Merge pull request 'feat(bot): ответы на вопросы о лечении из базы знаний КР (#64)' (#65) from feat/64-rag-dialog into dev

app/dialog/flows/core.py

@@ -39,6 +39,7 @@ def menu_reply(ctx: Ctx, text: str | None = None) -> Reply:
     buttons = [
         [Button(label=b.checkin, data="menu:checkin")],
         [Button(label=b.med_question, data="menu:med_question")],
+        [Button(label=b.kb_question, data="menu:kb_question")],
         [Button(label=b.contact_doctor, data="menu:contact_doctor")],
         [Button(label=b.revoke, data="consent:revoke")],
     ]

app/dialog/flows/knowledge.py

@@ -0,0 +1,102 @@
+"""Questions about the condition and its treatment → knowledge base of clinical guidelines (#64, RAG #60).
+
+    menu «Вопрос о лечении» ──▶ kb.question (free text) ──▶ rag.answer: KR quote + source, or «уточните у врача»
+    free text in idle that looks like a question (veteran_daily `text:idle`, after the check-in signals and the
+        drug question) ──▶ rag.answer, or the usual support phrase when the KB has no answer
+
+The engine has already scrubbed the text, checked consent and run the safety check (ACUTE never gets here).
+Questions about the patient's own drug stay with `med_qa` (instruction only, SPEC §7 п.5); this flow answers
+general questions from the guidelines, extractively and without doses (`app.llm.rag.qa`). Logs carry codes only.
+"""
+
+import logging
+import re
+from functools import lru_cache
+from pathlib import Path
+
+import yaml
+from pydantic import BaseModel, ConfigDict
+
+from app.config import get_settings
+from app.dialog.contracts import Event, Reply
+from app.dialog.flows.core import menu_reply
+from app.dialog.fsm import Ctx, Step, abandonable, handler_for, state, trigger
+from app.llm import rag
+from app.llm.med_qa import DOSE_QUESTION
+
+log = logging.getLogger(__name__)
+
+QUESTION_STATE = "kb.question"
+# `scrub()` placeholders and the punctuation around them: "[NAME], что такое ПТСР?" starts with "что такое".
+_PLACEHOLDERS = re.compile(r"^(?:\s*\[[A-Z_]+\][\s,.:;!-]*)+")
+
+
+class KnowledgeTexts(BaseModel):
+    model_config = ConfigDict(extra="forbid", frozen=True)
+
+    prompt: str
+    dose_redirect: str
+    question_starts: list[str]
+    small_talk: list[str]
+
+
+@lru_cache
+def load_knowledge_texts(path: Path | None = None) -> KnowledgeTexts:
+    path = path or get_settings().config_dir / "texts" / "knowledge.yaml"
+    return KnowledgeTexts.model_validate(yaml.safe_load(path.read_text(encoding="utf-8")))
+
+
+@lru_cache
+def _question_start() -> re.Pattern[str]:
+    starts = sorted(load_knowledge_texts().question_starts, key=len, reverse=True)
+    return re.compile(r"^\s*(?:" + "|".join(re.escape(s) for s in starts) + r")\b", re.IGNORECASE)
+
+
+@lru_cache
+def _small_talk() -> re.Pattern[str]:
+    return re.compile(r"\b(?:" + "|".join(re.escape(s) for s in load_knowledge_texts().small_talk) + r")\b", re.I)
+
+
+def looks_like_question(text: str) -> bool:
+    """A "?" anywhere or a question word at the start ("что такое флэшбек", "можно ли ..."); small talk
+    ("как дела?") is not a question for the knowledge base."""
+    text = _PLACEHOLDERS.sub("", text)
+    if _small_talk().search(text):
+        return False
+    return "?" in text or bool(_question_start().search(text))
+
+
+async def knowledge_answer(ctx: Ctx, *, explicit: bool) -> str | None:
+    """KB answer for `ctx.text`. No answer → «уточните у врача» if the patient asked via the menu (`explicit`),
+    `None` otherwise (the caller keeps its own reply, e.g. the support phrase)."""
+    if ctx.text is None:
+        return None
+    if DOSE_QUESTION.search(ctx.text.lower()):
+        log.info("kb: dose question redirected to doctor")
+        return load_knowledge_texts().dose_redirect
+    reply = await rag.answer(ctx.text, audience="patient")
+    if reply is not None:
+        log.info("kb: answered docs=%s", ",".join(c.doc for c in reply.citations))
+        return reply.text
+    log.info("kb: no answer explicit=%s", explicit)
+    return rag.fallback_text("patient") if explicit else None
+
+
+@trigger("button:menu:kb_question")
+async def _menu_question(ctx: Ctx, event: Event) -> Step:
+    return Step(replies=[Reply(text=load_knowledge_texts().prompt)], next_state=QUESTION_STATE)
+
+
+@state(QUESTION_STATE)
+async def _question(ctx: Ctx, event: Event) -> Step:
+    if event.type != "text" or ctx.text is None:
+        # A menu button or a reminder instead of the question: drop the question and let idle handle it.
+        step = Step(next_state="idle")
+        idle = handler_for("idle")
+        return step.then(await idle(ctx, event)) if idle is not None else step
+    text = await knowledge_answer(ctx, explicit=True)
+    return Step(replies=[menu_reply(ctx, text)], next_state="idle")
+
+
+# A question asked yesterday must not swallow today's message.
+abandonable("kb")

app/dialog/flows/veteran_daily.py

@@ -47,6 +47,7 @@ from app.db import repo
 from app.dialog.contracts import Button, Event, Reply
 from app.dialog.flows import intake
 from app.dialog.flows.core import menu_reply
+from app.dialog.flows.knowledge import knowledge_answer, looks_like_question
 from app.dialog.flows.meds import (
     answer_question,
     checkin_of,
@@ -423,11 +424,13 @@ async def _visit(ctx: Ctx, event: Event) -> Step:
 @trigger("text:idle", replace=True)
 async def _idle_text(ctx: Ctx, event: Event) -> Step:
     """Free text outside the check-in: drug question → instruction excerpt; side effects / quit intent / alcohol
-    → today's check-in (the rules react); otherwise the neutral support phrase with the menu."""
+    → today's check-in (the rules react); another question → clinical guidelines (knowledge base); otherwise
+    the neutral support phrase with the menu."""
     assert ctx.text is not None
     facts = await ctx.gateway.extract_checkin(ctx.text)
     if facts is None:
-        return Step(replies=[menu_reply(ctx, await ctx.gateway.phrase("support"))])
+        text = await knowledge_answer(ctx, explicit=False) if looks_like_question(ctx.text) else None
+        return Step(replies=[menu_reply(ctx, text or await ctx.gateway.phrase("support"))])
     fields: dict[str, Any] = {}
     if facts.quit_intent:
         fields["quit_intent"] = True
@@ -445,6 +448,8 @@ async def _idle_text(ctx: Ctx, event: Event) -> Step:
         await save_checkin(ctx, **fields)
     if text is None and facts.med_question and not facts.quit_intent:
         text = await answer_question(ctx, await current_prescription(ctx))
+    if text is None and not fields and looks_like_question(ctx.text):
+        text = await knowledge_answer(ctx, explicit=False)  # general question → clinical guidelines (#64)
     if text is None:
         text = await ctx.gateway.phrase("support")
     return Step(replies=[menu_reply(ctx, text)], notify_monitoring=bool(fields or facts.side_effects))

app/dialog/texts.py

@@ -47,6 +47,7 @@ class AgeGateTexts(_Strict):
 class IdleButtons(_Strict):
     checkin: str
     med_question: str
+    kb_question: str
     contact_doctor: str
     revoke: str
 

app/llm/rag/qa.py

@@ -56,6 +56,7 @@ PRESCRIBING = re.compile(
 )
 # "что такое флэшбек" -> the term; its definition "Флэшбек (flashback – англ.) – ..." is preferred.
 DEFINITION_QUESTION = re.compile(r"^\s*что (?:такое|значит|означает|это за)\s+(.+?)[?.!\s]*$", re.IGNORECASE)
+LEADING_PLACEHOLDERS = re.compile(r"^(?:\s*\[[A-Z_]+\][\s,.:;!-]*)+")  # "[NAME], что такое ПТСР?"
 DEFINITION_HEAD = re.compile(r"\s[–—-]\s(?![^(]*\))|\s[–—-]?\s?это\s")  # dash outside "(… – англ.)", or "это"
 DEFINITION_HEAD_CHARS = 60
 DEFINITION_BONUS = 2.0  # multiplier for a sentence that starts with the asked term (picks it within a chunk)
@@ -174,7 +175,7 @@ def _allowed(sentence: str | None, audience: Audience) -> bool:
 
 
 def _defined_term(question: str) -> set[str] | None:
-    m = DEFINITION_QUESTION.match(question)
+    m = DEFINITION_QUESTION.match(LEADING_PLACEHOLDERS.sub("", question))
     return set(tokens(m.group(1))) or None if m else None
 
 

config/texts/dialog.yaml

@@ -63,6 +63,7 @@ idle:
   buttons:
     checkin: "Как я сегодня"
     med_question: "Вопрос о лекарстве"
+    kb_question: "Вопрос о лечении"
     contact_doctor: "Связаться с врачом"
     revoke: "Отозвать согласие"
   # Заглушки кнопок меню; ветки ветерана (#36/#37) и med_qa (#15) подменяют их своими триггерами.

config/texts/knowledge.yaml

@@ -0,0 +1,51 @@
+# Вопросы пациента о состоянии и лечении → база знаний по клиническим рекомендациям (#64, RAG #60).
+# Зона clinical вычитывает формулировки. Сам ответ — только цитата из КР с источником (config/texts/rag.yaml
+# задаёт его обрамление); здесь — тексты диалога и признаки того, что свободное сообщение — вопрос.
+
+# Кнопка меню «Вопрос о лечении» → бот просит написать вопрос.
+prompt: |-
+  Напишите ваш вопрос о состоянии или лечении — например, «что такое флэшбек?» или «помогает ли психотерапия?».
+  Я отвечу цитатой из клинических рекомендаций Минздрава и укажу источник. Вопросы о дозе и о вашем препарате решает врач.
+
+# Вопрос о дозе (увеличить, уменьшить, полтаблетки): бот не отвечает по существу — это решение врача.
+dose_redirect: "Дозу и схему приёма подбирает только ваш врач — менять их самостоятельно нельзя. Могу передать ваш вопрос врачу: нажмите «Связаться с врачом»."
+
+# Свободное сообщение считается вопросом, если в нём есть «?» или оно начинается с одного из этих слов.
+# Тогда бот ищет ответ в базе знаний; если ответа нет — отвечает обычной фразой поддержки.
+question_starts:
+  - "что такое"
+  - "что значит"
+  - "что означает"
+  - "почему"
+  - "зачем"
+  - "как"
+  - "когда"
+  - "сколько"
+  - "можно ли"
+  - "нужно ли"
+  - "надо ли"
+  - "помогает ли"
+  - "поможет ли"
+  - "правда ли"
+  - "бывает ли"
+  - "что делать"
+  - "что будет"
+  - "чем"
+  - "где"
+  - "есть ли"
+
+# Бытовые фразы — не вопросы о лечении, даже со «?»: «как дела?», «как ты?». База знаний для них не ищется
+# (иначе «как дела» находит «линия делит» — у слов одна основа), бот отвечает обычной фразой поддержки.
+small_talk:
+  - "как дела"
+  - "как ты"
+  - "как вы"
+  - "как жизнь"
+  - "как поживаете"
+  - "что нового"
+  - "кто ты"
+  - "ты кто"
+  - "ты бот"
+  - "ты робот"
+  - "привет"
+  - "спасибо"

tests/test_dialog_knowledge.py

@@ -0,0 +1,188 @@
+"""Knowledge base in the dialog (#64): «Вопрос о лечении» menu, questions in idle free text, routing order
+(safety → check-in signals → drug question → guidelines → support)."""
+
+from collections.abc import AsyncIterator
+from datetime import datetime
+from typing import Any
+
+import pytest
+
+from app.config import Settings
+from app.core import clock
+from app.db import repo, session
+from app.db.bootstrap import init_db
+from app.db.clinical import Prescription
+from app.dialog.contracts import Event, Reply
+from app.dialog.engine import Engine
+from app.dialog.flows.knowledge import QUESTION_STATE, load_knowledge_texts, looks_like_question
+from app.integrations.base import DrugSection, integrations
+from app.llm import med_qa as med_qa_module
+from app.llm.gateway import LLMGateway
+from app.llm.med_qa import load_med_qa_texts
+from app.llm.mock import MockTasks
+from app.llm.rag import fallback_text
+from app.llm.rag import qa as rag_qa
+from app.llm.texts import render_phrase
+from app.services.contracts import NoopMonitoring
+
+TOKEN = "k" * 32
+MEM = "sqlite+aiosqlite:///:memory:"
+NOW = datetime(2026, 10, 7, 8, 0)
+SUPPORT = render_phrase("support", {})
+KB = load_knowledge_texts()
+
+
+class FakeDrugInfo:
+    async def sections(self, inn: str) -> list[DrugSection]:
+        text = "Улучшение обычно отмечается через 2–4 недели регулярного приёма."
+        return [DrugSection(inn=inn, section="onset", text=text, source="синтетика")]
+
+
+@pytest.fixture
+async def engine(monkeypatch: pytest.MonkeyPatch) -> AsyncIterator[Engine]:
+    monkeypatch.setattr(clock, "get_settings", lambda: Settings(_env_file=None, demo_clock=True))
+    monkeypatch.setattr(integrations, "drug_info", FakeDrugInfo())
+    monkeypatch.setattr(med_qa_module, "get_gateway", lambda: LLMGateway(MockTasks()))
+    monkeypatch.setattr(rag_qa, "get_gateway", lambda: LLMGateway(MockTasks()))
+    clock.set_now(NOW)
+    await session.configure(identity_url=MEM, clinical_url=MEM)
+    await init_db()
+    async with session.get_clinical_session() as s:
+        await repo.upsert_profile(
+            s, token=TOKEN, doctor_id="doc", modules=["veteran"], sex="m", birth_year=1990, therapy_start=NOW.date()
+        )
+        s.add(Prescription(token=TOKEN, drug="сертралин", dose="50 мг", schedule={"times": ["08:00"]}))
+        await repo.set_dialog_state(s, token=TOKEN, state="idle", data={"consent": True}, now=clock.now())
+    yield Engine(gateway=LLMGateway(MockTasks()), monitoring=NoopMonitoring())
+    clock.reset()
+    await session.configure()
+
+
+async def press(e: Engine, data: str) -> list[Reply]:
+    return await e.handle(TOKEN, Event(type="button", button=data))
+
+
+async def say(e: Engine, t: str) -> list[Reply]:
+    return await e.handle(TOKEN, Event(type="text", text=t))
+
+
+async def state() -> tuple[str, dict[str, Any]]:
+    async with session.get_clinical_session() as s:
+        row = await repo.get_dialog_state(s, TOKEN)
+        assert row is not None
+        return row.state, dict(row.data)
+
+
+def buttons(replies: list[Reply]) -> list[str]:
+    return [b.data for r in replies for row in r.buttons for b in row]
+
+
+def joined(replies: list[Reply]) -> str:
+    return "\n".join(r.text for r in replies)
+
+
+# --- question detection --------------------------------------------------------------------------------
+
+
+@pytest.mark.parametrize(
+    ("text", "expected"),
+    [
+        ("что такое флэшбек", True),
+        ("Можно ли вылечить ПТСР", True),
+        ("кошмары снятся, это нормально?", True),
+        ("сегодня 6 из 10", False),
+        ("спал плохо", False),
+        ("какой-то день", False),  # "как" only as a whole word at the start
+        ("как дела?", False),  # small talk
+        ("Привет! Как ты?", False),
+        ("[NAME], что такое ПТСР", True),  # scrub placeholder before the question
+    ],
+)
+def test_looks_like_question(text: str, expected: bool) -> None:
+    assert looks_like_question(text) is expected
+
+
+# --- menu «Вопрос о лечении» ------------------------------------------------------------------------------
+
+
+async def test_menu_has_knowledge_button(engine: Engine) -> None:
+    r = await engine.handle(TOKEN, Event(type="start"))
+    assert "menu:kb_question" in buttons(r)
+
+
+async def test_menu_question_answers_from_guidelines(engine: Engine) -> None:
+    r = await press(engine, "menu:kb_question")
+    assert joined(r) == KB.prompt
+    assert (await state())[0] == QUESTION_STATE
+
+    r = await say(engine, "что такое флэшбек?")
+    text = joined(r)
+    assert "Флэшбек (flashback" in text
+    assert "Источник: КР ПТСР 2023, стр. 2" in text
+    assert rag_qa.load_rag_texts()["patient"].doctor_note in text
+    assert "menu:kb_question" in buttons(r)  # back to the menu
+    assert (await state())[0] == "idle"
+
+
+async def test_menu_question_without_answer_sends_to_doctor(engine: Engine) -> None:
+    await press(engine, "menu:kb_question")
+    r = await say(engine, "какой рецепт борща")
+    assert fallback_text("patient") in joined(r)
+    assert (await state())[0] == "idle"
+
+
+@pytest.mark.parametrize("question", ["можно увеличить дозу?", "можно пить полтаблетки?"])
+async def test_dose_question_redirected_to_doctor(engine: Engine, question: str) -> None:
+    await press(engine, "menu:kb_question")
+    r = await say(engine, question)
+    assert joined(r).startswith(KB.dose_redirect)
+    assert "клинических рекомендациях" not in joined(r)
+
+
+async def test_button_instead_of_question_leaves_the_flow(engine: Engine) -> None:
+    await press(engine, "menu:kb_question")
+    r = await press(engine, "menu:contact_doctor")
+    assert "врач" in joined(r)
+    assert (await state())[0] == "idle"
+
+
+async def test_scrubbed_question_still_answered(engine: Engine) -> None:
+    await press(engine, "menu:kb_question")
+    r = await say(engine, "Иван Петров, 89161234567, что такое ПТСР?")
+    assert "ПТСР – это расстройство" in joined(r)
+
+
+# --- free text in idle ------------------------------------------------------------------------------------
+
+
+async def test_free_text_question_answered_from_guidelines(engine: Engine) -> None:
+    r = await say(engine, "помогает ли психотерапия при ПТСР?")
+    text = joined(r)
+    assert "В клинических рекомендациях Минздрава сказано:" in text
+    assert "Источник: КР ПТСР 2023" in text
+
+
+@pytest.mark.parametrize("message", ["сегодня обычный день", "как дела?"])
+async def test_free_text_without_answer_keeps_support(engine: Engine, message: str) -> None:
+    r = await say(engine, message)
+    assert joined(r) == SUPPORT
+    assert "Источник" not in joined(r)
+
+
+async def test_drug_question_stays_with_instruction(engine: Engine) -> None:
+    # SPEC §7 п.5: the patient's own drug is answered only from its instruction, never from the guidelines.
+    r = await say(engine, "через сколько подействуют таблетки?")
+    text = joined(r)
+    assert text.startswith(load_med_qa_texts().intro)
+    assert "клинических рекомендациях" not in text
+
+
+async def test_side_effect_question_goes_to_checkin(engine: Engine) -> None:
+    r = await say(engine, "меня тошнит от таблеток, это нормально?")
+    assert "клинических рекомендациях" not in joined(r)
+
+
+async def test_acute_risk_never_reaches_knowledge_base(engine: Engine) -> None:
+    r = await say(engine, "зачем жить? хочу умереть")
+    assert any(x.kind == "crisis" for x in r)
+    assert "клинических рекомендациях" not in joined(r)