feat: Timeline Navigator — contexte historique 3 temporalités COVID → aujourd'hui
- 2 nouvelles tables SQLite : market_events + timeline_context
- 32 événements historiques seedés (long/medium/short de feb 2020 à juin 2026)
- timeline_service.py : bootstrap, get_events_for_date, génération commentaires GPT-4o-mini
- /api/timeline router : GET /day/{date}, GET /events, POST /generate/{date}, POST /bootstrap
- Timeline.tsx : navigateur date avec strip visuel, 3 panneaux contextuels, catalogue d'événements
- Sidebar : entrée Timeline avec icône Layers
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
OpenSquared
@@ -3,6 +3,7 @@ from fastapi.middleware.cors import CORSMiddleware
|
||||
from routers import market_data, geopolitical, options, backtest, ai, portfolio, config, patterns, journal, cycle as cycle_router, profiles as profiles_router, reasoning as reasoning_router, knowledge as knowledge_router, options_vol as options_vol_router, analytics as analytics_router, risk as risk_router
|
||||
from routers import pattern_lab as pattern_lab_router
|
||||
from routers import specialist_desks as specialist_desks_router
|
||||
from routers import timeline as timeline_router
|
||||
from routers import logs as logs_router
|
||||
from routers import var as var_router
|
||||
from routers import reports as reports_router
|
||||
@@ -121,6 +122,7 @@ app.include_router(reports_router.router)
|
||||
app.include_router(institutional_router.router)
|
||||
app.include_router(pattern_lab_router.router)
|
||||
app.include_router(specialist_desks_router.router)
|
||||
app.include_router(timeline_router.router)
|
||||
|
||||
|
||||
@app.get("/")
|
||||
|
||||
113
backend/routers/timeline.py
Normal file
113
backend/routers/timeline.py
Normal file
@@ -0,0 +1,113 @@
|
||||
"""
|
||||
Timeline Navigator — historical market context browser.
|
||||
Prefix: /api/timeline
|
||||
"""
|
||||
import logging
|
||||
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
|
||||
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.get("/day/{ref_date}")
|
||||
def get_day_context(ref_date: str) -> Dict[str, Any]:
|
||||
"""Return cached or freshly generated timeline context for a date (YYYY-MM-DD)."""
|
||||
from services.database import get_events_for_date, get_timeline_context
|
||||
import json
|
||||
|
||||
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]:
|
||||
"""Generate (or regenerate) AI commentary for a specific date and cache it."""
|
||||
from services.timeline_service import get_or_generate_context
|
||||
import json
|
||||
|
||||
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": json.loads(ctx.get("evidence_headlines", "[]") or "[]") if isinstance(ctx.get("evidence_headlines"), str) else 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]:
|
||||
"""Seed historical market events (idempotent — only runs if table is empty)."""
|
||||
from services.timeline_service import bootstrap_events as _bootstrap
|
||||
count = _bootstrap()
|
||||
return {"seeded": count, "status": "ok" if count > 0 else "already_seeded"}
|
||||
@@ -734,6 +734,34 @@ def init_db():
|
||||
UNIQUE(asset, fetch_date)
|
||||
)""")
|
||||
|
||||
# ── Timeline Navigator ────────────────────────────────────────────────────
|
||||
c.execute("""CREATE TABLE IF NOT EXISTS market_events (
|
||||
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
||||
name TEXT NOT NULL,
|
||||
start_date TEXT NOT NULL,
|
||||
end_date TEXT,
|
||||
level TEXT NOT NULL,
|
||||
category TEXT NOT NULL DEFAULT 'macro',
|
||||
description TEXT NOT NULL DEFAULT '',
|
||||
market_impact TEXT DEFAULT '',
|
||||
affected_assets TEXT DEFAULT '[]',
|
||||
impact_score REAL DEFAULT 0.5,
|
||||
parent_event_id INTEGER REFERENCES market_events(id),
|
||||
created_at TEXT DEFAULT (datetime('now'))
|
||||
)""")
|
||||
|
||||
c.execute("""CREATE TABLE IF NOT EXISTS timeline_context (
|
||||
ref_date TEXT PRIMARY KEY,
|
||||
long_event_id INTEGER REFERENCES market_events(id),
|
||||
medium_event_id INTEGER REFERENCES market_events(id),
|
||||
short_event_id INTEGER REFERENCES market_events(id),
|
||||
long_commentary TEXT DEFAULT '',
|
||||
medium_commentary TEXT DEFAULT '',
|
||||
short_commentary TEXT DEFAULT '',
|
||||
evidence_headlines TEXT DEFAULT '[]',
|
||||
generated_at TEXT DEFAULT (datetime('now'))
|
||||
)""")
|
||||
|
||||
# Seed desk defaults (idempotent)
|
||||
_desk_count = c.execute("SELECT COUNT(*) FROM asset_class_configs").fetchone()[0]
|
||||
if _desk_count == 0:
|
||||
@@ -4431,3 +4459,121 @@ def update_report_result(
|
||||
return True
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
# ── Timeline Navigator ────────────────────────────────────────────────────────
|
||||
|
||||
def save_market_event(ev: Dict[str, Any]) -> int:
|
||||
import json
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.execute("""INSERT INTO market_events
|
||||
(name, start_date, end_date, level, category, description, market_impact,
|
||||
affected_assets, impact_score, parent_event_id)
|
||||
VALUES (?,?,?,?,?,?,?,?,?,?)""",
|
||||
(ev["name"], ev["start_date"], ev.get("end_date"),
|
||||
ev["level"], ev.get("category", "macro"), ev.get("description", ""),
|
||||
ev.get("market_impact", ""), json.dumps(ev.get("affected_assets", [])),
|
||||
ev.get("impact_score", 0.5), ev.get("parent_event_id")))
|
||||
conn.commit()
|
||||
return cur.lastrowid
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def update_market_event(event_id: int, ev: Dict[str, Any]) -> bool:
|
||||
import json
|
||||
conn = get_conn()
|
||||
try:
|
||||
conn.execute("""UPDATE market_events SET
|
||||
name=?, start_date=?, end_date=?, level=?, category=?, description=?,
|
||||
market_impact=?, affected_assets=?, impact_score=?
|
||||
WHERE id=?""",
|
||||
(ev["name"], ev["start_date"], ev.get("end_date"),
|
||||
ev["level"], ev.get("category", "macro"), ev.get("description", ""),
|
||||
ev.get("market_impact", ""), json.dumps(ev.get("affected_assets", [])),
|
||||
ev.get("impact_score", 0.5), event_id))
|
||||
conn.commit()
|
||||
return True
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_all_market_events() -> List[Dict[str, Any]]:
|
||||
conn = get_conn()
|
||||
try:
|
||||
rows = conn.execute(
|
||||
"SELECT * FROM market_events ORDER BY start_date DESC"
|
||||
).fetchall()
|
||||
return [dict(r) for r in rows]
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_events_for_date(ref_date: str) -> Dict[str, Any]:
|
||||
"""Return the active long/medium/short events for a given date."""
|
||||
conn = get_conn()
|
||||
try:
|
||||
def query_level(level: str) -> Optional[Dict]:
|
||||
row = conn.execute("""
|
||||
SELECT * FROM market_events
|
||||
WHERE level=? AND start_date<=?
|
||||
AND (end_date IS NULL OR end_date>=?)
|
||||
ORDER BY start_date DESC LIMIT 1""",
|
||||
(level, ref_date, ref_date)).fetchone()
|
||||
return dict(row) if row else None
|
||||
|
||||
return {
|
||||
"long": query_level("long"),
|
||||
"medium": query_level("medium"),
|
||||
"short": query_level("short"),
|
||||
}
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_timeline_context(ref_date: str) -> Optional[Dict[str, Any]]:
|
||||
conn = get_conn()
|
||||
try:
|
||||
row = conn.execute(
|
||||
"SELECT * FROM timeline_context WHERE ref_date=?", (ref_date,)
|
||||
).fetchone()
|
||||
return dict(row) if row else None
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def save_timeline_context(ctx: Dict[str, Any]) -> None:
|
||||
import json
|
||||
conn = get_conn()
|
||||
try:
|
||||
conn.execute("""INSERT INTO timeline_context
|
||||
(ref_date, long_event_id, medium_event_id, short_event_id,
|
||||
long_commentary, medium_commentary, short_commentary,
|
||||
evidence_headlines, generated_at)
|
||||
VALUES (?,?,?,?,?,?,?,?,datetime('now'))
|
||||
ON CONFLICT(ref_date) DO UPDATE SET
|
||||
long_event_id=excluded.long_event_id,
|
||||
medium_event_id=excluded.medium_event_id,
|
||||
short_event_id=excluded.short_event_id,
|
||||
long_commentary=excluded.long_commentary,
|
||||
medium_commentary=excluded.medium_commentary,
|
||||
short_commentary=excluded.short_commentary,
|
||||
evidence_headlines=excluded.evidence_headlines,
|
||||
generated_at=datetime('now')""",
|
||||
(ctx["ref_date"],
|
||||
ctx.get("long_event_id"), ctx.get("medium_event_id"), ctx.get("short_event_id"),
|
||||
ctx.get("long_commentary", ""), ctx.get("medium_commentary", ""),
|
||||
ctx.get("short_commentary", ""),
|
||||
json.dumps(ctx.get("evidence_headlines", []))))
|
||||
conn.commit()
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def count_market_events() -> int:
|
||||
conn = get_conn()
|
||||
try:
|
||||
return conn.execute("SELECT COUNT(*) FROM market_events").fetchone()[0]
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
396
backend/services/timeline_service.py
Normal file
396
backend/services/timeline_service.py
Normal file
@@ -0,0 +1,396 @@
|
||||
"""
|
||||
Timeline Navigator Service.
|
||||
|
||||
Manages the historical market event database (COVID → today) and generates
|
||||
3-level (long/medium/short) temporal context with AI commentary for any date.
|
||||
"""
|
||||
import json
|
||||
import logging
|
||||
from datetime import date, datetime, timedelta
|
||||
from typing import Any, Dict, List, Optional
|
||||
|
||||
from services.database import (
|
||||
count_market_events,
|
||||
get_all_market_events,
|
||||
get_events_for_date,
|
||||
get_timeline_context,
|
||||
save_market_event,
|
||||
save_timeline_context,
|
||||
)
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
# ── Seed event catalogue ──────────────────────────────────────────────────────
|
||||
# Each entry: name, start_date, end_date (None=ongoing), level, category,
|
||||
# description, market_impact, affected_assets (list), impact_score
|
||||
|
||||
_SEED_EVENTS: List[Dict[str, Any]] = [
|
||||
# ───── LONG-TERM STRUCTURAL EVENTS ─────
|
||||
{
|
||||
"name": "COVID-19 Crash", "level": "long", "category": "macro",
|
||||
"start_date": "2020-02-20", "end_date": "2020-03-23",
|
||||
"description": "Pandémie mondiale — liquidation forcée de tous les actifs risqués, gel des marchés du crédit, volatilité extrême (VIX>80).",
|
||||
"market_impact": "SPX -34% en 33 jours. VIX 85. Oil négatif avril. Fuite vers USD et T-Bills.",
|
||||
"affected_assets": ["SPX", "VIX", "CL", "HYG", "USD", "Gold"],
|
||||
"impact_score": 1.0,
|
||||
},
|
||||
{
|
||||
"name": "Fed QE Infinity & ZIRP", "level": "long", "category": "monetary",
|
||||
"start_date": "2020-03-23", "end_date": "2022-03-16",
|
||||
"description": "Fed à 0% + QE illimité. Injection de $4T+ en 2 ans. Régime de taux zéro et liquidité abondante pour tous les actifs.",
|
||||
"market_impact": "Bull market actions/crypto/immobilier. Compression des primes de risque. Carry trades extrêmes.",
|
||||
"affected_assets": ["SPX", "BTC", "Gold", "HYG", "TLT", "EUR/USD"],
|
||||
"impact_score": 0.95,
|
||||
},
|
||||
{
|
||||
"name": "Inflation Surge", "level": "long", "category": "macro",
|
||||
"start_date": "2021-04-01", "end_date": "2023-06-01",
|
||||
"description": "CPI US monte de 2% à 9,1% (juin 2022). Chocs d'offre post-COVID + demande stimulée + énergie. Fin de l'ère TINA.",
|
||||
"market_impact": "Fin du bull bond. Rotation value/energy. Pression sur multiples tech.",
|
||||
"affected_assets": ["TLT", "GC", "CL", "XLE", "EUR/USD", "TIPS"],
|
||||
"impact_score": 0.9,
|
||||
},
|
||||
{
|
||||
"name": "Fed Hike Cycle", "level": "long", "category": "monetary",
|
||||
"start_date": "2022-03-16", "end_date": "2023-07-26",
|
||||
"description": "525bps de hausse en 17 mois — cycle de resserrement le plus rapide depuis Volcker. Pivot du QE vers QT.",
|
||||
"market_impact": "Krach obligataire 2022 (-25% TLT). Tech -35%. Strength USD. Recession fears.",
|
||||
"affected_assets": ["TLT", "QQQ", "SPX", "USD", "HYG", "GC"],
|
||||
"impact_score": 0.95,
|
||||
},
|
||||
{
|
||||
"name": "Fed Pause (Peak Rates)", "level": "long", "category": "monetary",
|
||||
"start_date": "2023-07-26", "end_date": "2024-09-18",
|
||||
"description": "Fed maintient 5,25-5,50% pendant 14 mois. Marché anticipe baisse, puis repousse sans cesse. Atterrissage en douceur.",
|
||||
"market_impact": "Rally équités (SPX +26% 2023). Curve inversion. Gold accumulation.",
|
||||
"affected_assets": ["SPX", "GC", "USD", "TLT", "2Y UST"],
|
||||
"impact_score": 0.7,
|
||||
},
|
||||
{
|
||||
"name": "Fed Easing Cycle", "level": "long", "category": "monetary",
|
||||
"start_date": "2024-09-18", "end_date": None,
|
||||
"description": "Fed commence à baisser : -25bps sept 2024, -25bps nov 2024, -25bps déc 2024. Pause 2025. Nouveau cycle expansif.",
|
||||
"market_impact": "USD faiblesse. Or new ATH. Courbe se dépente. Risque inflation secondaire.",
|
||||
"affected_assets": ["USD", "GC", "TLT", "EUR/USD", "EM FX"],
|
||||
"impact_score": 0.75,
|
||||
},
|
||||
{
|
||||
"name": "Trump 2.0 & Tariff Era", "level": "long", "category": "geopolitical",
|
||||
"start_date": "2025-01-20", "end_date": None,
|
||||
"description": "Retour de Trump. Tarifs universels 10% + tarifs China jusqu'à 145%. Restructuration de l'ordre commercial mondial.",
|
||||
"market_impact": "VIX spike. USD strength initial. Disruption chaînes supply. Inflation tarif.",
|
||||
"affected_assets": ["SPX", "USD", "GC", "CL", "EM FX", "ZC", "ZS"],
|
||||
"impact_score": 0.85,
|
||||
},
|
||||
{
|
||||
"name": "US-Iran Military Conflict", "level": "long", "category": "geopolitical",
|
||||
"start_date": "2025-06-13", "end_date": None,
|
||||
"description": "Frappes US sur sites nucléaires iraniens. Escalade militaire directe USA-Iran en Moyen-Orient.",
|
||||
"market_impact": "CL +15% initial. Gold safe haven. Risque Hormuz premium. Risk-off global.",
|
||||
"affected_assets": ["CL", "NG", "GC", "USD", "VIX", "Tanker stocks"],
|
||||
"impact_score": 0.9,
|
||||
},
|
||||
|
||||
# ───── MEDIUM-TERM REGIME EVENTS ─────
|
||||
{
|
||||
"name": "COVID V-Bottom Recovery", "level": "medium", "category": "macro",
|
||||
"start_date": "2020-03-23", "end_date": "2021-01-01",
|
||||
"description": "Rebond historique post-crash. Fed backstop + stimulus fiscal. SPX récupère tous les pertes en 5 mois.",
|
||||
"market_impact": "Tech FAANG leaders. Growth stocks premium. Small cap lagging.",
|
||||
"affected_assets": ["SPX", "QQQ", "IWM", "Gold"],
|
||||
"impact_score": 0.85,
|
||||
},
|
||||
{
|
||||
"name": "Pfizer Vaccine Rotation", "level": "medium", "category": "macro",
|
||||
"start_date": "2020-11-09", "end_date": "2021-06-01",
|
||||
"description": "Annonce vaccin Pfizer (+95% efficacité). Mega-rotation value vs growth. Cyclicals/energy reprennent.",
|
||||
"market_impact": "Banks +20% en 2 sem. Energy recovery. Growth stocks stagnent.",
|
||||
"affected_assets": ["XLF", "XLE", "IWM", "QQQ", "Oil"],
|
||||
"impact_score": 0.7,
|
||||
},
|
||||
{
|
||||
"name": "GameStop & Meme Frenzy", "level": "medium", "category": "market",
|
||||
"start_date": "2021-01-13", "end_date": "2021-02-05",
|
||||
"description": "Short squeeze GameStop orchestré par retail Reddit. Systemic risk sur HF leverage. Brokers restriction d'achats.",
|
||||
"market_impact": "GME +2700%. HF losses $10B+. Systemic risk perceptions élevées.",
|
||||
"affected_assets": ["GME", "AMC", "HYG", "VIX"],
|
||||
"impact_score": 0.5,
|
||||
},
|
||||
{
|
||||
"name": "Archegos Collapse", "level": "medium", "category": "market",
|
||||
"start_date": "2021-03-26", "end_date": "2021-04-10",
|
||||
"description": "Family office Archegos default sur $20B de swaps TRS. Forced selling sur VIACOM/Discovery/GS/MS/CS.",
|
||||
"market_impact": "CS -15%, Nomura -14%. Leverage disclosure concerns. Media stocks -30%.",
|
||||
"affected_assets": ["CS", "Nomura", "Media sector"],
|
||||
"impact_score": 0.55,
|
||||
},
|
||||
{
|
||||
"name": "China Tech Crackdown", "level": "medium", "category": "geopolitical",
|
||||
"start_date": "2021-07-01", "end_date": "2022-03-01",
|
||||
"description": "Pékin régule les géants tech chinois (Alibaba, Didi, Tencent). Suspension IPO Didi. Education privée bannie.",
|
||||
"market_impact": "KWEB -60%. Alibaba -70%. Risque réglementaire chinois systémique.",
|
||||
"affected_assets": ["BABA", "BIDU", "JD", "KWEB", "FXI"],
|
||||
"impact_score": 0.7,
|
||||
},
|
||||
{
|
||||
"name": "Evergrande Crisis", "level": "medium", "category": "macro",
|
||||
"start_date": "2021-09-01", "end_date": "2022-01-01",
|
||||
"description": "Evergrande (2ème promoteur immobilier chinois) en défaut sur $300B de dettes. Contagion secteur immo chinois.",
|
||||
"market_impact": "HY spread China +500bps. CNY pressure. Contagion EM partielle.",
|
||||
"affected_assets": ["CNY", "FXI", "HY China bonds", "Iron Ore"],
|
||||
"impact_score": 0.65,
|
||||
},
|
||||
{
|
||||
"name": "Ukraine Invasion", "level": "medium", "category": "geopolitical",
|
||||
"start_date": "2022-02-24", "end_date": None,
|
||||
"description": "Invasion russe de l'Ukraine. Sanctions massives Russia. Disruption grains, gaz, engrais. Réarmement Europe.",
|
||||
"market_impact": "NG Europe x10. Wheat +60%. Gold safe haven. EUR weakness. Defense rally.",
|
||||
"affected_assets": ["NG", "ZW", "GC", "EUR/USD", "Defense stocks"],
|
||||
"impact_score": 0.9,
|
||||
},
|
||||
{
|
||||
"name": "CPI Peak 9.1% & Jackson Hole 2022", "level": "medium", "category": "monetary",
|
||||
"start_date": "2022-06-10", "end_date": "2022-10-01",
|
||||
"description": "CPI US 9,1% — peak inflation 40 ans. Jackson Hole Powell hawkish extrême. Marchés capitulent sur pivot.",
|
||||
"market_impact": "SPX -25% 2022. TLT -30%. EUR/USD parité. DXY 114.",
|
||||
"affected_assets": ["SPX", "TLT", "EUR/USD", "USD", "GC"],
|
||||
"impact_score": 0.85,
|
||||
},
|
||||
{
|
||||
"name": "SVB Collapse & Banking Crisis", "level": "medium", "category": "market",
|
||||
"start_date": "2023-03-08", "end_date": "2023-05-01",
|
||||
"description": "Silicon Valley Bank run sur les dépôts. FDIC intervention. Credit Suisse absorbe par UBS. Contagion régionale US.",
|
||||
"market_impact": "KBE -25%. TLT +10% (flight to safety). VIX spike. Credit tightening.",
|
||||
"affected_assets": ["KBE", "TLT", "VIX", "EUR/USD", "Gold"],
|
||||
"impact_score": 0.7,
|
||||
},
|
||||
{
|
||||
"name": "AI Boom / Nvidia Mania", "level": "medium", "category": "market",
|
||||
"start_date": "2023-05-25", "end_date": None,
|
||||
"description": "Nvidia guidance x3 — début du mega-cycle IA. ChatGPT + Capex hyperscalers. Concentration marchés extrême.",
|
||||
"market_impact": "NVDA +800%. Mag7 domination. SPX concentration record. Small cap underperform.",
|
||||
"affected_assets": ["NVDA", "MSFT", "GOOGL", "QQQ", "SMH"],
|
||||
"impact_score": 0.85,
|
||||
},
|
||||
{
|
||||
"name": "Israel-Hamas War", "level": "medium", "category": "geopolitical",
|
||||
"start_date": "2023-10-07", "end_date": None,
|
||||
"description": "Attaque Hamas + contre-offensive Gaza. Escalade Hezbollah, Yemen Houthis, Iran proxy. Risque Hormuz.",
|
||||
"market_impact": "CL spike initial +5%. Gold safe haven $2100. Défense stocks +15%.",
|
||||
"affected_assets": ["CL", "GC", "Defense", "ILS"],
|
||||
"impact_score": 0.7,
|
||||
},
|
||||
{
|
||||
"name": "BOJ Fin du YCC", "level": "medium", "category": "monetary",
|
||||
"start_date": "2024-03-19", "end_date": None,
|
||||
"description": "Bank of Japan abandonne le Yield Curve Control et hausse taux à 0.1% (premier hike en 17 ans). JPY carry unwind.",
|
||||
"market_impact": "JPY +5% en 1 mois. Carry trade unwind global. Nikkei -12% août 2024.",
|
||||
"affected_assets": ["JPY", "Nikkei", "USD/JPY", "JGB", "EM carry"],
|
||||
"impact_score": 0.75,
|
||||
},
|
||||
{
|
||||
"name": "Jackson Hole Pivot 2024", "level": "medium", "category": "monetary",
|
||||
"start_date": "2024-08-23", "end_date": "2024-12-31",
|
||||
"description": "Powell annonce officiellement la fin du cycle de resserrement et signale des baisses imminentes.",
|
||||
"market_impact": "USD -2%. Gold ATH. Equities rally. Curve bull steepening.",
|
||||
"affected_assets": ["USD", "GC", "SPX", "TLT", "EUR/USD"],
|
||||
"impact_score": 0.75,
|
||||
},
|
||||
{
|
||||
"name": "Trump Tariff Shock (Liberation Day)", "level": "medium", "category": "geopolitical",
|
||||
"start_date": "2025-04-02", "end_date": "2025-04-09",
|
||||
"description": "Annonce tarifs universels 10% + tarifs pays spécifiques jusqu'à 50%. Pire choc tarifaire depuis 1930.",
|
||||
"market_impact": "SPX -15% en 4 séances. VIX 60. Credit spreads +200bps.",
|
||||
"affected_assets": ["SPX", "VIX", "USD", "GC", "HYG"],
|
||||
"impact_score": 0.9,
|
||||
},
|
||||
{
|
||||
"name": "90-Day Tariff Pause", "level": "medium", "category": "geopolitical",
|
||||
"start_date": "2025-04-09", "end_date": "2025-07-09",
|
||||
"description": "Trump annonce pause 90 jours sur la majorité des tarifs réciproqués pour permettre des négociations.",
|
||||
"market_impact": "SPX +9.5% en 1 séance (3ème meilleure journée historique). VIX collapse.",
|
||||
"affected_assets": ["SPX", "VIX", "USD", "EUR/USD", "EM"],
|
||||
"impact_score": 0.8,
|
||||
},
|
||||
{
|
||||
"name": "Iran Ceasefire Talks", "level": "medium", "category": "geopolitical",
|
||||
"start_date": "2025-06-20", "end_date": None,
|
||||
"description": "Négociations de cessez-le-feu US-Iran sous médiation Oman + Suisse. Dé-escalade progressive.",
|
||||
"market_impact": "CL -8%. Hormuz risk premium réduit. Gold consolidation.",
|
||||
"affected_assets": ["CL", "GC", "USD", "VIX"],
|
||||
"impact_score": 0.7,
|
||||
},
|
||||
|
||||
# ───── SHORT-TERM CATALYST EVENTS ─────
|
||||
{
|
||||
"name": "Fed Meeting — Pause Confirmée", "level": "short", "category": "monetary",
|
||||
"start_date": "2025-05-07", "end_date": "2025-05-21",
|
||||
"description": "FOMC maintient les taux — attend clarification tarifs avant de bouger. Powell data-dependent.",
|
||||
"market_impact": "USD stable. Marchés attendent. Curve stable.",
|
||||
"affected_assets": ["USD", "TLT", "SPX"],
|
||||
"impact_score": 0.5,
|
||||
},
|
||||
{
|
||||
"name": "US-China Trade Truce", "level": "short", "category": "geopolitical",
|
||||
"start_date": "2025-05-12", "end_date": "2025-08-12",
|
||||
"description": "Réduction tarifs US-China : 145%→30% (US) et 125%→10% (China) pendant 90 jours suite aux négociations de Genève.",
|
||||
"market_impact": "SPX +3%. USD -1%. Relief rally. Chaînes supply se reconstituent.",
|
||||
"affected_assets": ["SPX", "USD", "CNY", "ZS", "ZC"],
|
||||
"impact_score": 0.7,
|
||||
},
|
||||
{
|
||||
"name": "US Frappe Iran — Escalade", "level": "short", "category": "geopolitical",
|
||||
"start_date": "2025-06-13", "end_date": "2025-06-20",
|
||||
"description": "Frappes US-Israël sur Natanz et Fordow. Iran riposte sur bases US région. Risk-off maximal.",
|
||||
"market_impact": "CL +15%. Gold +5%. VIX 35. USD strength.",
|
||||
"affected_assets": ["CL", "GC", "VIX", "USD", "Defense"],
|
||||
"impact_score": 0.9,
|
||||
},
|
||||
{
|
||||
"name": "Iran Ceasefire Signal", "level": "short", "category": "geopolitical",
|
||||
"start_date": "2025-06-20", "end_date": "2025-07-10",
|
||||
"description": "Signal de cessez-le-feu Iran-US via Oman. Début dé-escalade. Marchés absorbent la prime de risque.",
|
||||
"market_impact": "CL -8% en 2 séances. VIX retombe. Risk-on partiel.",
|
||||
"affected_assets": ["CL", "VIX", "GC", "SPX"],
|
||||
"impact_score": 0.65,
|
||||
},
|
||||
{
|
||||
"name": "SNB Meeting — Easing Signalé", "level": "short", "category": "monetary",
|
||||
"start_date": "2026-06-19", "end_date": "2026-07-03",
|
||||
"description": "Swiss National Bank réunion — signal de poursuite du cycle d'assouplissement. CHF sous pression.",
|
||||
"market_impact": "CHF -0.5%. EUR/CHF hausse. Francs positions réduites.",
|
||||
"affected_assets": ["CHF", "EUR/CHF", "USD/CHF"],
|
||||
"impact_score": 0.45,
|
||||
},
|
||||
]
|
||||
|
||||
|
||||
def bootstrap_events() -> int:
|
||||
"""Seed the market_events table. Idempotent — only runs if table is empty."""
|
||||
if count_market_events() > 0:
|
||||
logger.info("[Timeline] market_events already seeded, skip bootstrap")
|
||||
return 0
|
||||
saved = 0
|
||||
for ev in _SEED_EVENTS:
|
||||
try:
|
||||
save_market_event(ev)
|
||||
saved += 1
|
||||
except Exception as e:
|
||||
logger.warning(f"[Timeline] Failed to save event '{ev['name']}': {e}")
|
||||
logger.info(f"[Timeline] Bootstrapped {saved} market events")
|
||||
return saved
|
||||
|
||||
|
||||
def _days_since(ref_date: str, start_date: str) -> int:
|
||||
try:
|
||||
d0 = datetime.strptime(start_date[:10], "%Y-%m-%d").date()
|
||||
d1 = datetime.strptime(ref_date[:10], "%Y-%m-%d").date()
|
||||
return (d1 - d0).days
|
||||
except Exception:
|
||||
return 0
|
||||
|
||||
|
||||
def generate_daily_commentary(
|
||||
ref_date: str,
|
||||
long_ev: Optional[Dict],
|
||||
medium_ev: Optional[Dict],
|
||||
short_ev: Optional[Dict],
|
||||
) -> Dict[str, str]:
|
||||
"""Call GPT-4o-mini to generate 3 French trading commentaries for ref_date."""
|
||||
import os
|
||||
from openai import OpenAI
|
||||
|
||||
api_key = os.environ.get("OPENAI_API_KEY", "")
|
||||
if not api_key:
|
||||
return {
|
||||
"long_commentary": "IA non configurée — ajoutez votre clé OpenAI dans la configuration.",
|
||||
"medium_commentary": "",
|
||||
"short_commentary": "",
|
||||
}
|
||||
|
||||
client = OpenAI(api_key=api_key)
|
||||
|
||||
def _ev_summary(ev: Optional[Dict], level: str) -> str:
|
||||
if not ev:
|
||||
return f"[Pas d'événement {level} actif]"
|
||||
days = _days_since(ref_date, ev["start_date"])
|
||||
return (
|
||||
f"Nom: {ev['name']}\n"
|
||||
f"Depuis: {ev['start_date']} ({days} jours)\n"
|
||||
f"Description: {ev['description']}\n"
|
||||
f"Impact marché: {ev.get('market_impact', '')}"
|
||||
)
|
||||
|
||||
prompt = f"""Tu es un analyste macro senior en trading d'options géopolitiques.
|
||||
Date d'analyse: {ref_date}
|
||||
|
||||
Contexte temporel à cette date:
|
||||
|
||||
== LONG TERME (régime structurel) ==
|
||||
{_ev_summary(long_ev, 'long terme')}
|
||||
|
||||
== MOYEN TERME (régime courant) ==
|
||||
{_ev_summary(medium_ev, 'moyen terme')}
|
||||
|
||||
== COURT TERME (catalyseur récent) ==
|
||||
{_ev_summary(short_ev, 'court terme')}
|
||||
|
||||
Pour chaque niveau temporel, rédige UN SEUL paragraphe de 3-4 phrases en français.
|
||||
Explique ce que signifie ce contexte pour un trader d'options sur commodités/indices/forex à cette date précise.
|
||||
Sois concret : implications pour le vol, le positionnement, les risques.
|
||||
|
||||
FORMAT DE RÉPONSE (JSON strict):
|
||||
{{
|
||||
"long_commentary": "...",
|
||||
"medium_commentary": "...",
|
||||
"short_commentary": "..."
|
||||
}}"""
|
||||
|
||||
try:
|
||||
response = client.chat.completions.create(
|
||||
model="gpt-4o-mini",
|
||||
messages=[{"role": "user", "content": prompt}],
|
||||
response_format={"type": "json_object"},
|
||||
temperature=0.4,
|
||||
max_tokens=800,
|
||||
)
|
||||
result = json.loads(response.choices[0].message.content)
|
||||
return {
|
||||
"long_commentary": result.get("long_commentary", ""),
|
||||
"medium_commentary": result.get("medium_commentary", ""),
|
||||
"short_commentary": result.get("short_commentary", ""),
|
||||
}
|
||||
except Exception as e:
|
||||
logger.warning(f"[Timeline] GPT commentary failed for {ref_date}: {e}")
|
||||
return {
|
||||
"long_commentary": f"Erreur génération IA: {e}",
|
||||
"medium_commentary": "",
|
||||
"short_commentary": "",
|
||||
}
|
||||
|
||||
|
||||
def get_or_generate_context(ref_date: str) -> Dict[str, Any]:
|
||||
"""Return timeline context for a date — from cache or freshly generated."""
|
||||
cached = get_timeline_context(ref_date)
|
||||
if cached and cached.get("long_commentary"):
|
||||
events = get_events_for_date(ref_date)
|
||||
return {**cached, "events": events}
|
||||
|
||||
events = get_events_for_date(ref_date)
|
||||
commentaries = generate_daily_commentary(
|
||||
ref_date,
|
||||
events.get("long"),
|
||||
events.get("medium"),
|
||||
events.get("short"),
|
||||
)
|
||||
|
||||
ctx = {
|
||||
"ref_date": ref_date,
|
||||
"long_event_id": events["long"]["id"] if events.get("long") else None,
|
||||
"medium_event_id": events["medium"]["id"] if events.get("medium") else None,
|
||||
"short_event_id": events["short"]["id"] if events.get("short") else None,
|
||||
**commentaries,
|
||||
"evidence_headlines": [],
|
||||
}
|
||||
save_timeline_context(ctx)
|
||||
return {**ctx, "events": events}
|
||||
Reference in New Issue
Block a user