vault backup: 2026-05-11 17:49:28

This commit is contained in:
Vadym Samoilenko 2026-05-11 17:49:28 +01:00
parent ffc675b23b
commit 08783253e5
2 changed files with 4 additions and 0 deletions

View file

@ -78,6 +78,7 @@ Standalone Python CLI tool that benchmarks OpenAI Assistants against RAG knowled
## Change Log
| Date | Requested | Changed | Files |
|------|-----------|---------|-------|
| 2026-05-11 | Report generation | Executed evaluations for Internal Banners, Social Posts, Display Banners, PPC | rag_test_app/results (4 report.html files) |
| 2026-05-11 | Report generation | Eliminated duplicate evaluation calls, cached results in self.evaluation | rag_test_app/evaluate.py, report_generator.py |
| 2026-05-11 | Report regeneration | Config paths updated, evaluation caching optimized | evaluation.py, config.yaml |
| 2026-05-11 | Report restart with Excel questions | Reorganized question files structure, updated config paths | config files, question directory structure |

View file

@ -605,3 +605,6 @@ tags: [daily]
- 17:28 | `barclays-rag-report`
- **Asked:** Run evaluation reports using questions from the Excel file.
- **Done:** Generated 4 Barclays reports (Internal Banners, Social Posts, Display Banners, PPC) with evaluation results and optimized caching.
- 17:48 | `barclays-rag-report`
- **Asked:** Restart evaluation reports using questions from the Excel file (Eval Questions Barclays May 2026.xlsx)
- **Done:** Generated fresh reports for four marketing categories with significantly improved quality, accuracy, and completeness scores