rehan953 commited on
Commit
ef56da6
·
verified ·
1 Parent(s): f035ea9

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +7 -0
app.py CHANGED
@@ -1057,6 +1057,13 @@ def _extract_textlayer_rows(pdf_path: str, page_num: int):
1057
  if not desc:
1058
  continue
1059
 
 
 
 
 
 
 
 
1060
  rows.append({"date": date_str, "post_date": post_date_str, "desc": desc, "amount": amount_candidate})
1061
 
1062
  # Guard: require at least 2 rows to avoid false positives on non-transaction pages
 
1057
  if not desc:
1058
  continue
1059
 
1060
+ # Guard: description must contain at least one letter.
1061
+ # Rows where description is purely digits/dates/amounts are
1062
+ # DAILY BALANCE rows (e.g. "170,198.04 01-13 45,442.31 01-24"),
1063
+ # not real transactions — skip them.
1064
+ if not re.search(r"[A-Za-z]", desc):
1065
+ continue
1066
+
1067
  rows.append({"date": date_str, "post_date": post_date_str, "desc": desc, "amount": amount_candidate})
1068
 
1069
  # Guard: require at least 2 rows to avoid false positives on non-transaction pages