MedITron/PTSRaiAgent
4
0
Fork 0
Code Issues 21 Pull Requests Actions Packages Projects Releases Wiki Activity
134 Commits 20 Branches 0 Tags
bfb0c9e7bbe37f3cb856163cd3e7c077df271ba4
Commit Graph
5 Commits
Author SHA1 Message Date
ElebobaandClaude Opus 5.5 e72378f178 fix(bot): agent polish — instant small talk, capped tool calls, analogues, plain text, definitions allowed
CI / ruff · pytest (pull_request) Successful in 2m43s
Details 
- Greetings / chit-chat answered without the LLM (instant, free, no medical findings dumped into «привет»).
- At most 4 tool calls per turn (one «как дела?» turn fired 23 parallel searches).
- «My drug» is never asked back; analogues → the drug's group, replacement only by the doctor.
- Markdown stripped to plain text with «•»; no sources footer under short replies.
- Diagnosis filter no longer blocks definitions («ПТСР — это расстройство…»); diagnosis-like statements about
  the patient are still blocked.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
 2026-10-04 22:43:18 +03:00
ElebobaandClaude Opus 5.5 0bbe1e0932 fix: web demo token tok-andrey shares Andrey's ЕМИАС record; REVISION written only after a healthy deploy
CI / ruff · pytest (pull_request) Successful in 2m44s
Details 
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
 2026-10-04 17:52:23 +03:00
ElebobaandClaude Opus 5.5 ade93d959b feat(bot): GigaChat function-calling provider for the chat agent; never 500 on provider errors
CI / ruff · pytest (pull_request) Successful in 2m15s
Details 
OpenRouter returns 403 «Access denied by security policy» from the RF server, and parsing that string error
crashed the request (500 instead of the fallback answer). Error parsing is robust now and any agent exception
degrades to the extractive answer. GigaChat tool provider (GigaChat-2-Pro by default) converts the agent's
OpenAI-style history to function_call / role=function; auto mode prefers GigaChat when credentials are set.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
 2026-10-04 15:46:21 +03:00
ElebobaandClaude Opus 5.5 9ec61b9a03 feat(bot): chat agent on Gemini 2.5 Flash-Lite — grounded first turn, current-prescription defaults
CI / ruff · pytest (pull_request) Successful in 2m5s
Details 
Live-tested on OpenRouter: Flash-Lite answers in 3–8 s with correct findings at ~$0.10/$0.40 per 1M tokens
(qwen3.7-flash and gpt-oss-120b took 26–32 s). Force a tool call on the first turn, give the model the
current prescription, let check_drug_for_patient default to it and return composition, intolerance check
and general interactions (alcohol); retry once on an empty final message.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
 2026-10-04 14:59:38 +03:00
ElebobaandClaude Opus 5.5 6dda5f9b30 feat(bot): tool-using chat agent over RAG, ЕМИАС record and drug-safety reference
CI / ruff · pytest (pull_request) Successful in 2m10s
Details 
Free-text treatment and medication questions go to an LLM agent with read-only tools bound to the current
patient: guideline search (RAG, doses stripped), ЕМИАС record mock (diagnoses, comorbidities, intolerances,
other specialists' drugs, current prescription), drug info (composition, contraindications, interactions)
and a deterministic patient-specific check. Providers: OpenRouter (temporary, synthetic data only) and an
offline mock; GigaChat to follow. Scrubbed input only, forbidden-output filter, any failure falls back to
the existing extractive RAG / instruction answers. Off unless OPENROUTER_API_KEY or CHAT_AGENT=mock.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
 2026-10-04 14:39:16 +03:00