What's Changed

Release history for VA Logistics Optimization Platform

Release 2 v002 2026-02-26
Classification engine Quarantine system Backfill report Mobile layout Release notes page E2E test suite Bulk optimization (17x) Structured logging
New Features
  • 4-category classification engine assigns slips to machine, PO box, non-continental, and accessory zones.
  • Quarantine system with duplicate-recipient and address-match rules, release workflow, and auto-evaluation on upload.
  • Smart batching groups classified slips by category with SKU-signature matching and item count tallies.
  • Backfill report showing warehouse replenishment needs with configurable par levels and case-based calculations.
  • Image upload support (PNG/JPG/TIFF) alongside PDF, with auto-batching on upload completion.
  • Slip preview popover on hover showing recipient and item details without page navigation.
  • What's Changed page showing versioned release notes with category groupings, highlight chips, and version badge in navigation.
Improvements
  • Mobile-responsive CSS for all pages including Dashboard, Batches, and Backfill views.
  • Quick access reorganized into two-row layout with batch category filter buttons.
  • Inline confirmation dialogs replace browser confirm() for destructive actions.
  • QR code popover for batch labels with encoded item tallies.
  • Item count display added to recent batches table on dashboard.
Performance
  • Bulk database operations replace N+1 queries: demo data generation reduced from ~60s to 3.6s (17x speedup).
  • Single-transaction bulk inserts for slips and items (1000 slips + 5000 items in 24ms).
  • Pre-loaded item lookup eliminates per-slip queries in classification, batching, and quarantine evaluation.
  • Docker auto-seeding: warehouse config (498 items) seeds automatically on container startup via entrypoint.sh.
Infrastructure
  • Docker support for Archive Flask app with gunicorn worker configuration.
  • SpecKit remediation across Archive specs for cross-artifact consistency.
  • Generated 16 strategic demo PDFs covering all 4 categories and both quarantine rules.
  • Comprehensive E2E test suite: 5 Playwright suites (31 tests) covering smoke tests, PDF upload, batching, slips/quarantine, and inventory/fulfillment views.
  • Comprehensive structured logging with Azure Application Insights integration, flat custom dimensions, and OTel auto-instrumentation.
  • Log operation wrapper on 30+ database functions with timing, error tracking, and HIPAA-safe dimensions.
  • All 5 scripts converted from print() to structured logging with consistent formatters and filters.
  • Docker json-file log driver with 10MB/3-file rotation limits to prevent disk fill.
  • 25 logging tests covering setup, formatters, filters, environment loading, and log_operation() timing.
Bug Fixes
  • Demo data generation fixed for backfill report compatibility.
  • Documentation inconsistencies fixed with port standardized to 5001.
  • Corrected demo PDF batch count from 9 to 8 across all documentation and scripts.
Release 1 v001 2026-02-12
OCR pipeline REST API PostgreSQL backend 174 tests
New Features
  • Complete OCR-to-batch pipeline: PDF upload, PaddleOCR extraction, SKU grouping, SmartBatch ID assignment.
  • FastAPI REST API with X-API-Key authentication, upload/batch/status/query endpoints.
  • PostgreSQL database layer with SQLAlchemy ORM, Alembic migrations, and scan/batch workflow.
  • CLI interface (Click) with scan, batch, status, and process commands.
  • Web upload UI with CORS support and enhanced error handling.
  • Azure Application Insights integration with custom dimensions for OCR metrics.
  • eSCL scanner integration plan for direct UI-triggered document scanning.
Improvements
  • Full document parsing extracts SDC ID, recipient, line items, and order details from packing slips.
  • SmartBatch ID generation with date-based naming convention and strict field validation.
  • Comprehensive OCR accuracy testing framework with benchmark reports.
Infrastructure
  • Docker Compose deployment with PostgreSQL 16 and PaddleOCR sidecar dependency.
  • 174 tests across 11 test files with pytest-xdist parallel execution.
  • SpecKit framework for spec-driven development with constitution and templates.
  • Structured logging with App Insights export and service name tagging.
Breaking Changes
  • OCR engine migrated from Tesseract to EasyOCR to PaddleOCR API during development.