: I) where F: FnMut(Result<Transcribed, String>) + Send + 'static, I: IntoIterator<Item = Vec<f32>> {
+   channel_and_close(flags, handler_fn, stream, || {});
+}
+
+fn channel_and_close<F, I, G>(flags: Flags, handler_fn: F, stream: I, mut close_fn: G) where F: FnMut(Result<Transcribed, String>) + Send + 'static, I: IntoIterator<Item = Vec<f32>>, G: FnMut() + Send + 'static {
+   let w = new_service(
+      flags.model_path,
+      flags.model_buffer,
+      flags.threads,
+      flags.stream_head,
+      flags.stream_tail,
+      handler_fn,
+   ).unwrap();
+   let stream_retain = (flags.stream_retain * 16_000.0) as usize;
+   match &flags.debug {
+      true => { File::create("/tmp/page.rawf32audio").unwrap(); },
+      false => {},
+   };
+   let mut buffer = vec![];
+   for data in stream {
+      data.iter().for_each(|x| buffer.push(*x));
+      if buffer.len() >= (flags.stream_step * 16_000) as usize {
+         w.transcribe_async(&buffer).unwrap();
+
+         match &flags.debug {
+            true => {
+               let mut f = File::options().append(true).open("/tmp/page.rawf32audio").unwrap();
+               let mut wav_data = vec![];
+               for i in buffer.iter() {
+                  for j in i.to_le_bytes() {
+                     wav_data.push(j);
+                  }
+               }
+               f.write_all(wav_data.as_slice()).unwrap();
+            },
+            false => {},
+         };
+
+         for i in 0..stream_retain {
+            buffer[i] = buffer[buffer.len() - stream_retain + i];
+         }
+         buffer.truncate(stream_retain);
+      }
+   }
+   if buffer.len() > 0 {
+      w.transcribe(&buffer);
+   }
+   close_fn();
+}
+
+struct Service {
+   jobs: std::sync::mpsc::SyncSender<ATranscribe>,
+}
+
+fn new_service<F>(model_path: Option<String>, model_buffer: Option<Vec<u8>>, threads: i32, stream_head: f32, stream_tail: f32, handler_fn: F) -> Result<Service, String> where F: FnMut(Result<Transcribed, String>) + Send + 'static {
+   match new_engine(model_path, model_buffer, threads) {
+      Ok(engine) => {
+         let mut whisper = new_whisper_impl(engine, stream_head, stream_tail, handler_fn);
+         let (send, recv) = std::sync::mpsc::sync_channel(100);
+         thread::spawn(move || { whisper.transcribe_asyncs(recv); });
+         Ok(Service{jobs: send})
+      },
+      Err(msg) => Err(format!("failed to initialize engine: {}", msg)),
+   }
+}
+
+impl Service {
+   fn transcribe(&self, data: &Vec<f32>) {
+      let (send, recv) = std::sync::mpsc::sync_channel(0);
+      self._transcribe_async(data, Some(send)).unwrap();
+      recv.recv().unwrap();
+   }
+
+   fn transcribe_async(&self, data: &Vec<f32>) -> Result<(), String> {
+      self._transcribe_async(data, None)
+   }
+
+   fn _transcribe_async(&self, data: &Vec<f32>, ack: Option<std::sync::mpsc::SyncSender<bool>>) -> Result<(), String> {
+      match self.jobs.try_send(ATranscribe{
+         data: data.clone().to_vec(),
+         ack: ack,
+      }) {
+         Ok(_) => Ok(()),
+         Err(msg) => Err(format!("failed to enqueue transcription: {}", msg)),
+      }
+   }
+}
+
+struct Impl {
+   engine: Engine,
+   stream_head: f32,
+   stream_tail: f32,
+   handler_fn: Option<Box<dyn FnMut(Result<Transcribed, String>) + Send + 'static>>
+}
+
+fn new_whisper_impl<F>(engine: Engine, stream_head: f32, stream_tail: f32, handler_fn: F) -> Impl where F: FnMut(Result<Transcribed, String>) + Send + 'static {
+   Impl {
+      engine: engine,
+      stream_head: stream_head,
+      stream_tail: stream_tail,
+      handler_fn: Some(Box::new(handler_fn)),
+   }
+}
+
+impl Impl {
+   fn transcribe_asyncs(&mut self, recv: std::sync::mpsc::Receiver<ATranscribe>) {
+      loop {
+         match recv.recv() {
+            Ok(job) => {
+               let result = self.transcribe(&job).is_ok();
+               match job.ack {
+                  Some(ack) => {
+                     ack.send(result).unwrap();
+                  },
+                  None => (),
+               };
+            }
+            Err(_) => return,
+         };
+      }
+   }
+
+   fn transcribe(&mut self, a_whisper: &ATranscribe) -> Result<(), ()> {
+      match self.engine.transcribe(&a_whisper.data) {
+         Ok(result) => {
+            self.on_success(&result);
+            Ok(())
+         },
+         Err(msg) => {
+            self.on_error(msg.to_string());
+            Err(())
+         },
+      }
+   }
+
+   fn on_success(&mut self, whispered: &Transcribed) {
+      let result = whispered
+         .after(&(self.stream_head * 100.0))
+         .before(&(self.stream_tail * 100.0));
+      if result.to_string().trim().len() > 0 {
+         (self.handler_fn.as_mut().unwrap())(Ok(result));
+      }
+   }
+
+   fn on_error(&mut self, msg: String) {
+      (self.handler_fn.as_mut().unwrap())(Err(format!("failed to transcribe: {}", &msg)));
+   }
+}
+
+struct Engine {
+   ctx: WhisperContext,
+   threads: i32,
+}
+
+fn new_engine(model_path: Option<String>, model_buffer: Option<Vec<u8>>, threads: i32) -> Result<Engine, String> {
+   let whisper_context_result = match model_path {
+      Some(model_path) => WhisperContext::new(&model_path),
+      None => WhisperContext::new_from_buffer(&model_buffer.unwrap()),
+   };
+   match whisper_context_result {
+      Ok(ctx) => Ok(Engine{ctx: ctx, threads: threads}),
+      Err(msg) => Err(format!("failed to load model: {}", msg)),
+   }
+}
+
+impl Engine {
+   fn transcribe(&self, data: &Vec<f32>) -> Result<Transcribed, String> {
+      match self._transcribe(data) {
+         Ok(transcribed) => Ok(transcribed),
+         Err(msg) => Err(format!("{}", msg)),
+      }
+   }
+
+   fn _transcribe(&self, data: &Vec<f32>) -> Result<Transcribed, WhisperError> {
+      let mut params = FullParams::new(SamplingStrategy::Greedy { best_of: 0 });
+      params.set_no_context(true);
+      params.set_n_threads(self.threads);
+      params.set_translate(false);
+      params.set_detect_language(false);
+      params.set_language(Some("en"));
+      params.set_print_special(false);
+      params.set_print_progress(false);
+      params.set_print_realtime(false);
+      params.set_print_timestamps(false);
+
+      let mut state = self.ctx.create_state()?;
+      state.full(params, &data[..])?;
+
+      let mut result = new_whispered();
+      let num_segments = state.full_n_segments()?;
+      for i in 0..num_segments {
+         let data = state.full_get_segment_text(i)?;
+         let start = state.full_get_segment_t0(i)?;
+         let stop = state.full_get_segment_t1(i)?;
+         result.push(data, start, stop);
+      }
+
+      Ok(result)
+   }
+}
+
+struct ATranscribe {
+   data: Vec<f32>,
+   ack: Option<std::sync::mpsc::SyncSender<bool>>,
+}
+
+#[derive(Clone, Debug)]
+pub struct Transcribed {
+   data: Vec<ATranscribed>,
+}
+
+#[derive(Clone, Debug)]
+struct ATranscribed {
+   pub data: String,
+   pub offset: i64,
+   pub length: i64,
+}
+
+fn new_whispered() -> Transcribed {
+   Transcribed{data: vec![]}
+}
+
+fn new_a_whispered(data: String, start: i64, stop: i64) -> ATranscribed {
+   ATranscribed{
+      data: data,
+      offset: start.clone(),
+      length: stop - start,
+   }
+}
+
+impl Transcribed {
+   pub fn to_string(&self) -> String {
+      let mut result = "".to_string();
+      for i in 0..self.data.len() {
+         result = format!("{} {}", result, &self.data[i].data);
+      }
+      result
+   }
+
+   fn after(&self, t: &f32) -> Transcribed {
+      let mut result = new_whispered();
+      self.data
+         .iter()
+         .filter(|x| x.offset as f32 >= *t)
+         .for_each(|x| result.data.push(x.clone()));
+      result
+   }
+
+   fn before(&self, t: &f32) -> Transcribed {
+      let mut result = new_whispered();
+      let end = match self.data.iter().map(|x| x.offset + x.length).max() {
+         Some(x) => x,
+         None => 1,
+      };
+      let t = (end as f32) - *t;
+      self.data
+         .iter()
+         .filter(|x| ((x.offset) as f32) <= t)
+         .for_each(|x| result.data.push(x.clone()));
+      result
+   }
+
+   fn push(&mut self, data: String, start: i64, stop: i64) {
+      let words: Vec<_> = data.split_whitespace().collect();
+      let per_word = (stop - start) / (words.len() as i64);
+      for i in 0..words.len() {
+         let start = (i as i64) * per_word;
+         let stop = start.clone() + per_word;
+         self.data.push(new_a_whispered(words[i].to_string(), start, stop));
+      }
+   }
+}
+
+#[cfg(test)]
+mod tests {
+   use super::*;
+
+   #[test]
+   fn test_transcribe_tiny_jfk_wav_whisper_rs_wav_channel() {
+      wav_channel(
+         Flags {
+            model_path: None,
+            model_buffer: Some(include_bytes!("../../models/ggml-tiny.en.bin").to_vec()),
+            threads: 8,
+            stream_step: 30,
+            stream_retain: 0.0,
+            stream_head: 0.0,
+            stream_tail: 0.0,
+            wav: Some("../gitea-whisper-rs/sys/whisper.cpp/bindings/go/samples/jfk.wav".to_string()),
+            debug: false,
+            stream_device: None,
+         },
+         move | result | {
+            assert!(result.is_ok());
+            assert_eq!(result.unwrap().to_string(), " And so my fellow Americans ask not what your country can do for you ask what you can do for your country.");
+         },
+      );
+   }
+
+   #[test]
+   fn test_transcribe_tiny_jfk_wav_whisper_rs() {
+      wav(
+         Flags {
+            model_path: None,
+            model_buffer: Some(include_bytes!("../../models/ggml-tiny.en.bin").to_vec()),
+            threads: 8,
+            stream_step: 0,
+            stream_retain: 0.0,
+            stream_head: 0.0,
+            stream_tail: 0.0,
+            wav: Some("../gitea-whisper-rs/sys/whisper.cpp/bindings/go/samples/jfk.wav".to_string()),
+            debug: false,
+            stream_device: None,
+         },
+         | result | {
+            assert!(result.is_ok());
+            assert_eq!(result.unwrap().to_string(), " And so my fellow Americans ask not what your country can do for you ask what you can do for your country.");
+         },
+         "../gitea-whisper-rs/sys/whisper.cpp/bindings/go/samples/jfk.wav".to_string(),
+      );
+   }
+}

rust-whisper-lib/target/doc/rust_whisper_lib/all.html

@@ -0,0 +1 @@
+<!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="List of all items in this crate"><title>List of all items in this crate</title><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/SourceSerif4-Regular-46f98efaafac5295.ttf.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/FiraSans-Regular-018c141bf0843ffd.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/FiraSans-Medium-8f9a781e4970d388.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/SourceCodePro-Regular-562dcc5011b6de7d.ttf.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/SourceSerif4-Bold-a2c9cd1067f8b328.ttf.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/SourceCodePro-Semibold-d899c5a5c4aeb14a.ttf.woff2"><link rel="stylesheet" href="../static.files/normalize-76eba96aa4d2e634.css"><link rel="stylesheet" href="../static.files/rustdoc-fa3bb1812debf86c.css"><meta name="rustdoc-vars" data-root-path="../" data-static-root-path="../static.files/" data-current-crate="rust_whisper_lib" data-themes="" data-resource-suffix="" data-rustdoc-version="1.74.0 (79e9716c9 2023-11-13)" data-channel="1.74.0" data-search-js="search-8be46b629f5f14a8.js" data-settings-js="settings-74424d7eec62a23e.js" ><script src="../static.files/storage-fec3eaa3851e447d.js"></script><script defer src="../static.files/main-c5bd66d33317d69f.js"></script><noscript><link rel="stylesheet" href="../static.files/noscript-5d8b3c7633ad77ba.css"></noscript><link rel="alternate icon" type="image/png" href="../static.files/favicon-16x16-8b506e7a72182f1c.png"><link rel="alternate icon" type="image/png" href="../static.files/favicon-32x32-422f7d1d52889060.png"><link rel="icon" type="image/svg+xml" href="../static.files/favicon-2c020d218678b618.svg"></head><body class="rustdoc mod"><!--[if lte IE 11]><div class="warning">This old browser is unsupported and will most likely display funky things.</div><![endif]--><nav class="mobile-topbar"><button class="sidebar-menu-toggle">&#9776;</button><a class="logo-container" href="../rust_whisper_lib/index.html"><img class="rust-logo" src="../static.files/rust-logo-151179464ae7ed46.svg" alt="logo"></a></nav><nav class="sidebar"><a class="logo-container" href="../rust_whisper_lib/index.html"><img class="rust-logo" src="../static.files/rust-logo-151179464ae7ed46.svg" alt="logo"></a><h2 class="location"><a href="#">Crate rust_whisper_lib</a></h2><div class="sidebar-elems"><section><ul class="block"><li><a href="#structs">Structs</a></li><li><a href="#functions">Functions</a></li></ul></section></div></nav><main><div class="width-limiter"><nav class="sub"><form class="search-form"><span></span><input class="search-input" name="search" aria-label="Run search in the documentation" autocomplete="off" spellcheck="false" placeholder="Click or press ‘S’ to search, ‘?’ for more options…" type="search"><div id="help-button" title="help" tabindex="-1"><a href="../help.html">?</a></div><div id="settings-menu" tabindex="-1"><a href="../settings.html" title="settings"><img width="22" height="22" alt="Change settings" src="../static.files/wheel-7b819b6101059cd0.svg"></a></div></form></nav><section id="main-content" class="content"><h1>List of all items</h1><h3 id="structs">Structs</h3><ul class="all-items"><li><a href="struct.Flags.html">Flags</a></li><li><a href="struct.Transcribed.html">Transcribed</a></li></ul><h3 id="functions">Functions</h3><ul class="all-items"><li><a href="fn.channel.html">channel</a></li><li><a href="fn.wav.html">wav</a></li></ul></section></div></main></body></html>

rust-whisper-lib/target/doc/rust_whisper_lib/index.html

@@ -0,0 +1 @@
+<!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="API documentation for the Rust `rust_whisper_lib` crate."><title>rust_whisper_lib - Rust</title><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/SourceSerif4-Regular-46f98efaafac5295.ttf.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/FiraSans-Regular-018c141bf0843ffd.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/FiraSans-Medium-8f9a781e4970d388.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/SourceCodePro-Regular-562dcc5011b6de7d.ttf.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/SourceSerif4-Bold-a2c9cd1067f8b328.ttf.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../static.files/SourceCodePro-Semibold-d899c5a5c4aeb14a.ttf.woff2"><link rel="stylesheet" href="../static.files/normalize-76eba96aa4d2e634.css"><link rel="stylesheet" href="../static.files/rustdoc-fa3bb1812debf86c.css"><meta name="rustdoc-vars" data-root-path="../" data-static-root-path="../static.files/" data-current-crate="rust_whisper_lib" data-themes="" data-resource-suffix="" data-rustdoc-version="1.74.0 (79e9716c9 2023-11-13)" data-channel="1.74.0" data-search-js="search-8be46b629f5f14a8.js" data-settings-js="settings-74424d7eec62a23e.js" ><script src="../static.files/storage-fec3eaa3851e447d.js"></script><script defer src="../crates.js"></script><script defer src="../static.files/main-c5bd66d33317d69f.js"></script><noscript><link rel="stylesheet" href="../static.files/noscript-5d8b3c7633ad77ba.css"></noscript><link rel="alternate icon" type="image/png" href="../static.files/favicon-16x16-8b506e7a72182f1c.png"><link rel="alternate icon" type="image/png" href="../static.files/favicon-32x32-422f7d1d52889060.png"><link rel="icon" type="image/svg+xml" href="../static.files/favicon-2c020d218678b618.svg"></head><body class="rustdoc mod crate"><!--[if lte IE 11]><div class="warning">This old browser is unsupported and will most likely display funky things.</div><![endif]--><nav class="mobile-topbar"><button class="sidebar-menu-toggle">&#9776;</button><a class="logo-container" href="../rust_whisper_lib/index.html"><img class="rust-logo" src="../static.files/rust-logo-151179464ae7ed46.svg" alt="logo"></a></nav><nav class="sidebar"><a class="logo-container" href="../rust_whisper_lib/index.html"><img class="rust-logo" src="../static.files/rust-logo-151179464ae7ed46.svg" alt="logo"></a><h2 class="location"><a href="#">Crate rust_whisper_lib</a></h2><div class="sidebar-elems"><ul class="block"><li class="version">Version 0.1.0</li><li><a id="all-types" href="all.html">All Items</a></li></ul><section><ul class="block"><li><a href="#structs">Structs</a></li><li><a href="#functions">Functions</a></li></ul></section></div></nav><main><div class="width-limiter"><nav class="sub"><form class="search-form"><span></span><input class="search-input" name="search" aria-label="Run search in the documentation" autocomplete="off" spellcheck="false" placeholder="Click or press ‘S’ to search, ‘?’ for more options…" type="search"><div id="help-button" title="help" tabindex="-1"><a href="../help.html">?</a></div><div id="settings-menu" tabindex="-1"><a href="../settings.html" title="settings"><img width="22" height="22" alt="Change settings" src="../static.files/wheel-7b819b6101059cd0.svg"></a></div></form></nav><section id="main-content" class="content"><div class="main-heading"><h1>Crate <a class="mod" href="#">rust_whisper_lib</a><button id="copy-path" title="Copy item path to clipboard"><img src="../static.files/clipboard-7571035ce49a181d.svg" width="19" height="18" alt="Copy item path"></button></h1><span class="out-of-band"><a class="src" href="../src/rust_whisper_lib/lib.rs.html#1-352">source</a> · <button id="toggle-all-docs" title="collapse all docs">[<span>&#x2212;</span>]</button></span></div><h2 id="structs" class="small-section-header"><a href="#structs">Structs</a></h2><ul class="item-table"><li><div class="item-name"><a class="struct" href="struct.Flags.html" title="struct rust_whisper_lib::Flags">Flags</a></div></li><li><div class="item-name"><a class="struct" href="struct.Transcribed.html" title="struct rust_whisper_lib::Transcribed">Transcribed</a></div></li></ul><h2 id="functions" class="small-section-header"><a href="#functions">Functions</a></h2><ul class="item-table"><li><div class="item-name"><a class="fn" href="fn.channel.html" title="fn rust_whisper_lib::channel">channel</a></div></li><li><div class="item-name"><a class="fn" href="fn.wav.html" title="fn rust_whisper_lib::wav">wav</a></div></li></ul></section></div></main></body></html>

rust-whisper-lib/target/doc/rust_whisper_lib/sidebar-items.js

@@ -0,0 +1 @@
+window.SIDEBAR_ITEMS = {"fn":["channel","wav"],"struct":["Flags","Transcribed"]};

snowboy-2022/build.sh

@@ -1,99 +0,0 @@
-#! /bin/bash
-
-echo https://github.com/seasalt-ai/snowboy
-cd "$(dirname "$(realpath "$BASH_SOURCE")")"
-
-set -e
-set -o pipefail
-
-if [ ! -d ./snowboy.git.d ]; then
-   git clone https://github.com/seasalt-ai/snowboy snowboy.git.d
-fi
-
-timeout 2 docker version &> /dev/null
-if ! docker images | grep snowboy-pmdl.*latest &> /dev/null; then
-   pushd snowboy.git.d
-   docker build -t snowboy-pmdl:latest .
-   popd
-fi
-
-export HOTWORD="${HOTWORD:-${TRAIN:-default_hotword}}"
-
-if [ -n "$TRAIN" ] || [ ! -d ./model ] || [ ! -f ./model/$HOTWORD.pmdl ]; then
-   mkdir -p model
-   pushd model
-   rm -f ./record{1,2,3}.wav || true
-   echo "record 3 instances of '$HOTWORD'" >&2
-   for i in 1 2 3; do
-      read -p "[$i/3] ready? you get 3 seconds."
-      (
-         timeout 3 rec \
-            -r 16000 \
-            -c 1 \
-            -b 16 \
-            -e signed-integer \
-            -t wav \
-            record$i.wav
-      ) || true
-      ls record$i.wav
-   done
-   popd
-
-   docker run \
-      --rm \
-      -it \
-      -v "$(realpath ./model)":/snowboy-master/examples/Python/model \
-      snowboy-pmdl:latest
-   mv ./model/hotword.pmdl ./model/$HOTWORD.pmdl
-   if [ -n "$TRAIN" ]; then
-      exit 0
-   fi
-fi
-
-if false; then
-   if ! which swig; then
-      brew install swig
-   fi
-   pip3 install pyaudio
-   pushd snowboy.git.d/swig/Python3/
-   make
-   popd
-
-   cd snowboy.git.d/examples/Python3/
-   echo '
-   import snowboydecoder
-   import datetime
-   detected_callback = lambda *args: print(datetime.datetime.now(), "GOTCHA")
-   d = snowboydecoder.HotwordDetector("../../../model/'"$HOTWORD"'.pmdl", sensitivity=0.5, audio_gain=1)
-   d.start(detected_callback)
-   ' > breel.py
-   echo GO
-   cleanup() {
-      echo OK IM DONE NOW
-   }
-   trap cleanup EXIT
-   python3 ./breel.py
-else
-   resources="$(realpath snowboy.git.d/resources/common.res)"
-   hotword="$(realpath ./model/$HOTWORD.pmdl)"
-   GOPROXY= go build -o snowboy
-   if [ -z "$PUSH" ]; then
-      ./snowboy \
-         -ms "$hotword/$HOTWORD" \
-         -r "$resources" \
-         -s 0.5 \
-         "$@"
-   else
-      echo '
-FROM registry-app.eng.qops.net:5001/imported/alpine:3.16
-WORKDIR /main/
-COPY ./snowboy.git.d/resources/common.res ./
-COPY ./model/hotword.pmdl ./
-COPY ./snowboy ./
-ENTRYPOINT ["sh", "-c", "true; echo copying /main/ to /mnt/; cp /main/* /mnt/"]
-CMD []
-      ' > Dockerfile
-      docker build -t registry-app.eng.qops.net:5001/breel/snowboy:latest .
-      docker push registry-app.eng.qops.net:5001/breel/snowboy:latest
-   fi
-fi

snowboy-2022/go.mod

@@ -1,13 +0,0 @@
-module snowboy
-
-go 1.19
-
-require (
-	github.com/brentnd/go-snowboy v0.0.0-20190301212623-e19133c572af
-	github.com/gordonklaus/portaudio v0.0.0-20221027163845-7c3b689db3cc
-)
-
-require (
-	github.com/Kitt-AI/snowboy v1.3.0 // indirect
-	github.com/stretchr/testify v1.8.1 // indirect
-)

snowboy-2022/go.sum

@@ -1,22 +0,0 @@
-github.com/Kitt-AI/snowboy v1.3.0 h1:PjBVN84M/9tAzDBQXILAKMoJMxt/fT0nhJ1rhKtVRUc=
-github.com/Kitt-AI/snowboy v1.3.0/go.mod h1:sDzzMXFQ1wFkXkZaX/ant0xJsizGVq/9hyKb7ZB3cNI=
-github.com/brentnd/go-snowboy v0.0.0-20190301212623-e19133c572af h1:ijY5OHNQs3CdzTN2XT+zByIsR1QVyXTvOUSkQcBm6pw=
-github.com/brentnd/go-snowboy v0.0.0-20190301212623-e19133c572af/go.mod h1:XcT4k8Tn9hrM5SLVvu5hNQbAC6GojXM0MXz1Rt8CL68=
-github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
-github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
-github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
-github.com/gordonklaus/portaudio v0.0.0-20221027163845-7c3b689db3cc h1:yYLpN7bJxKYILKnk20oczGQOQd2h3/7z7/cxdD9Se/I=
-github.com/gordonklaus/portaudio v0.0.0-20221027163845-7c3b689db3cc/go.mod h1:WY8R6YKlI2ZI3UyzFk7P6yGSuS+hFwNtEzrexRyD7Es=
-github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
-github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
-github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
-github.com/stretchr/objx v0.4.0/go.mod h1:YvHI0jy2hoMjB+UWwv71VJQ9isScKT/TqJzVSSt89Yw=
-github.com/stretchr/objx v0.5.0/go.mod h1:Yh+to48EsGEfYuaHDzXPcE3xhTkx73EhmCGUpEOglKo=
-github.com/stretchr/testify v1.7.1/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
-github.com/stretchr/testify v1.8.0/go.mod h1:yNjHg4UonilssWZ8iaSj1OCr/vHnekPRkoO+kdMU+MU=
-github.com/stretchr/testify v1.8.1 h1:w7B6lhMri9wdJUVmEZPGGhZzrYTPvgJArz7wNPgYKsk=
-github.com/stretchr/testify v1.8.1/go.mod h1:w2LPCIKwWwSfY2zedu0+kehJoqGctiVI29o6fzry7u4=
-gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
-gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
-gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
-gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

snowboy-2022/main.go

@@ -1,122 +0,0 @@
-// This example streams the microphone thru Snowboy to listen for the hotword,
-// by using the PortAudio interface.
-//
-// HOW TO USE:
-// 	go run examples/Go/listen/main.go [path to snowboy resource file] [path to snowboy hotword file]
-//
-package main
-
-import (
-	"bytes"
-	"encoding/binary"
-	"flag"
-	"fmt"
-	"log"
-	"path"
-	"strings"
-	"time"
-
-	"github.com/brentnd/go-snowboy"
-	"github.com/gordonklaus/portaudio"
-)
-
-// Sound represents a sound stream implementing the io.Reader interface
-// that provides the microphone data.
-type Sound struct {
-	stream *portaudio.Stream
-	data   []int16
-}
-
-// Init initializes the Sound's PortAudio stream.
-func (s *Sound) Init() {
-	inputChannels := 1
-	outputChannels := 0
-	sampleRate := 16000
-	s.data = make([]int16, 1024)
-
-	// initialize the audio recording interface
-	err := portaudio.Initialize()
-	if err != nil {
-		fmt.Errorf("Error initialize audio interface: %s", err)
-		return
-	}
-
-	// open the sound input stream for the microphone
-	stream, err := portaudio.OpenDefaultStream(inputChannels, outputChannels, float64(sampleRate), len(s.data), s.data)
-	if err != nil {
-		fmt.Errorf("Error open default audio stream: %s", err)
-		return
-	}
-
-	err = stream.Start()
-	if err != nil {
-		fmt.Errorf("Error on stream start: %s", err)
-		return
-	}
-
-	s.stream = stream
-}
-
-// Close closes down the Sound's PortAudio connection.
-func (s *Sound) Close() {
-	s.stream.Close()
-	portaudio.Terminate()
-}
-
-// Read is the Sound's implementation of the io.Reader interface.
-func (s *Sound) Read(p []byte) (int, error) {
-	s.stream.Read()
-
-	buf := &bytes.Buffer{}
-	for _, v := range s.data {
-		binary.Write(buf, binary.LittleEndian, v)
-	}
-
-	copy(p, buf.Bytes())
-	return len(p), nil
-}
-
-func main() {
-	resources := flag.String("r", "", "path to the .res file")
-	models := flag.String("ms", "", "comma delimited path to the .?mdl file/output")
-	sensitivity := flag.Float64("s", 0.45, "0..1")
-	quiet := flag.Bool("q", false, "emit '1' on detect else silent")
-	flag.Parse()
-
-	if *resources == "" || *models == "" {
-		panic("all flags must be set")
-	}
-
-	// open the mic
-	mic := &Sound{}
-	mic.Init()
-	defer mic.Close()
-
-	// open the snowboy detector
-	d := snowboy.NewDetector(*resources)
-	defer d.Close()
-
-	// set the handlers
-	for _, modelStrC := range strings.Split(*models, ",") {
-		modelStr := modelStrC
-		d.HandleFunc(snowboy.NewHotword(path.Dir(modelStr), float32(*sensitivity)), func(string) {
-			if !*quiet {
-				log.Println(path.Base(modelStr))
-			}
-			fmt.Println(path.Base(modelStr))
-		})
-	}
-
-	d.HandleSilenceFunc(1*time.Second, func(string) {
-		if !*quiet {
-			log.Println("...")
-		}
-	})
-
-	// display the detector's expected audio format
-	sr, nc, bd := d.AudioFormat()
-	log.Printf("sample rate=%d, num channels=%d, bit depth=%d\n", sr, nc, bd)
-
-	// start detecting using the microphone
-	d.ReadAndDetect(mic)
-}

snowboy-2022/pyautogui.sh

@@ -1,82 +0,0 @@
-#! /bin/bash
-
-main() {
-   cleanup() {
-      killall -9 $(jobs -p)
-      killall snowboy
-   }
-   trap cleanup EXIT
-
-   if [ ! -e /tmp/stt.fifo ]; then
-      mkfifo /tmp/stt.fifo
-   fi
-
-   echo starting in
-   for ((i=2; i>0; i--)); do
-      echo "...$i..."
-      sleep 1
-   done
-
-   local models=($(
-      cat pyautogui.yaml \
-         | gojq -r -c --yaml-input '
-            to_entries[] | "model/"+.key+".pmdl/"+.key
-         ' \
-      | tr '\n' ',' \
-      | sed 's/,$//'
-   ))
-   echo models=$models
-
-   ./snowboy -r resources.res -ms $models "$@" > /tmp/stt.fifo &
-
-   python3 -c '
-import pyautogui
-import time
-
-keys = set()
-
-def toggle(key):
-   global keys
-   if key in keys:
-      release(key)
-   else:
-      hold(key)
-
-def hold(key):
-   global keys
-   for keyin in [todrop for todrop in keys]:
-      if keyin != key:
-         release(keyin)
-   keys = set()
-   keys.add(key)
-   print()
-   print("pressing", key)
-   print()
-   pyautogui.keyDown(key)
-
-def release(key):
-   print()
-   print("releasing", key)
-   print()
-   pyautogui.keyUp(key)
-
-def main():
-   with open("/tmp/stt.fifo", "r") as q:
-      for line in q:
-         handle(line.strip())
-
-import yaml
-mapping = yaml.safe_load(open("./pyautogui.yaml", "r"))
-print(mapping)
-
-def handle(cmd):
-   global mapping
-   hold(mapping.get(cmd))
-
-main()
-   '
-}
-
-if [ "$0" == "$BASH_SOURCE" ]; then
-   main "$@"
-fi

snowboy-2022/pyautogui.yaml

@@ -1,5 +0,0 @@
-up: w
-down: s
-left: a
-right: d
-jump: w

whisper-2023/hotwords.py

@@ -1,304 +0,0 @@
-import speech_recognition as sr
-import time
-import threading
-import queue
-import signal
-import sys
-import os
-import requests
-import yaml
-
-def log(*args):
-   print(">", *args, file=sys.stderr)
-
-class Piper(threading.Thread):
-   def __init__(self, inq, outq):
-      threading.Thread.__init__(self)
-      self.inq = inq
-      self.outq = outq
-
-   def run(self):
-      while True:
-         got = self.inq.get()
-         if got is None:
-            break
-         self._run(got)
-      self.outq.put(None)
-
-class Manager(threading.Thread):
-   def __init__(self, outq):
-      threading.Thread.__init__(self)
-      self.outq = outq
-      inq = queue.Queue()
-      def catcher(sig, frame):
-         inq.put(None)
-      self.inq = inq
-      signal.signal(signal.SIGINT, catcher)
-
-   def run(self):
-      log("Manager.run: start")
-      self.inq.get()
-      self.outq.put(None)
-      log("Manager.run: stop")
-
-class Reader(threading.Thread):
-   def __init__(self, inq, outq):
-      threading.Thread.__init__(self)
-      self.name = os.environ.get("MIC_NAME", "pulse_monitor")
-      if not self.name:
-         for index, name in enumerate(sr.Microphone.list_microphone_names()):
-             print("[{0}] Microphone with name \"{1}\" found for `Microphone(device_index={0})`".format(index, name))
-         exit()
-      self.inq = inq
-      self.outq = outq
-
-   def run(self):
-      log("Reader.run: start")
-      try:
-         idx = [
-            idx for idx,v in enumerate(
-               sr.Microphone.list_microphone_names(),
-            ) if v in self.name.split(",")
-         ][0]
-         with sr.Microphone(device_index=idx) as mic:
-            while not self.should_stop():
-               try:
-                  self.outq.put(self._run(mic))
-               except Exception as e:
-                  if not "timed out" in str(e):
-                     log("Reader.run: error:", e)
-      except Exception as e:
-         log("Reader.run panic:", e)
-         log("microphones:", sr.Microphone.list_microphone_names())
-      finally:
-         self.outq.put(None)
-      log("Reader.run: stop")
-
-   def should_stop(self):
-      return not self.inq.empty()
-
-   def _run(self, mic):
-      mic_timeout = int(os.environ.get("MIC_TIMEOUT", 5))
-      r = sr.Recognizer()
-      return r.listen(
-         mic,
-         timeout=mic_timeout,
-         phrase_time_limit=mic_timeout,
-      )
-
-class Parser(threading.Thread):
-   def __init__(self, inq, outq):
-      threading.Thread.__init__(self)
-      self.inq = inq
-      self.outq = outq
-
-   def run(self):
-      log("Parser.run: start")
-      while True:
-         try:
-            clip = self.inq.get()
-            backlog = self.inq.qsize()
-            if backlog:
-               log("Parser.run backlog", backlog)
-            if clip is None:
-               break
-            self.outq.put(self._run(clip).strip())
-         except Exception as e:
-            log("Parser.run: error:", e)
-      self.outq.put(None)
-      log("Parser.run: stop")
-
-   def _run(self, clip):
-      r = sr.Recognizer()
-      return r.recognize_whisper(clip, language="english", model=os.environ.get("MODEL", "small.en")) # tiny.en=32x, base.en=16x, small.en=6x, medium.en=x2
-
-def load_dot_notation(v, s):
-   items = s.replace("[]", ".[]").split(".")
-   return _load_dot_notation(v, items)
-
-def _load_dot_notation(v, items):
-   for i in range(len(items)):
-      k = items[i]
-      if not k:
-         continue
-      if k == "[]":
-         if isinstance(v, list):
-            result = []
-            for j in v:
-               subresult = _load_dot_notation(j, items[i+1:])
-               if isinstance(subresult, list):
-                  result.extend(subresult)
-               else:
-                  result.append(subresult)
-            return result
-         else:
-            result = []
-            for j in v.values():
-               subresult = _load_dot_notation(j, items[i+1:])
-               if isinstance(subresult, list):
-                  result.extend(subresult)
-               else:
-                  result.append(subresult)
-            return result
-      else:
-         if isinstance(v, list):
-            v = v[int(k)]
-         else:
-            v = v[k]
-   return v
-
-def test_load_dot_notation():
-   for i in [
-      "a" == load_dot_notation("a", "."),
-      ["a"] == load_dot_notation(["a"], "."),
-      "b" == load_dot_notation({"a":"b"}, ".a"),
-      "c" == load_dot_notation({"a":{"b":"c"}}, ".a.b"),
-      "c" == load_dot_notation({"a":{"b":["c"]}}, ".a.b.0"),
-      ["c","d"] == load_dot_notation({"a":{"b":"c"}, "a2":{"b":"d"}}, ".[].b"),
-      ["c","d"] == load_dot_notation({"a":{"b":["c"], "b2":["d"]}}, ".a.[].0"),
-      ["c","d"] == load_dot_notation({"a":{"b":["c"], "b2":["d"]}}, ".a[].0"),
-      ["c","d"] == load_dot_notation(["c", "d"], "."),
-      ["c","d"] == load_dot_notation(["c", "d"], "[]"),
-   ]:
-      if not i:
-         raise Exception(i)
-test_load_dot_notation()
-
-class Reactor(threading.Thread):
-   def __init__(self, inq, outq):
-      threading.Thread.__init__(self)
-      self.inq = inq
-      self.outq = outq
-      self.load_hotwords = Reactor.new_load_hotwords()
-      log(f"hotwords: {self.load_hotwords()}")
-
-   def new_load_hotwords():
-      p = os.environ.get("HOTWORDS", None)
-      if not p:
-         def load_nothing():
-            return []
-         return load_nothing
-
-      try:
-         if "@" in p:
-            def load_hotwords_in_yaml_file():
-               with open(p.split("@")[0], "r") as f:
-                  v = yaml.safe_load(f)
-                  v = load_dot_notation(v, p.split("@")[-1])
-                  return ["".join(i.strip().lower().split()) for i in v if i]
-            load_hotwords_in_yaml_file()
-            return load_hotwords_in_yaml_file
-         else:
-            def load_hotwords_in_file():
-               with open(p, "r") as f:
-                  return ["".join(i.strip().lower().split()) for i in f.readlines()]
-            load_hotwords_in_file()
-            return load_hotwords_in_file
-      except Exception as e:
-         log(f"$HOTWORDS {p} is not a file: {e}")
-
-      hotwords = ["".join(i.lower().strip().split()) for i in p.split("\/\/")]
-      log(f'$HOTWORDS: {hotwords}')
-      def load_hotwords_as_literal():
-         return hotwords
-      return load_hotwords_as_literal
-
-   def run(self):
-      log("Reactor.run: start")
-      while True:
-         text = self.inq.get()
-         if text is None:
-            break
-         self.handle(text)
-      self.outq.put(None)
-      log("Reactor.run: stop")
-
-   def handle(self, text):
-      hotwords = self.load_hotwords()
-      if os.environ.get("DEBUG", None):
-         log(f"seeking {hotwords} in {text}")
-      if not hotwords:
-         if not os.environ.get("HOTWORDS", None):
-            print(text)
-         else:
-            log(text)
-         return
-      cleantext = "".join([i for i in "".join(text.lower().split()) if i.isalpha()])
-      for i in hotwords:
-         if i in cleantext:
-            #log(f"Reactor.handle: found hotword '{i}' in '{text}' as '{cleantext}'")
-            self.outq.put((i, text))
-
-class Actor(threading.Thread):
-   def __init__(self, inq):
-      threading.Thread.__init__(self)
-      self.inq = inq
-      self.handle = self.handle_stderr
-      if os.environ.get("STDOUT", "") == "true":
-         self.handle = self.handle_stdout
-      elif os.environ.get("SIGUSR2", ""):
-         self.pid = int(environ["SIGUSR2"])
-         self.handle = self.handle_signal
-      elif os.environ.get("URL", ""):
-         self.url = environ["URL"]
-         self.handle = self.handle_url
-         self.headers = [i.split("=")[:2] for i in os.environ.get("HEADERS", "").split("//") if i]
-         self.body = os.environ.get("BODY", '{"hotword":"{{hotword}}","context":"{{context}}"}')
-         log(self.headers)
-
-   def run(self):
-      log("Actor.run: start")
-      while True:
-         got = self.inq.get()
-         if got is None:
-            break
-         self.handle(got[0], got[1])
-      log("Actor.run: stop")
-
-   def handle_stderr(self, hotword, context):
-      log(f"'{hotword}' in '{context}'")
-
-   def handle_stdout(self, hotword, context):
-      log(context)
-      print(hotword)
-
-   def handle_signal(self, hotword, context):
-      self.handle_stderr(hotword, context)
-      os.kill(self.pid, signal.SIGUSR2)
-
-   def handle_url(self, hotword, context):
-      self.handle_stderr(hotword, context)
-      try:
-         headers = {}
-         for i in self.headers:
-            key = i[0]
-            value = i[1]
-            value = value.replace("{{hotword}}", hotword)
-            value = value.replace("{{context}}", context)
-            headers[key] = value
-         body = self.body
-         body = body.replace("{{hotword}}", hotword)
-         body = body.replace("{{context}}", context)
-         if os.environ.get("DEBUG", "") :
-            log("POST", self.url, headers, body)
-         requests.post(self.url, headers=headers, data=body)
-      except Exception as e:
-         log("Actor.handle_url:", e)
-
-def main():
-   managerToParserQ = queue.Queue(maxsize=1)
-   readerToParserQ = queue.Queue(maxsize=10)
-   parserToReactorQ = queue.Queue(maxsize=10)
-   reactorToActorQ = queue.Queue(maxsize=10)
-   threads = [
-      Manager(managerToParserQ),
-      Reader(managerToParserQ, readerToParserQ),
-      Parser(readerToParserQ, parserToReactorQ),
-      Reactor(parserToReactorQ, reactorToActorQ),
-      Actor(reactorToActorQ),
-   ]
-   [t.start() for t in threads]
-   [t.join() for t in threads]
-
-if __name__ == "__main__":
-   main()

whisper-2023/install.sh

@@ -1,7 +0,0 @@
-#! /bin/bash
-
-sudo apt install portaudio19-dev python3-pyaudio
-python3 -m pip install git+https://github.com/openai/whisper.git soundfile PyAudio SpeechRecognition
-
-#sudo apt-get install python3 python3-all-dev python3-pip build-essential swig git libpulse-dev libasound2-dev
-#python3 -m pip install pocketsphinx

whisper-cpp-2023/Dockerfile

@@ -1,20 +0,0 @@
-FROM debian:buster as builder
-
-RUN apt -y update && apt -y install build-essential wget ffmpeg
-
-WORKDIR /tmp/whisper-cpp.git.d
-RUN wget https://github.com/ggerganov/whisper.cpp/archive/refs/tags/v1.2.1.tar.gz \
-   && tar -xf ./*.tar.gz \
-   && mv ./whisper*/ ./git.d
-WORKDIR /tmp/whisper-cpp.git.d/git.d
-RUN make && make samples
-
-FROM debian:buster
-RUN apt -y update && apt -y install curl
-COPY --from=builder /tmp/whisper-cpp.git.d/git.d/ /whisper-cpp.git.d/
-WORKDIR /whisper.d
-RUN bash /whisper-cpp.git.d/models/download-ggml-model.sh tiny.en
-
-ENTRYPOINT []
-CMD /whisper-cpp.git.d/main -m /whisper-cpp.git.d/models/ggml-tiny.en.bin -f /whisper-cpp.git.d/samples/gb1.wav -t 4
-

whisper-cpp-2023/c-whisper

@@ -1 +0,0 @@
-git.d/main

whisper-cpp-2023/compare.md

@@ -1,14 +0,0 @@
-# git.d/samples/mm0.wav (30s)
-
-| model | threads | rust          | c           |
-| ----- | ------- | ------------- | ----------- |
-| tiny  | 1       | 4.4s@122%     | 4.9s@125%   |
-| tiny  | 2       | 2.7s@210%     | 3.3s@190%   |
-| tiny  | 4       | 2.0s@-        | 2.9s@400%   |
-| tiny  | 8       | 2.0s@-        | 3.1s@700%   |
-| small | 1       | 23.9s@175%    | 28.5s@205%  |
-| small | 2       | 14.9s@347%    | 19.2s@330%  |
-| small | 4       | 12.3s@515%    | 22.1s@530%  |
-| base  | 1       | 8.7s@150%     | 10.2s@155%  |
-| base  | 2       | 5.1s@240%     | 7.1s@270%   |
-| base  | 4       | 3.8s@370%     | 6.0s@430%   |

whisper-cpp-2023/compare.sh

@@ -1,7 +0,0 @@
-export P=${1:-1}
-export MODEL=${2:-models/ggml-tiny.en.bin}
-export WAV=${3:-git.d/samples/jfk.wav}
-echo === RUST
-time rust-whisper 2>&1 | grep -v ^whisper_ | grep ..
-echo === C
-time ./c-whisper -m $MODEL -f $WAV -t $P 2>&1 | grep -v ^whisper_ | grep -v ^system_info | grep -v ^main: | grep ..

whisper-cpp-2023/libwhisper.a

@@ -1 +0,0 @@
-git.d/libwhisper.a

whisper-cpp-2023/models

@@ -1 +0,0 @@
-git.d/models

whisper-cpp-2023/native.sh

@@ -1,24 +0,0 @@
-#! /bin/bash
-
-if [ ! -d ./git.d/.git ]; then
-   git clone https://github.com/ggerganov/whisper.cpp.git git.d
-fi
-
-cd ./git.d
-if [ ! -f ./samples/gb1.wav ]; then
-   make samples
-fi
-if [ ! -f ./main ]; then
-   make
-fi
-if [ ! -f ./stream ]; then
-   make stream
-fi
-if [ ! -f ./models/ggml-${MODEL:-tiny.en}.bin ]; then
-   bash ./models/download-ggml-model.sh ${MODEL:-tiny.en}
-fi
-if [ -n "$STREAM" ]; then
-   ./stream -m ./models/ggml-${MODEL:-tiny.en}.bin -t 8 --step 500 --length ${MIC_TIMEOUT:-2}000 $(test -n "$MIC_ID" && echo -c "$MIC_ID")
-else
-   time ./main -m ./models/ggml-${MODEL:-tiny.en}.bin -f ./samples/gb1.wav -t 4
-fi

whisper-cpp-2023/rust.d/src/main.rs

@@ -1,37 +0,0 @@
-use whisper_rs::{WhisperContext, FullParams, SamplingStrategy};
-
-fn main() {
-   let mut ctx = WhisperContext::new(
-      &std::env::var("MODEL").unwrap_or(String::from("../models/ggml-tiny.en.bin"))
-   ).expect("failed to load model");
-
-   // create a params object
-   let mut params = FullParams::new(SamplingStrategy::Greedy { best_of: 0 });
-   params.set_n_threads(
-      std::env::var("P").unwrap_or(String::from("1")).parse::<i32>().expect("$P must be a number")
-   );
-   params.set_translate(false);
-   params.set_language(Some("en"));
-   params.set_print_special(false);
-   params.set_print_progress(false);
-   params.set_print_realtime(false);
-   params.set_print_timestamps(false);
-
-   let (header, data) = wav::read(&mut std::fs::File::open(
-      &std::env::var("WAV").unwrap_or(String::from("../git.d/samples/jfk.wav"))
-   ).expect("failed to open .wav")).expect("failed to decode .wav");
-   assert!(header.channel_count == 1);
-   assert!(header.sampling_rate == 16000);
-   let data16 = data.as_sixteen().expect("wav is not 32bit floats");
-   let audio_data = &whisper_rs::convert_integer_to_float_audio(&data16);
-
-   ctx.full(params, &audio_data[..])
-      .expect("failed to run model");
-
-   let num_segments = ctx.full_n_segments();
-   for i in 0..num_segments {
-      let segment = ctx.full_get_segment_text(i).expect("failed to get segment");
-      print!("{} ", segment);
-   }
-   println!("");
-}

whisper-cpp-2023/whisper.h

@@ -1 +0,0 @@
-git.d/whisper.h