From 1119c46b97101398183ede18e5b55e742f66ab51 Mon Sep 17 00:00:00 2001
From: bel <bel@bel.bel>
Date: Tue, 21 Feb 2023 12:39:22 -0700
Subject: [PATCH] cluster.Chars has .path

---
 cluster.py      | 5 +++--
 main.py         | 6 +++---
 test_cluster.py | 2 +-
 3 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/cluster.py b/cluster.py
index 4319e76..c8c0619 100644
--- a/cluster.py
+++ b/cluster.py
@@ -1,7 +1,8 @@
 import config
 
 class Chars:
-   def __init__(self, chars, page):
+   def __init__(self, path, chars, page):
+      self.path = path
       self.chars = chars
       self.page = page
       self.n = 0
@@ -44,7 +45,7 @@ class Chars:
          if result and result[-1]._box().overlaps(Box.from_char(char), clearance=median_x_delta_when_same_y):
             result[-1].merge_in(char)
          else:
-            result.append(Chars([char], self.page))
+            result.append(Chars(self.path, [char], self.page))
       result = [i for i in result if i.n > 2]
 
       # any clusters shorter than median character and high/lower are header/footer
diff --git a/main.py b/main.py
index 4e241fc..f9b09fd 100644
--- a/main.py
+++ b/main.py
@@ -16,7 +16,7 @@ def de_column_ify():
    with pdfplumber.open(config.INPUT) as pdf:
       with ThreadPool(4) as pool:
          for i in range(len(pdf.pages)):
-            pool.apply_async(de_columnify_page, (q, pdf.pages[i], ))
+            pool.apply_async(de_columnify_page, (q, config.INPUT, pdf.pages[i], ))
          for i in range(len(pdf.pages)):
             log("getting", i, "of", len(pdf.pages))
             got = q.get()
@@ -24,8 +24,8 @@ def de_column_ify():
                if got_i.chars:
                   debug.debug_show(debug.debug_im(got_i.page))
 
-def de_columnify_page(q, page):
-   result = cluster.Chars(page.chars, page).divide_into_columns()
+def de_columnify_page(q, path, page):
+   result = cluster.Chars(path, page.chars, page).divide_into_columns()
    log("putting", page.page_number)
    q.put((page, result))
 
diff --git a/test_cluster.py b/test_cluster.py
index ffac0d9..adda9ab 100644
--- a/test_cluster.py
+++ b/test_cluster.py
@@ -28,7 +28,7 @@ class TestChars(unittest.TestCase):
                #   for i in range(len(words))
                #])
                #continue
-               got = cluster.Chars(page.chars, page).divide_into_columns()
+               got = cluster.Chars(p, page.chars, page).divide_into_columns()
                print(p)
                debug.draw_boxes(page, [
                   {