shira-hermes/api/routes/smart_assistant.py

"""SmartAssistant chat endpoint — same HTTP contract as ai-gateway."""

from __future__ import annotations

import asyncio
import os
import logging

from fastapi import APIRouter, Request, HTTPException
from pydantic import BaseModel

from api.services.prompt_builder import build_case_prompt, build_office_prompt
from api.services.agent_runner import AgentRunner
from api.services.skills import list_skills
from api.services.memory import load_user_profile, load_case_memory
from api.services.learner import extract_skills

logger = logging.getLogger("shira.api")

router = APIRouter()

# Hold references to background tasks so they don't get garbage-collected
_background_tasks: set[asyncio.Task] = set()


def _get_api_key() -> str:
    return os.environ.get("API_KEY", "")


def _verify_auth(request: Request):
    """Verify X-Api-Key header if API_KEY is configured."""
    expected = _get_api_key()
    if not expected:
        return  # No auth configured
    provided = request.headers.get("X-Api-Key", "")
    if provided != expected:
        logger.warning(
            "[auth] Key mismatch — expected=%s... provided=%s...",
            expected[:8] if expected else "NONE",
            provided[:8] if provided else "NONE",
        )
        raise HTTPException(status_code=401, detail="Unauthorized")


def _get_runner() -> AgentRunner:
    return AgentRunner(
        ai_gateway_url=os.environ.get("AI_GATEWAY_URL", "http://localhost:3000"),
        ai_gateway_api_key=os.environ.get("AI_GATEWAY_API_KEY", ""),
        model=os.environ.get("CLAUDE_MODEL", "sonnet"),
        max_tokens=int(os.environ.get("MAX_OUTPUT_TOKENS", "4096")),
        max_iterations=int(os.environ.get("MAX_AGENT_ITERATIONS", "10")),
    )


@router.post("/api/smart-assistant/chat")
async def smart_assistant_chat(request: Request):
    _verify_auth(request)

    body = await request.json()
    message = body.get("message", "")
    context = body.get("context", {})
    conversation_id = body.get("conversationId", "")
    conversation_history = body.get("conversationHistory", [])
    mode = body.get("mode", "case")

    # EspoCRM credentials
    espocrm = body.get("espocrm", {})
    espocrm_url = espocrm.get("url") or os.environ.get("ESPOCRM_URL", "")
    espocrm_api_key = espocrm.get("apiKey") or os.environ.get("ESPOCRM_API_KEY", "")

    if not message and not conversation_history:
        return {"text": "no message", "actions": [], "conversationId": conversation_id}

    # Extract IDs for memory loading
    case_id = (context.get("case") or {}).get("id") or (context.get("drillDown", {}).get("case", {}).get("id"))
    user_id = (context.get("currentUser") or {}).get("id")

    # Load cross-conversation memory and skills
    user_profile = load_user_profile(user_id) if user_id else ""
    case_memory_notes = load_case_memory(case_id) if case_id else ""
    available_skills = list_skills()

    # Build system prompt with memory and skills injected
    if mode == "office":
        system_prompt = build_office_prompt(
            context,
            user_profile=user_profile,
            available_skills=available_skills,
        )
    else:
        system_prompt = build_case_prompt(
            context,
            user_profile=user_profile,
            case_memory_notes=case_memory_notes,
            available_skills=available_skills,
        )

    # Build messages from conversation history
    messages = []
    for m in conversation_history:
        if m.get("role") == "tool_result":
            messages.append({
                "role": "user",
                "content": f'[תוצאת כלי {m.get("toolName", "")}]:\n{m.get("content", "")}',
            })
        else:
            messages.append({
                "role": "assistant" if m.get("role") == "assistant" else "user",
                "content": m.get("content", ""),
            })

    if message:
        messages.append({"role": "user", "content": message})

    logger.info(
        '[smart-assistant] mode=%s conv=%s espocrm=%s msg="%s..."',
        mode, conversation_id, espocrm_url, message[:50],
    )

    runner = _get_runner()
    text = await runner.run(
        system_prompt=system_prompt,
        messages=messages,
        context=context,
        espocrm_url=espocrm_url,
        espocrm_api_key=espocrm_api_key,
    )

    logger.info("[smart-assistant] response: %s...", text[:80])

    # Trigger self-learning in the background (non-blocking)
    conversation_messages = runner.get_messages()
    if conversation_messages:
        task = asyncio.create_task(_learn_in_background(conversation_messages))
        _background_tasks.add(task)
        task.add_done_callback(_background_tasks.discard)

    return {
        "text": text,
        "actions": [],
        "executedActions": [],
        "conversationId": conversation_id,
    }


async def _learn_in_background(messages: list[dict]) -> None:
    """Run skill extraction in background — errors are swallowed and logged."""
    try:
        await extract_skills(messages)
    except Exception as e:
        logger.error("[learner] background learning failed: %s", e)