Files
PTSRaiAgent/app/llm/agent/providers.py
T
ElebobaandClaude Opus 5.5 d1f92407f9
CI / ruff · pytest (pull_request) Successful in 2m43s
feat: live agent answers (SSE streaming) + «Чат» tab on phones
- POST /dialog/{token}/events/stream: Server-Sent Events with the agent's steps («Смотрю вашу карту ЕМИАС…»,
  «Ищу в клинических рекомендациях…»), the answer as the model writes it (OpenRouter stream=true), then the
  final replies (same as /events, incl. the forbidden-output filter and sources). X-Accel-Buffering: no.
- Web: free-text messages use the stream — a live bubble with steps, typing dots and a cursor; falls back to
  the plain request if the stream breaks.
- Phones: the bottom tab bar gets «Чат» (it had no way back to the chat).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-05 12:52:19 +03:00

345 lines
16 KiB
Python

"""Tool-calling chat providers for the agent.
* `OpenRouterProvider` — OpenAI-compatible `/chat/completions` with `tools`. TEMPORARY for the prototype on
synthetic data: OpenRouter is a foreign service, forbidden for real PII (152-FZ art. 18 p. 5). Replace with
GigaChat (function calling) before any real patient — same `ToolChatProvider` contract.
* `MockAgentProvider` — offline, deterministic: plans tool calls from keywords and composes an answer from the
tool results. Keeps the demo working without network (ADR-004).
"""
import json
import logging
import re
from typing import Any, Protocol
import httpx
from pydantic import BaseModel, Field
from app.config import Settings
from app.llm.agent import stream
log = logging.getLogger(__name__)
class ToolCall(BaseModel):
id: str
name: str
arguments: dict[str, Any] = Field(default_factory=dict)
class AssistantTurn(BaseModel):
content: str | None = None
tool_calls: list[ToolCall] = Field(default_factory=list)
class ToolChatProvider(Protocol):
name: str
async def chat(self, messages: list[dict[str, Any]], tools: list[dict[str, Any]]) -> AssistantTurn: ...
class ProviderError(RuntimeError):
pass
class OpenRouterProvider:
name = "openrouter"
def __init__(self, settings: Settings, client: httpx.AsyncClient | None = None) -> None:
self._key = settings.openrouter_api_key
self._model = settings.openrouter_model
self._url = settings.openrouter_base_url.rstrip("/") + "/chat/completions"
self._timeout = settings.chat_agent_timeout_s
self._proxy = settings.openrouter_proxy_url or None
self._client = client
async def chat(self, messages: list[dict[str, Any]], tools: list[dict[str, Any]]) -> AssistantTurn:
if stream.active():
return await self._chat_stream(messages, tools)
body: dict[str, Any] = {"model": self._model, "messages": messages, "tools": tools, "temperature": 0.1}
body["max_tokens"] = 900
if not any(m["role"] == "tool" for m in messages):
body["tool_choice"] = "required" # never answer from the model's own memory: ground first
headers = {"Authorization": f"Bearer {self._key}", "X-Title": "PTSRaiAgent prototype (synthetic data)"}
try:
if self._client is not None:
resp = await self._client.post(self._url, json=body, headers=headers, timeout=self._timeout)
else:
async with httpx.AsyncClient(timeout=self._timeout, proxy=self._proxy) as client:
resp = await client.post(self._url, json=body, headers=headers)
except httpx.HTTPError as exc:
raise ProviderError(type(exc).__name__) from exc
if resp.status_code != 200:
# Provider error bodies are generic (no patient text) — log the message to make failures diagnosable.
try:
err = resp.json().get("error", "")
detail = str(err.get("message", "") if isinstance(err, dict) else err)[:200]
except (ValueError, AttributeError):
detail = resp.text[:200]
log.error("openrouter http %s: %s", resp.status_code, detail)
raise ProviderError(f"http {resp.status_code}")
try:
msg = resp.json()["choices"][0]["message"]
except (ValueError, KeyError, IndexError) as exc:
raise ProviderError("bad response") from exc
calls = []
for c in msg.get("tool_calls") or []:
fn = c.get("function") or {}
try:
args = json.loads(fn.get("arguments") or "{}")
except json.JSONDecodeError:
args = {}
calls.append(ToolCall(id=c.get("id") or fn.get("name", "call"), name=fn.get("name", ""), arguments=args))
usage = resp.json().get("usage") or {}
log.info("openrouter ok calls=%d tokens=%s", len(calls), usage.get("total_tokens"))
return AssistantTurn(content=msg.get("content"), tool_calls=calls)
class GigaChatToolProvider:
"""GigaChat function calling (RF provider, 152-FZ compliant). Converts the OpenAI-style history the agent keeps:
assistant `tool_calls` → `function_call`, `tool` results → role `function` with the function name.
GigaChat returns at most one function call per turn — the agent loop handles that naturally."""
name = "gigachat"
def __init__(self, settings: Settings, client: Any | None = None) -> None:
self._settings = settings
self._client = client
def _get_client(self) -> Any:
if self._client is None:
from gigachat import GigaChat
s = self._settings
self._client = GigaChat(
credentials=s.gigachat_credentials,
scope=s.gigachat_scope,
model=s.gigachat_agent_model,
verify_ssl_certs=s.gigachat_verify_ssl,
timeout=s.chat_agent_timeout_s,
)
return self._client
@staticmethod
def to_messages(messages: list[dict[str, Any]]) -> list[Any]:
from gigachat.models import FunctionCall, Messages, MessagesRole
out = []
for m in messages:
role = m["role"]
if role == "system":
out.append(Messages(role=MessagesRole.SYSTEM, content=m["content"]))
elif role == "user":
out.append(Messages(role=MessagesRole.USER, content=m["content"]))
elif role == "assistant":
calls = m.get("tool_calls") or []
for c in calls or [None]:
fc = None
if c is not None:
fc = FunctionCall(name=c["function"]["name"], arguments=json.loads(c["function"]["arguments"]))
out.append(Messages(role=MessagesRole.ASSISTANT, content=m.get("content") or "", function_call=fc))
elif role == "tool":
out.append(Messages(role=MessagesRole.FUNCTION, name=m.get("name"), content=m["content"]))
return out
@staticmethod
def to_functions(tools: list[dict[str, Any]]) -> list[Any]:
from gigachat.models import Function
return [Function.model_validate(t["function"]) for t in tools]
async def chat(self, messages: list[dict[str, Any]], tools: list[dict[str, Any]]) -> AssistantTurn:
from gigachat.models import Chat
chat = Chat(
messages=self.to_messages(messages),
functions=self.to_functions(tools),
function_call="auto",
temperature=0.1,
max_tokens=900,
)
try:
resp = await self._get_client().achat(chat)
except Exception as exc: # noqa: BLE001 — network/auth/HTTP errors of the SDK
raise ProviderError(type(exc).__name__) from exc
msg = resp.choices[0].message
calls = []
if msg.function_call is not None:
args = msg.function_call.arguments or {}
if isinstance(args, str):
try:
args = json.loads(args)
except json.JSONDecodeError:
args = {}
calls.append(ToolCall(id=msg.function_call.name, name=msg.function_call.name, arguments=args))
usage = getattr(resp, "usage", None)
log.info("gigachat agent ok calls=%d tokens=%s", len(calls), getattr(usage, "total_tokens", None))
return AssistantTurn(content=msg.content or None, tool_calls=calls)
# ------------------------------------------------------------------------------------------------ streaming
async def _iter_sse(resp: httpx.Response):
async for line in resp.aiter_lines():
if not line.startswith("data:"):
continue # comments («: OPENROUTER PROCESSING») and blank lines
data = line[5:].strip()
if data == "[DONE]":
return
try:
yield json.loads(data)
except json.JSONDecodeError:
continue
async def _openrouter_stream(self: "OpenRouterProvider", messages: list[dict[str, Any]], tools: list[dict[str, Any]]):
"""Same request with `stream: true`: content deltas go to the live channel, tool calls are assembled."""
body: dict[str, Any] = {
"model": self._model,
"messages": messages,
"tools": tools,
"temperature": 0.1,
"max_tokens": 900,
"stream": True,
}
if not any(m["role"] == "tool" for m in messages):
body["tool_choice"] = "required"
headers = {"Authorization": f"Bearer {self._key}", "X-Title": "PTSRaiAgent prototype (synthetic data)"}
content: list[str] = []
calls: dict[int, dict[str, Any]] = {}
try:
client = self._client or httpx.AsyncClient(timeout=self._timeout, proxy=self._proxy)
try:
async with client.stream("POST", self._url, json=body, headers=headers, timeout=self._timeout) as resp:
if resp.status_code != 200:
await resp.aread()
raise ProviderError(f"http {resp.status_code}")
async for chunk in _iter_sse(resp):
choice = (chunk.get("choices") or [{}])[0]
delta = choice.get("delta") or {}
if delta.get("content"):
content.append(delta["content"])
stream.emit("delta", delta["content"])
for tc in delta.get("tool_calls") or []:
slot = calls.setdefault(tc.get("index", 0), {"id": "", "name": "", "args": ""})
slot["id"] = tc.get("id") or slot["id"]
fn = tc.get("function") or {}
slot["name"] += fn.get("name") or ""
slot["args"] += fn.get("arguments") or ""
finally:
if self._client is None:
await client.aclose()
except httpx.HTTPError as exc:
raise ProviderError(type(exc).__name__) from exc
out = []
for i in sorted(calls):
c = calls[i]
try:
args = json.loads(c["args"] or "{}")
except json.JSONDecodeError:
args = {}
out.append(ToolCall(id=c["id"] or c["name"] or f"call{i}", name=c["name"], arguments=args))
if out and content:
stream.emit("reset", None) # a preface before tool calls is not the answer
log.info("openrouter stream ok calls=%d chars=%d", len(out), sum(map(len, content)))
return AssistantTurn(content="".join(content) or None, tool_calls=out)
OpenRouterProvider._chat_stream = _openrouter_stream # type: ignore[attr-defined]
# ------------------------------------------------------------------------------------------------ offline mock
_DRUG_WORDS = re.compile(
r"\b(сертралин\w*|пароксетин\w*|эсциталопрам\w*|флуоксетин\w*|ибупрофен\w*|нурофен\w*|золофт\w*)", re.I
)
_MED_HINT = re.compile(r"(таблет|лекарств|препарат|состав|можно ли|совмест|вместе|противопоказ|аллерг|лактоз)", re.I)
class MockAgentProvider:
"""Plans: record → (drug check for each mentioned drug, or the current prescription for medicine questions)
→ guidelines search; then writes a short answer strictly from the tool results."""
name = "mock"
async def chat(self, messages: list[dict[str, Any]], tools: list[dict[str, Any]]) -> AssistantTurn:
question = next(m["content"] for m in reversed(messages) if m["role"] == "user")
done = [m for m in messages if m["role"] == "tool"]
if not done:
drugs = [m.group(1).lower() for m in _DRUG_WORDS.finditer(question)]
calls = [ToolCall(id="c0", name="get_patient_record")]
calls += [
ToolCall(id=f"d{i}", name="check_drug_for_patient", arguments={"drug": d}) for i, d in enumerate(drugs)
]
if drugs and re.search(r"состав|лактоз|желатин|аллерг|непереносим", question, re.I):
calls += [
ToolCall(id=f"i{i}", name="get_drug_info", arguments={"drug": d}) for i, d in enumerate(drugs)
]
calls.append(ToolCall(id="s0", name="search_guidelines", arguments={"query": question}))
return AssistantTurn(tool_calls=calls)
mine = re.search(r"мо(и|ими|их|ей|я|ё) (таблет|лекарств|препарат)", question, re.I)
checked = [m for m in done if m.get("name") == "check_drug_for_patient"]
if (not checked or mine) and len(checked) < 3 and _MED_HINT.search(question):
record = _tool_json(done, "get_patient_record")
rx = (record or {}).get("current_prescriptions") or []
checked_args = {str(json.loads(m["content"]).get("inn", "")) for m in checked}
rx = [r for r in rx if not any(r["drug"].lower().startswith(c[:-1]) for c in checked_args if c)]
if rx:
return AssistantTurn(
tool_calls=[ToolCall(id="d9", name="check_drug_for_patient", arguments={"drug": rx[0]["drug"]})]
)
return AssistantTurn(content=_compose(done))
def _tool_json(done: list[dict[str, Any]], name: str) -> dict[str, Any] | None:
for m in done:
if m.get("name") == name:
try:
return json.loads(m["content"])
except json.JSONDecodeError:
return None
return None
def _compose(done: list[dict[str, Any]]) -> str:
parts: list[str] = []
seen: set[str] = set()
for m in done:
if m.get("name") != "check_drug_for_patient":
continue
data = json.loads(m["content"])
if not data.get("found") or data["inn"] in seen:
continue
seen.add(data["inn"])
findings = data.get("findings") or []
if findings:
parts.append(f"По вашей карте и инструкции к препарату «{data['inn']}» есть важное:")
parts += [f"• {f['text']}" for f in findings[:3]]
else:
parts.append(f"В справочнике не нашлось особых ограничений по «{data['inn']}» с учётом вашей карты.")
for m in done:
if m.get("name") != "get_drug_info":
continue
data = json.loads(m["content"])
comp = data.get("composition") if data.get("found") else None
if comp:
record = _tool_json(done, "get_patient_record") or {}
allergies = record.get("allergies") or []
if allergies:
names = ", ".join(a["name"] for a in allergies)
hit = {t for a in allergies for t in a.get("tags", [])} & set(comp.get("excipient_tags", []))
verdict = "это вещество есть" if hit else "этого вещества нет, но состав зависит от производителя"
parts.append(f"В вашей карте: {names}. В примере состава «{data['inn']}» {verdict}.")
excipients = ", ".join(comp["excipients"])
parts.append(f"Состав ({comp['form']}): {comp['active']}; вспомогательные: {excipients}. {comp['note']}")
search = _tool_json(done, "search_guidelines") or {}
results = search.get("results") or []
if results and not parts: # medicine findings already answer the question; a KR quote would dilute it
text = results[0]["text"].strip()
parts.append(f"В клинических рекомендациях: «{text[:400]}{'…' if len(text) > 400 else ''}»")
if not parts:
return ""
parts.append("Любые изменения в лечении обсуждайте с вашим врачом.")
return "\n".join(parts)