From 42605c706b7d90e6101840b5bba88bb3f4041447 Mon Sep 17 00:00:00 2001
From: Bel LaPointe <breel@qualtrics.com>
Date: Mon, 31 Jan 2022 15:27:18 -0700
Subject: [PATCH] get gitlab scrape scripts from hackathon

---
 spike/crawl/env.env               |  3 ++
 spike/crawl/gitlab-wiki/expand.sh | 64 +++++++++++++++++++++++++++++++
 spike/crawl/gitlab/expand.sh      | 64 +++++++++++++++++++++++++++++++
 3 files changed, 131 insertions(+)
 create mode 100644 spike/crawl/env.env
 create mode 100644 spike/crawl/gitlab-wiki/expand.sh
 create mode 100644 spike/crawl/gitlab/expand.sh

diff --git a/spike/crawl/env.env b/spike/crawl/env.env
new file mode 100644
index 0000000..3f494d6
--- /dev/null
+++ b/spike/crawl/env.env
@@ -0,0 +1,3 @@
+GITLAB_PAT=LkaMCFFfqdHd-r4Cubnw
+ODO_TOKEN=ac9a9e4d-9c6b-4049-9e8d-c8b97fe053aa
+GDOC_TOKEN=
diff --git a/spike/crawl/gitlab-wiki/expand.sh b/spike/crawl/gitlab-wiki/expand.sh
new file mode 100644
index 0000000..14594bb
--- /dev/null
+++ b/spike/crawl/gitlab-wiki/expand.sh
@@ -0,0 +1,64 @@
+#! /bin/bash
+
+urlencode() {
+    # urlencode <string>
+
+    old_lc_collate=$LC_COLLATE
+    LC_COLLATE=C
+
+    local length="${#1}"
+    for (( i = 0; i < length; i++ )); do
+        local c="${1:$i:1}"
+        case $c in
+            [a-zA-Z0-9.~_-]) printf '%s' "$c" ;;
+            *) printf '%%%02X' "'$c" ;;
+        esac
+    done
+
+    LC_COLLATE=$old_lc_collate
+}
+
+pat="${GITLAB_PAT:-"$(security find-generic-password -a "${USER}" -s GITLAB_PAT -w 2> /dev/null )"}"
+project="$(urlencode ${PROJECT:-"data-store/orchestration/runbooks"})"
+path="/api/v4/projects/$project/repository/tree"
+list() {
+   local query="recursive=true&path=$(test -n "$GITLAB_PATH" && echo "$GITLAB_PATH")"
+   curl -sS \
+      -H "Authorization: Bearer $pat" \
+      "https://gitlab-app.eng.qops.net/$path?$query"
+}
+list_wiki() {
+   curl -sS \
+      -H "Authorization: Bearer $pat" \
+      "https://gitlab-app.eng.qops.net/api/v4/projects/$project/wikis?with_content=0"
+}
+files=()
+trees=("${GITLAB_PATH:-"/"}")
+i=0
+while [ "$i" -lt "${#trees[@]}" ]; do
+   gitlab_path="${trees[i]}"
+   echo gitlab_path=$gitlab_path, i=$i, trees=${#trees[@]}, files=${#files[@]}... >&2
+   got="$(GITLAB_PATH=$gitlab_path list)"
+   for nested_path in $(echo "$got" | jq -c .[] | grep '"type":"tree"' | jq -r .path | while read -r line; do echo "$line" | base64; done | grep .); do
+      nested_path="$(echo "$nested_path" | base64 --decode)"
+      if echo "${trees[@]}" | grep -q "[ ^]$nested_path[ $]"; then
+         continue
+      fi
+      trees+=("$nested_path")
+   done
+   for nested_path in $(echo "$got" | jq -c .[] | grep '"type":"blob"' | jq -r .path | while read -r line; do echo "$line" | base64; done | grep .); do
+      nested_path="$(echo "$nested_path" | base64 --decode)"
+      if echo "${files[@]}" | grep -q "[ ^]$nested_path[ $]"; then
+         continue
+      fi
+      files+=("$nested_path")
+   done
+   i=$((i+1))
+   if ((i>5)); then
+      break
+   fi
+done
+for file in "${files[@]}"; do
+   echo "gitlab://$file"
+done | sort
+list_wiki | jq -c .[] | jq -r .title | sed 's/^/wiki:\/\//'
diff --git a/spike/crawl/gitlab/expand.sh b/spike/crawl/gitlab/expand.sh
new file mode 100644
index 0000000..14594bb
--- /dev/null
+++ b/spike/crawl/gitlab/expand.sh
@@ -0,0 +1,64 @@
+#! /bin/bash
+
+urlencode() {
+    # urlencode <string>
+
+    old_lc_collate=$LC_COLLATE
+    LC_COLLATE=C
+
+    local length="${#1}"
+    for (( i = 0; i < length; i++ )); do
+        local c="${1:$i:1}"
+        case $c in
+            [a-zA-Z0-9.~_-]) printf '%s' "$c" ;;
+            *) printf '%%%02X' "'$c" ;;
+        esac
+    done
+
+    LC_COLLATE=$old_lc_collate
+}
+
+pat="${GITLAB_PAT:-"$(security find-generic-password -a "${USER}" -s GITLAB_PAT -w 2> /dev/null )"}"
+project="$(urlencode ${PROJECT:-"data-store/orchestration/runbooks"})"
+path="/api/v4/projects/$project/repository/tree"
+list() {
+   local query="recursive=true&path=$(test -n "$GITLAB_PATH" && echo "$GITLAB_PATH")"
+   curl -sS \
+      -H "Authorization: Bearer $pat" \
+      "https://gitlab-app.eng.qops.net/$path?$query"
+}
+list_wiki() {
+   curl -sS \
+      -H "Authorization: Bearer $pat" \
+      "https://gitlab-app.eng.qops.net/api/v4/projects/$project/wikis?with_content=0"
+}
+files=()
+trees=("${GITLAB_PATH:-"/"}")
+i=0
+while [ "$i" -lt "${#trees[@]}" ]; do
+   gitlab_path="${trees[i]}"
+   echo gitlab_path=$gitlab_path, i=$i, trees=${#trees[@]}, files=${#files[@]}... >&2
+   got="$(GITLAB_PATH=$gitlab_path list)"
+   for nested_path in $(echo "$got" | jq -c .[] | grep '"type":"tree"' | jq -r .path | while read -r line; do echo "$line" | base64; done | grep .); do
+      nested_path="$(echo "$nested_path" | base64 --decode)"
+      if echo "${trees[@]}" | grep -q "[ ^]$nested_path[ $]"; then
+         continue
+      fi
+      trees+=("$nested_path")
+   done
+   for nested_path in $(echo "$got" | jq -c .[] | grep '"type":"blob"' | jq -r .path | while read -r line; do echo "$line" | base64; done | grep .); do
+      nested_path="$(echo "$nested_path" | base64 --decode)"
+      if echo "${files[@]}" | grep -q "[ ^]$nested_path[ $]"; then
+         continue
+      fi
+      files+=("$nested_path")
+   done
+   i=$((i+1))
+   if ((i>5)); then
+      break
+   fi
+done
+for file in "${files[@]}"; do
+   echo "gitlab://$file"
+done | sort
+list_wiki | jq -c .[] | jq -r .title | sed 's/^/wiki:\/\//'