Carlos Garcia 774c0cc062 fix: tighten receipt amount extraction prompt to reduce OCR misreads ...

Replaced 'pick the largest one' guidance with 'bottom-most total' and
'return 0 if no clear total found' to avoid picking line items or tips.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

2026-05-16 16:47:48 -04:00

..

agents

fix: tighten receipt amount extraction prompt to reduce OCR misreads

2026-05-16 16:47:48 -04:00

llm

fix: force JSON output for Ollama intent classification; fix attachment detection

2026-05-16 01:17:58 -04:00

logging_utils

feat(infra): add sweep coordinator, structured logging, test suite, and README

2026-04-12 18:08:11 -04:00

mcp

feat(mcp): add MCP gateway — 14 tools over SSE, all agent calls forced local

2026-04-15 16:45:49 -04:00

memory

feat: OCR via tesseract, dedup, category selection for expense receipts

2026-05-16 01:40:32 -04:00

migrations

fix: MemoryManager kwarg 'llm' -> 'llm_router'; fix alembic script_location

2026-04-24 17:43:03 -04:00

prompts

feat(agent): direct-answer fallback for non-Odoo questions

2026-04-24 23:27:06 -04:00

routers

fix: widen actions_taken to list[Any] and improve bot error replies

2026-05-16 16:31:45 -04:00

tools

fix: improve OCR accuracy for rotated/sideways receipt photos

2026-05-16 01:51:29 -04:00

app_state.py

feat(service): add FastAPI agent service with 5 routers and Docker setup

2026-04-12 17:54:28 -04:00

config.py

Switch default Ollama model to activeblue-chat (fine-tuned Llama 3.1 8B)

2026-05-14 13:31:22 +00:00

main.py

feat: file upload + expense report creation from Discuss attachments

2026-05-16 01:02:24 -04:00