aal-bd/po-reconciliation-filter

📦 PO-Based Order Reconciliation Filter System

Author: Mohammad Ali — ANANTA Automation Suite Platform: HuggingFace Docker Space Version: v4.1-HF Base Algorithm: Colab v4 FINAL (100% identical — no logic changes)

★ What This System Does

Filters a large Order Reconciliation Excel Report down to only the POs you need, preserving 100% of the original formatting — merged cells, fonts, colors, borders, column widths, row heights, and freeze panes — in the exact order specified.

System Capabilities

Pipeline

Step 1 — Parse Needed PO file        → ordered PO list
Step 2 — Scan Source (read-only)     → build PO→rows block map
Step 3 — Build ordered row list      → with deduplication
Step 4 — Load Source (full mode)     → access styles & merges
Step 5 — Create output workbook      → copy widths, freeze panes, tab color
Step 6 — Copy header row (row 1)     → with full style
Step 7 — Copy PO data rows           → in Needed PO file order
Step 8 — Translate merge ranges      → consecutive-row validation
Step 9 — Save output .xlsx           → download ready

v4 Grouped PO Detection

The source file sometimes places multiple POs in a single cell, separated by /:

7100026049/ 4300195616/4628 /4625/ 4605 /4604

This resolves to: 7100026049, 4300195616, 4300194628, 4300194625, 4300194605, 4300194604

Abbreviated PO Rule: 4628 → prefix of last full PO (430019) + 4628 = 4300194628

A row is included if any PO in its group matches the Needed PO list.

HF Space Name (Recommended)

po-reconciliation-filter

Full URL:

https://{your-hf-username}-po-reconciliation-filter.hf.space/

Deployment Instructions

1. Create the HF Space

1. 1.Go to huggingface.co/new-space
2. 2.Space name: po-reconciliation-filter
3. 3.SDK: Docker
4. 4.Visibility: Private (recommended for enterprise use)

2. Upload These 4 Files

Upload all 4 files to the root of your new Space:

• —app.py
• —Dockerfile
• —requirements.txt
• —README.md

3. Space will auto-build

HuggingFace will auto-detect the Dockerfile and start building. Build time: ~3–5 minutes (no Playwright — much faster than the RPA suite).

4. No configuration needed

All settings are embedded in app.py. No credentials, no API keys, no environment variables required.

Version History

v4.1-HF  — HuggingFace Docker deployment
           • Enterprise-grade UI redesign (dark log monitor, blue banner)
           • OUTPUT_DIR = /home/user/app/outputs (HF-compatible)
           • server_name="0.0.0.0", server_port=7860
           • No share=True (HF handles public URL)
           • All processing logic: IDENTICAL to Colab v4 FINAL

v4.0     — Colab FINAL release
           • Grouped PO detection & abbreviated PO expansion
           • Row deduplication for shared grouped POs
           • Full merge + format preservation

v3.x     — Previous format-preservation work

Gradio Version Note

Pinned to gradio==3.50.2 / gradio_client==0.6.1 — same as the enterprise-rpa-sewing-only reference deployment. This avoids the known Gradio/HuggingFace conflict affecting newer releases. Do not upgrade without testing on HF.