Add bank statement analyzer with automatic work storage

2026-07-13 10:12:19 +05:30
parent d391e9b443
commit bbc5afe1c0
23 changed files with 1047 additions and 0 deletions
@@ -0,0 +1,37 @@
+from __future__ import annotations
+import re, pandas as pd
+from pathlib import Path
+from .base import *
+from .common import find
+
+class AxisParser(BaseParser):
+    bank_name='Axis Bank'; parser_name='AxisParser'
+    @classmethod
+    def detect(cls,text): return 0.98 if 'SMART STATEMENT REPORT' in text.upper() and 'UTIB' in text.upper() else 0
+    def parse(self,path,text=None):
+        text=text or extract_text(path); meta=StatementMeta(bank_name=self.bank_name,source_file=Path(path).name,parser_name=self.parser_name,confidence='High')
+        # name is first meaningful line after report title
+        m=re.search(r'Smart Statement Report\s*\n\s*([^\n]+)',text,re.I); meta.customer_name=norm(m.group(1)) if m else ''
+        meta.account_number=find(r'Statement of Account No\s*-\s*([^\s]*)',text)
+        meta.ifsc=find(r'IFSC:\s*([A-Z0-9]+)',text)
+        m=re.search(r'for period\s*\((\d{2}/\d{2}/\d{4})\s+to\s+(\d{2}/\d{2}/\d{4})\)',text,re.I)
+        if m: meta.period_from=pd.to_datetime(m.group(1),dayfirst=True).strftime('%Y-%m-%d'); meta.period_to=pd.to_datetime(m.group(2),dayfirst=True).strftime('%Y-%m-%d')
+        meta.opening_balance=amount(find(r'Opening Balance:\s*INR\s*([\d,]+\.\d{2})',text))
+        pat=re.compile(r'^\s*(\d+)\s+(\d{2}/\d{2}/\d{4})\s+(\d{2}/\d{2}/\d{4})\s+(.*)$')
+        rows=[]; cur=None; page=1
+        for line in text.splitlines():
+            if '\f' in line: page+=line.count('\f')
+            m=pat.match(line)
+            if m:
+                if cur: rows.append(cur)
+                cur={'transaction_date':m.group(2),'value_date':m.group(3),'body':m.group(4),'source_page':page}
+            elif cur and line.strip() and not re.match(r'^(S\. No\.|Smart Statement|Page )',line.strip(),re.I): cur['body']+=' '+line.strip()
+        if cur: rows.append(cur)
+        out=[]
+        for r in rows:
+            b=norm(r['body']); ma=re.search(r'INR\s*([\d,]+\.\d{2})\s+(CR|DR)\s+INR\s*([\d,]+\.\d{2})',b,re.I)
+            if not ma: continue
+            txn=amount(ma.group(1)); typ=ma.group(2).upper(); bal=amount(ma.group(3)); narr=b[:ma.start()].strip(); ref=''
+            z=re.search(r'([A-Z0-9/-]{8,})',narr); ref=z.group(1) if z else ''
+            out.append({**r,'narration':narr,'reference_no':ref,'debit':txn if typ=='DR' else None,'credit':txn if typ=='CR' else None,'balance':bal})
+        return meta,finalize(pd.DataFrame(out),meta)