test: Add comprehensive test suite for GPU embedder validation

Browse Source

Test Scripts Added:
- test_gpu_mistral.py: Ingestion test with Mistral LLM (9 chunks in 1.2s)
- test_search_simple.js: Puppeteer search test (16 results found)
- test_chat_puppeteer.js: Puppeteer chat test (11 chunks, 5 sections)
- test_memories_conversations.js: Memories & conversations UI test

Test Results:
✅ Ingestion: GPU vectorization works (30-70x faster than Docker)
✅ Search: Semantic search functional with GPU embedder
✅ Chat: RAG chat with hierarchical search working
✅ Memories: API backend functional (10 results)
✅ Conversations: UI and search working

Screenshots Added:
- chat_page.png, chat_before_send.png, chat_response.png
- search_page.png, search_results.png
- memories_page.png, memories_search_results.png
- conversations_page.png, conversations_search_results.png

All tests validate the GPU embedder migration is production-ready.
GPU: NVIDIA RTX 4070, VRAM: 2.6 GB, Model: BAAI/bge-m3 (1024 dims)

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

...

This commit is contained in:

David Blanc Brioir

2026-01-09 11:51:01 +01:00

parent 17dfe213ed

commit 7340ce5540

13 changed files with 666 additions and 0 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

BIN

memories_search_results.png Normal file

Binary file not shown.

Side by Side After Width:  |  Height:  |  Size: 81 KiB