Python Developer for PDF Book → Database & Auto-Update Tool (ETL) System
Budget: $10 – $30 AUD
Description:
We need a Python developer to build a system that ingests manufacturer PDF price books, extracts structured pricing data, and stores it in a normalized database. The tool must detect updates when new editions are uploaded, highlight changes, and provide a lightweight admin UI for upload, preview, comparison, and export.
Phase 1 – Price Book ETL Tool
- Parse both digital and scanned PDFs (OCR fallback).
- Extract items, finishes, options, dimensions, pricing, and effective dates.
- Normalize outputs (Excel/CSV + SQL or Baserow).
- Auto-diff engine: update prices, insert/remove SKUs, log changes.
- Lightweight UI for upload, preview, compare, and export.
- Accuracy targets: ≥98% rows, ≥99% numeric values.
Phase 2 – Configuration & Pricing Tool
- Build configurable product sets (assemblies/kits) using parsed data.
- Apply rules: requires/excludes, weight limits, finish mappings, net add/percent add.
- Pricing engine: base + adders + discounts with full audit trail.
- Auto-reprice saved sets when new price books are uploaded.
- Exports: Quote-ready Excel/PDF, CSV for ERP, and REST API endpoints.
- Admin UI: drag/drop set builder, validation engine, approval workflow, versioning.
Technical Requirements
- Python 3.11+, pdfplumber/camelot/pdfminer.six, OCR fallback (Tesseract/LayoutParser).
- Postgres/MySQL (or Baserow API).
- Fuzzy matching (Levenshtein/TF-IDF).
- Lightweight web UI (Flask/Django/FastAPI; React optional).
What We Provide
- Two real-world price books (Hager & SELECT Hinges, 2025 editions).
- Vibe coded project as reference
We need a Python developer to build a system that ingests manufacturer PDF price books, extracts structured pricing data, and stores it in a normalized database. The tool must detect updates when new editions are uploaded, highlight changes, and provide a lightweight admin UI for upload, preview, comparison, and export.
Phase 1 – Price Book ETL Tool
- Parse both digital and scanned PDFs (OCR fallback).
- Extract items, finishes, options, dimensions, pricing, and effective dates.
- Normalize outputs (Excel/CSV + SQL or Baserow).
- Auto-diff engine: update prices, insert/remove SKUs, log changes.
- Lightweight UI for upload, preview, compare, and export.
- Accuracy targets: ≥98% rows, ≥99% numeric values.
Phase 2 – Configuration & Pricing Tool
- Build configurable product sets (assemblies/kits) using parsed data.
- Apply rules: requires/excludes, weight limits, finish mappings, net add/percent add.
- Pricing engine: base + adders + discounts with full audit trail.
- Auto-reprice saved sets when new price books are uploaded.
- Exports: Quote-ready Excel/PDF, CSV for ERP, and REST API endpoints.
- Admin UI: drag/drop set builder, validation engine, approval workflow, versioning.
Technical Requirements
- Python 3.11+, pdfplumber/camelot/pdfminer.six, OCR fallback (Tesseract/LayoutParser).
- Postgres/MySQL (or Baserow API).
- Fuzzy matching (Levenshtein/TF-IDF).
- Lightweight web UI (Flask/Django/FastAPI; React optional).
What We Provide
- Two real-world price books (Hager & SELECT Hinges, 2025 editions).
- Vibe coded project as reference