woody
/
legco_ai_assistant
1
1
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
67 Commits 1 Branch 0 Tags 2.6 MiB
Commit Graph

5 Commits

Author SHA1 Message Date
Woody 51640201f3 test(backend): update query tests for sub-question generation (sub-phase 2.3) 
Update prompt assertion in decomposer test and field assertions in query endpoint tests to match extracted_questions rename.

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
 2026-04-24 16:24:10 +08:00
Woody 33b960f786 fix(backend): extract JSON from markdown code blocks in LLM responses 
The LLM (Qwen3.5 via OpenRouter) returns JSON wrapped in markdown code blocks:

```json

["project manager", "limits", ...]

```

But the code was trying to parse this directly with json.loads(), causing:

- QueryDecomposer to return empty keywords

- RelevanceFilter to fail with "Expecting value: line 1 column 1"

Changes:

- Added _extract_json_from_markdown() helper function to both modules

- Strips markdown code block markers (```json and ```) before JSON parsing

- Added unit tests for markdown code block handling

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)

Co-authored-by: Sisyphus \u003cclio-agent@sisyphuslabs.ai\u003e
 2026-04-23 16:28:07 +08:00
Woody be5e75e67c test(backend): update unit tests for LLM monitoring changes 
- Fixed MockLLMClient to accept step_name parameter

- Updated test mocks for OpenAI SDK structure

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
 2026-04-23 14:52:41 +08:00
Woody 351950f512 test(backend): update Phase 1 test suite 
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
 2026-04-23 13:27:40 +08:00
Woody 181f4eca5b feat: Phase 1.3 query pipeline with decomposition, relevance filter, and response 
- Add QueryDecomposer: extracts keywords from question via LLM JSON response
- Add RelevanceFilter: batch scores chunks 0-10, filters by threshold
- Add POST /api/v1/query endpoint with full 3-step pipeline:
  1. QueryDecomposer.decompose() → keywords
  2. RAGService.retrieve() → chunks from ChromaDB
  3. RelevanceFilter.filter() → score and filter chunks
  4. RAGService.generate_response() → bullet-point answer
- Fix SourceMetadata.upload_date type from datetime to str for flexibility
- Test-first: 13 new tests pass (5 decomposer, 5 relevance filter, 3 query endpoint)
- All Phase 1 tests: 41 passed, 2 skipped
 2026-04-22 17:19:21 +08:00