Update TextProcessor.py

TextProcessor.py

@@ -1,4 +1,3 @@
-# This file converts the images into text
 import pandas as pd
 import numpy as np
 import fitz
@@ -53,20 +52,21 @@ def paragraphs(folderpath):
             current_supertabbed = x1s[j] - baselines[pg_inds[j]] >= 14
             prior_more_left = (x1s[j] - baselines[pg_inds[j]]) - (x1s[j - 1] - baselines[pg_inds[j - 1]]) > 7
             prior_right_margin = x1s[j - 1] > prior_median
+            current_right_margin = x1s[j] > prior_median
 
             is_section_header = (prior_is_section_header or current_is_section_header or prior_is_asterisk or current_is_asterisk or prior_is_date)
             is_the_classic = (prior_endswith_period and current_tabbed and prior_more_left and not prior_supertabbed)  # Note: Supertabbing oofs stuff
             is_start_blockquote = (prior_endswith_period and current_supertabbed and prior_more_left and not is_inblock)
             is_after_blockquote = (prior_endswith_period and not current_supertabbed and is_inblock)
             is_after_disposition = (prior_right_margin and current_tabbed)
+            is_disposition = (current_right_margin and ("affirm" in line_texts[j].lower()) or "reverse" in line_texts[j].lower() or "vacate" in line_texts[j].lower() or "so ordered" in line_texts[j].lower())
 
             if is_start_blockquote and not is_section_header:
                 is_inblock = True
-                # print("START BLOCK")
             if is_after_blockquote or prior_is_date:  # This date hack helps ensure that slip opinion headers do not get caught
                 is_inblock = False
 
-            if is_section_header or is_the_classic or is_start_blockquote or is_after_blockquote or is_after_disposition:
+            if is_section_header or is_the_classic or is_start_blockquote or is_after_blockquote or is_after_disposition or is_disposition:
                 paras.append(para)
                 para = []
             #     print('\n')
@@ -96,5 +96,4 @@ def process_file(folderpath):
         x1, y1, x2, y2, para_first_line, pg_ind = indent
         image = cv2.imread(folderpath + '/' + str(pg_ind) + '-processed.png')
         cv2.circle(image, (x1 - 15, int(0.5 * (y1 + y2))), radius=1, color=(240, 32, 160), thickness=2)
-        cv2.imwrite(folderpath + '/' + str(pg_ind) + '-processed.png', image)
-
+        cv2.imwrite(folderpath + '/' + str(pg_ind) + '-processed.png', image)