OpenFin/backend/routers/timeline.py

"""
Timeline Navigator — historical market context browser.
Prefix: /api/timeline
"""
import json
import logging
import os
from typing import Any, Dict, List, Optional

from fastapi import APIRouter, HTTPException
from pydantic import BaseModel

logger = logging.getLogger(__name__)

router = APIRouter(prefix="/api/timeline", tags=["timeline"])


class EventCreate(BaseModel):
    name: str
    start_date: str
    end_date: Optional[str] = None
    level: str  # long | medium | short
    category: str = "macro"
    description: str = ""
    market_impact: str = ""
    affected_assets: List[str] = []
    impact_score: float = 0.5
    absorption_pct: Optional[float] = None
    relevant_indicators: List[Dict[str, Any]] = []
    parent_event_id: Optional[int] = None


class EventUpdate(EventCreate):
    pass


@router.get("/events")
def list_events() -> List[Dict[str, Any]]:
    from services.database import get_all_market_events
    return get_all_market_events()


@router.post("/events")
def create_event(body: EventCreate) -> Dict[str, Any]:
    from services.database import save_market_event
    new_id = save_market_event(body.dict())
    return {"id": new_id, "status": "created"}


@router.put("/events/{event_id}")
def update_event(event_id: int, body: EventUpdate) -> Dict[str, Any]:
    from services.database import update_market_event
    ok = update_market_event(event_id, body.dict())
    if not ok:
        raise HTTPException(404, "Event not found")
    return {"status": "updated"}


@router.delete("/events/{event_id}")
def delete_event(event_id: int) -> Dict[str, Any]:
    from services.database import delete_market_event
    delete_market_event(event_id)
    return {"status": "deleted"}


@router.post("/events/{event_id}/ai-enrich")
def ai_enrich_event(event_id: int) -> Dict[str, Any]:
    """Ask GPT-4o-mini to suggest absorption_pct + relevant_indicators for this event."""
    from services.database import get_all_market_events, update_market_event
    from datetime import datetime

    api_key = os.environ.get("OPENAI_API_KEY", "")
    if not api_key:
        raise HTTPException(400, "OpenAI API key not configured")

    # Fetch event
    all_evs = get_all_market_events()
    ev = next((e for e in all_evs if e["id"] == event_id), None)
    if not ev:
        raise HTTPException(404, "Event not found")

    from openai import OpenAI
    client = OpenAI(api_key=api_key)

    today = datetime.utcnow().strftime("%Y-%m-%d")
    days_elapsed = max(0, (datetime.strptime(today, "%Y-%m-%d") -
                           datetime.strptime(ev["start_date"][:10], "%Y-%m-%d")).days)
    is_ongoing = not ev.get("end_date")

    prompt = f"""Tu es un analyste macro senior spécialisé en options géopolitiques.

Événement à analyser:
- Nom: {ev['name']}
- Niveau: {ev['level']} (long=structurel, medium=régime, short=catalyseur)
- Catégorie: {ev['category']}
- Début: {ev['start_date']}
- Fin: {ev.get('end_date') or 'en cours'}
- Jours écoulés depuis début: {days_elapsed}j
- Description: {ev['description']}
- Impact marché: {ev.get('market_impact', '')}
- Assets affectés: {ev.get('affected_assets', '[]')}

Ta mission:
1. Estime le taux d'absorption (0-100%) de cet événement par le marché à ce jour.
   Absorption = à quel point l'impact attendu est déjà "pricé" par les marchés.
   Exemples: 0%=pas encore pricé, 50%=à moitié absorbé, 100%=fully priced in.

2. Suggère 3-5 indicateurs techniques PERTINENTS pour suivre cet événement dans le contexte d'un trader d'options.
   Pour chaque indicateur:
   - symbol: ticker yfinance (ex: "CL=F", "GC=F", "SPY", "^VIX")
   - indicator: type ("MA10", "MA20", "MA50", "MA100", "price", "RSI14")
   - label: nom affiché (ex: "Brent MA100", "Gold prix spot", "VIX 20j avg")
   - rationale: 1 phrase pourquoi cet indicateur est pertinent POUR CET ÉVÉNEMENT PRÉCIS

Règle: choisis les indicateurs selon le niveau temporel:
- Long terme → préfère MA100, MA50 (tendances structurelles)
- Moyen terme → MA20, MA50
- Court terme → MA10, MA20, prix spot

FORMAT JSON STRICT:
{{
  "absorption_pct": <0-100>,
  "absorption_rationale": "<1 phrase expliquant pourquoi ce niveau>",
  "relevant_indicators": [
    {{"symbol": "...", "indicator": "...", "label": "...", "rationale": "..."}}
  ]
}}"""

    try:
        response = client.chat.completions.create(
            model="gpt-4o-mini",
            messages=[{"role": "user", "content": prompt}],
            response_format={"type": "json_object"},
            temperature=0.3,
            max_tokens=600,
        )
        result = json.loads(response.choices[0].message.content)

        # Update event in DB
        ev_update = dict(ev)
        ev_update["affected_assets"] = json.loads(ev.get("affected_assets", "[]") or "[]")
        ev_update["relevant_indicators"] = result.get("relevant_indicators", [])
        ev_update["absorption_pct"] = result.get("absorption_pct")
        update_market_event(event_id, ev_update)

        return {
            "event_id": event_id,
            "absorption_pct": result.get("absorption_pct"),
            "absorption_rationale": result.get("absorption_rationale", ""),
            "relevant_indicators": result.get("relevant_indicators", []),
            "status": "enriched",
        }
    except Exception as e:
        logger.error(f"[Timeline] ai_enrich failed for event {event_id}: {e}")
        raise HTTPException(500, str(e))


@router.get("/day/{ref_date}")
def get_day_context(ref_date: str) -> Dict[str, Any]:
    from services.database import get_events_for_date, get_timeline_context

    events = get_events_for_date(ref_date)
    cached = get_timeline_context(ref_date)

    result: Dict[str, Any] = {
        "ref_date": ref_date,
        "events": events,
        "long_commentary": "",
        "medium_commentary": "",
        "short_commentary": "",
        "evidence_headlines": [],
        "generated_at": None,
        "has_commentary": False,
    }
    if cached:
        result.update({
            "long_commentary": cached.get("long_commentary", ""),
            "medium_commentary": cached.get("medium_commentary", ""),
            "short_commentary": cached.get("short_commentary", ""),
            "evidence_headlines": json.loads(cached.get("evidence_headlines", "[]") or "[]"),
            "generated_at": cached.get("generated_at"),
            "has_commentary": bool(cached.get("long_commentary")),
        })
    return result


@router.post("/generate/{ref_date}")
def generate_commentary(ref_date: str) -> Dict[str, Any]:
    from services.timeline_service import get_or_generate_context

    try:
        ctx = get_or_generate_context(ref_date)
        return {
            "ref_date": ref_date,
            "long_commentary": ctx.get("long_commentary", ""),
            "medium_commentary": ctx.get("medium_commentary", ""),
            "short_commentary": ctx.get("short_commentary", ""),
            "evidence_headlines": ctx.get("evidence_headlines", []),
            "events": ctx.get("events", {}),
            "has_commentary": True,
        }
    except Exception as e:
        logger.error(f"[Timeline] generate_commentary failed for {ref_date}: {e}")
        raise HTTPException(500, str(e))


@router.post("/bootstrap")
def bootstrap_events() -> Dict[str, Any]:
    from services.timeline_service import bootstrap_events as _bootstrap
    count = _bootstrap()
    return {"seeded": count, "status": "ok" if count > 0 else "already_seeded"}