Challenge
Unstructured CDSL PDF statements made Demat holding and mutual fund extraction tedious and error-prone.
Constraints
- —No API access to CDSL PDFs
- —Mixed layouts across statement variants
- —Process repeated by non-technical staff monthly
- —Output needed to be compatible with Excel automation
Architecture
- •Python parser using PyMuPDF and tabula-py
- •Tabular extraction normalized into typed records
- •Pandas transformations before Excel output
- •Packaged as Tauri desktop app for client deployment
Key Decisions
- •Kept parsing deterministic to minimize hard-to-debug PDF drift
- •Built Excel export path first because client team sort/filter in spreadsheets
Impact
Desktop executable successfully deployed at the client, converting raw CDSL PDFs into structured Excel reports automatically.
Internal enterprise tool
Stack
PythonTauriPyMuPDFtabula-pyPandas