Update app.py

app.py

@@ -1076,78 +1076,126 @@ def _extract_textlayer_rows(pdf_path: str, page_num: int):
 
 def _extract_jb_summary_sections(text_layer: str, needs_checks: bool = True, needs_dab: bool = True) -> str:
     """
-    Extract Johnson Bank Checks and Daily Account Balance sections from the PDF
-    text layer and return them as HTML tables.
+    Parse Johnson Bank Checks and Daily Account Balance from pdfminer column-separated
+    page text and return as HTML tables.
 
-    Only fires when MM-DD date format is detected (Johnson Bank unique identifier).
-    needs_checks / needs_dab flags prevent re-extraction when already present.
+    pdfminer outputs multi-column sections as separate column lists:
+        Date      Number    Amount
+        04-14     3259      45.72
+    becomes:
+        "Date\n04-14\n04-09\n\nNumber\n3259\n3260\n\nAmount\n45.72\n1,628.00"
+
+    Uses pdfminer (always available) not pymupdf.
     """
     if not text_layer:
         return ""
 
-    _CHK_SEC = re.compile(r"Checks\s*\(", re.IGNORECASE)
-    _DAB_SEC = re.compile(r"^Daily\s+Account\s+Balance\s*$", re.IGNORECASE | re.MULTILINE)
-    _CHK_ROW = re.compile(r"^(\d{2}-\d{2})\s+(\*?\d+)\s+(\d{1,3}(?:,\d{3})*\.\d{2})\s*$")
-    _BAL_ROW = re.compile(r"^(\d{2}-\d{2})\s+(\d{1,3}(?:,\d{3})*\.\d{2})\s*$")
+    _DATE_RE  = re.compile(r"^\d{2}-\d{2}$")
+    _MONEY_RE = re.compile(r"^\d{1,3}(?:,\d{3})*\.\d{2}$")
+    _CHECK_RE = re.compile(r"^\*?\d+$")
 
-    lines  = text_layer.splitlines()
+    lines = [l.strip() for l in text_layer.splitlines()]
     output = []
 
-    # ── Checks section ─────────────────────────────────────────────────────
+    # ── Checks ───────────────────────────────────────────────────────────
     if needs_checks:
         checks = []
-        in_checks = False
-        for line in lines:
-            stripped = line.strip()
-            if _CHK_SEC.search(stripped):
-                in_checks = True
-                continue
-            if in_checks:
-                if _DAB_SEC.match(stripped):
-                    break
-                m = _CHK_ROW.match(stripped)
-                if m:
-                    checks.append((m.group(1), m.group(2), m.group(3)))
+        i = 0
+        while i < len(lines):
+            if lines[i].lower().startswith("checks"):
+                i += 1
+                while i < len(lines):
+                    if "daily account balance" in lines[i].lower():
+                        break
+                    if lines[i] == "Date":
+                        i += 1
+                        dates = []
+                        while i < len(lines) and _DATE_RE.match(lines[i]):
+                            dates.append(lines[i]); i += 1
+                        while i < len(lines) and lines[i] != "Number":
+                            i += 1
+                        i += 1
+                        numbers = []
+                        while i < len(lines):
+                            if _CHECK_RE.match(lines[i]):
+                                numbers.append(lines[i]); i += 1
+                            elif lines[i] == "":
+                                i += 1
+                            else:
+                                break
+                        while i < len(lines) and lines[i] != "Amount":
+                            i += 1
+                        i += 1
+                        amounts = []
+                        while i < len(lines):
+                            if _MONEY_RE.match(lines[i]):
+                                amounts.append(lines[i]); i += 1
+                            elif lines[i] == "":
+                                i += 1
+                            else:
+                                break
+                        for d, n, a in zip(dates, numbers, amounts):
+                            checks.append((d, n, a))
+                    else:
+                        i += 1
+                break
+            i += 1
 
         if checks:
-            rows_html = "\n".join(
+            rows = "\n".join(
                 "<tr><td>" + d + "</td><td>" + n + "</td><td>" + a + "</td></tr>"
                 for d, n, a in checks
             )
             output.append("Checks")
             output.append(
                 "<table>\n<tr><th>Date</th><th>Number</th><th>Amount</th></tr>\n"
-                + rows_html + "\n</table>"
+                + rows + "\n</table>"
             )
 
-    # ── Daily Account Balance section ───────────────────────────────────────
+    # ── Daily Account Balance ─────────────────────────────────────────────
     if needs_dab:
         balances = []
-        in_bal = False
-        for line in lines:
-            stripped = line.strip()
-            if _DAB_SEC.match(stripped):
-                in_bal = True
-                continue
-            if in_bal:
-                m = _BAL_ROW.match(stripped)
-                if m:
-                    balances.append((m.group(1), m.group(2)))
+        i = 0
+        while i < len(lines):
+            if "daily account balance" in lines[i].lower():
+                i += 1
+                while i < len(lines):
+                    if lines[i] == "Date":
+                        i += 1
+                        dates = []
+                        while i < len(lines) and _DATE_RE.match(lines[i]):
+                            dates.append(lines[i]); i += 1
+                        while i < len(lines) and lines[i] != "Balance":
+                            i += 1
+                        i += 1
+                        bals = []
+                        while i < len(lines):
+                            if _MONEY_RE.match(lines[i]):
+                                bals.append(lines[i]); i += 1
+                            elif lines[i] == "":
+                                i += 1
+                            else:
+                                break
+                        for d, b in zip(dates, bals):
+                            balances.append((d, b))
+                    else:
+                        i += 1
+                break
+            i += 1
 
         if balances:
-            rows_html = "\n".join(
+            rows = "\n".join(
                 "<tr><td>" + d + "</td><td>" + b + "</td></tr>"
                 for d, b in balances
             )
             output.append("Daily Account Balance")
             output.append(
                 "<table>\n<tr><th>Date</th><th>Balance</th></tr>\n"
-                + rows_html + "\n</table>"
+                + rows + "\n</table>"
             )
 
     return "\n\n".join(output)
 
-
 def _patch_ocr_with_textlayer(page_md: str, pdf_path: str, page_num: int) -> str:
     """
     Compare OCR table rows against PDF text-layer rows and inject any rows
@@ -2369,13 +2417,25 @@ def run_ocr(uploaded_file):
                         needs_checks = "checks" not in stabilized.lower() or "<th>Number</th>" not in stabilized
                         needs_dab    = "daily account balance" not in stabilized.lower()
                         if needs_checks or needs_dab:
-                            import pymupdf as _fitz4d
-                            _doc4d  = _fitz4d.open(path)
-                            _txt4d  = _doc4d[page_num].get_text()
-                            _doc4d.close()
-                            jb_extra = _extract_jb_summary_sections(_txt4d, needs_checks, needs_dab)
-                            if jb_extra:
-                                stabilized = stabilized.rstrip() + "\n\n" + jb_extra
+                            # Try pdfminer first (column-aware), fall back to pymupdf
+                            _txt4d = ""
+                            try:
+                                from pdfminer.high_level import extract_text as _pm_extract
+                                _txt4d = _pm_extract(path, page_numbers=[page_num])
+                            except Exception:
+                                pass
+                            if not _txt4d:
+                                try:
+                                    import pymupdf as _fitz4d
+                                    _d4 = _fitz4d.open(path)
+                                    _txt4d = _d4[page_num].get_text()
+                                    _d4.close()
+                                except Exception:
+                                    pass
+                            if _txt4d:
+                                jb_extra = _extract_jb_summary_sections(_txt4d, needs_checks, needs_dab)
+                                if jb_extra:
+                                    stabilized = stabilized.rstrip() + "\n\n" + jb_extra
                     except Exception:
                         pass  # safe no-op — never break other PDFs