From 54372c88f334f4940b485164787c3c499266ed69 Mon Sep 17 00:00:00 2001
From: bel <bel@bel.bel>
Date: Tue, 21 Feb 2023 12:15:30 -0700
Subject: [PATCH] dont think i will use but page to plaintext

---
 main.py | 23 ++++++++++++++++++++++-
 1 file changed, 22 insertions(+), 1 deletion(-)

diff --git a/main.py b/main.py
index f9a0332..64ad4d5 100644
--- a/main.py
+++ b/main.py
@@ -22,12 +22,33 @@ def de_column_ify():
             got = q.get()
             for got_i in got[1]:
                if got_i.chars:
-                  debug.debug_show(debug.debug_im(got_i.page))
+                  True or debug.debug_show(debug.debug_im(got_i.page))
+               textify(got_i.page)
 
 def de_columnify_page(q, page):
    result = cluster.Chars(page.chars, page).divide_into_columns()
    log("putting", page.page_number)
    q.put((page, result))
 
+def textify(page):
+   lines = page.extract_text(layout=True).split("\n")
+   leading_spaces = []
+   for line in lines:
+      leading_spaces.append(len(line) - len(line.lstrip()))
+   median_leading_spaces = sorted(leading_spaces)[len(leading_spaces)//2]
+
+   paragraphs = []
+   for line in lines:
+      if line.startswith(median_leading_spaces * " "):
+         line = line[median_leading_spaces:]
+      if line.startswith(" "):
+         line = "\t" + line[1:]
+      if not paragraphs or line.startswith("\t"):
+         paragraphs.append([])
+         paragraphs.append([])
+      paragraphs[-1].append(line.rstrip())
+   for paragraph in paragraphs:
+      print(" ".join(paragraph))
+
 if __name__ == "__main__":
    main()