Oliver/ai_qc
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
212 commits 29 branches 11 tags 50 MiB
Commit graph

3 commits

Author SHA1 Message Date
nickviljoen
9eed569587 Tone down OCR from authoritative to supplementary to reduce false positives 
OCR measurements were causing the LLM to over-rely on bounding box numbers
and fail correct assets on minor measurement inaccuracies. Changes:
- All prompts now say "supplementary data" not "authoritative/primary source"
- LLM instructed to prioritise visual assessment, use OCR to confirm/question
- Alignment tolerance widened from 1.5% to 3% of width
- OCR context footer softened with accuracy caveat (~5-10px margin of error)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-04-02 13:54:00 +02:00
nickviljoen
9f9777240a Add OCR layout measurement module for precise spatial QC checks 
Adds Tesseract-based OCR pre-processing that computes pixel-level text
positions, margins, spacing, and alignment before LLM analysis. This
enables detection of subtle layout differences that vision models miss
(e.g. 2.8% vs 6.4% headline margin, 83px vs 39px date gap).

OCR measurements injected into 10 checks across all client profiles:
- Amazon: margins, typography, headline_layout
- Static General: element_alignment, safety_area, visual_hierarchy_general,
  text_readability_general, text_edge_clearance
- L'Oreal: text_readability
- Diageo/Unilever KV: visual_hierarchy

Non-blocking: if Tesseract is unavailable, checks run with visual
estimation only. Production requires: sudo apt install tesseract-ocr

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-04-02 11:00:07 +02:00
nickviljoen
3fec052c12 Create frontend and backend folder structure for deployment 
Organized the application into separate frontend and backend directories for cleaner deployment and better separation of concerns.

Frontend Directory (frontend/):
- index.html: Single-page web interface (renamed from web_ui.html)
- README.md: Frontend deployment guide
- Total size: ~113 KB (self-contained)
- Smart base path detection (works at / or /ai_qc/)
- No configuration changes required

Backend Directory (backend/):
- All Python files (api_server.py, llm_config.py, etc.)
- visual_qc_apps/: 33 QC check modules
- profiles/: 6 QC profile configurations
- brand_guidelines/: Reference asset storage
- config/: Environment configurations
- scripts/: Deployment automation
- uploads/, output/: Data directories
- requirements.txt, ai_qc.service, apache_config.conf
- Complete documentation

New Documentation:
- FOLDER_STRUCTURE.md: Comprehensive guide to new structure
- frontend/README.md: Frontend deployment instructions
- backend/BACKEND_README.md: Backend deployment guide

Deployment Mapping:
- frontend/ → /var/www/html/ai_qc/ (web root)
- backend/ → /opt/ai_qc/ (application directory)

Benefits:
- Clear separation of concerns
- Backend code not in web-accessible directory
- Independent frontend/backend updates
- Matches server's existing patterns (/opt/veo3, /opt/voice2text)
- Industry-standard architecture
- Easy to deploy and maintain

Original files preserved in root directory for reference.
Ready for production deployment following MIGRATION_GUIDE.md.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
 2025-11-06 11:55:53 +02:00