Viktor Barzin 25c948c933 examples: disable qwen3 thinking output in llama-swap requests ... qwen3-8b emits <think>...</think> chain-of-thought before the JSON response by default, which trips the JSON parser and forces ~50% of posts to escalate to claude-agent-service (discovered during first bulk ingest, 2026-06-05). Fix: pass chat_template_kwargs.enable_thinking=false in the request body for the Tier 1 (llama-swap) call. Claude calls are unaffected. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>		2026-06-05 08:40:25 +00:00
..
__init__.py	examples: RawPost + ExtractedExample + Summary Pydantic schemas	2026-05-28 22:12:01 +00:00
cli.py	examples: serialize LLM calls via Semaphore (default 1) to dodge CAS busy-lock	2026-06-02 13:54:15 +00:00
filters.py	examples: regex pre-filter (MONEY_RE + LOCATION_RE)	2026-05-28 22:14:59 +00:00
llm_extract.py	examples: disable qwen3 thinking output in llama-swap requests	2026-06-05 08:40:25 +00:00
models.py	examples: RawPost + ExtractedExample + Summary Pydantic schemas	2026-05-28 22:12:01 +00:00
praw_source.py	examples: async PRAW wrapper → RawPost	2026-05-28 22:16:53 +00:00
service.py	examples: service.upsert_example + summary_for_country	2026-05-28 22:28:53 +00:00