Initial implementation of Options Sidekick

Full-stack iOS options trading assistant:
- Python FastAPI backend with SQLite, APScheduler (15-min position monitor),
  APNs push notifications, and yfinance market data integration
- Signal engine: IV Rank (rolling HV proxy), SMA-50/200, swing-based
  support/resistance, earnings detection, signal strength scoring and
  noise-resistant SHA hash for change detection
- Recommendation engine: covered call and cash-secured put strike/expiry
  selection across 0DTE, 1DTE, weekly, and monthly horizons
- REST API: /devices, /portfolio, /recommendations, /positions, /signals, /alerts
- iOS SwiftUI app (iOS 17+): dashboard, recommendations, trades, portfolio,
  and alerts tabs with push notification deep-linking
- Unit + integration tests for signal engine and API layer

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/.env.example

@@ -0,0 +1,7 @@
+DATABASE_URL=sqlite:///./app.db
+APNS_KEY_PATH=./AuthKey_XXXXXXXXXX.p8
+APNS_KEY_ID=XXXXXXXXXX
+APNS_TEAM_ID=XXXXXXXXXX
+APNS_BUNDLE_ID=com.yourname.options-sidekick
+APNS_USE_SANDBOX=true
+RECOMMENDATION_THROTTLE_MINUTES=5

backend/Procfile

@@ -0,0 +1 @@
+web: uvicorn app.main:app --host 0.0.0.0 --port $PORT

backend/app/config.py

@@ -0,0 +1,19 @@
+from pydantic_settings import BaseSettings
+
+
+class Settings(BaseSettings):
+    database_url: str = "sqlite:///./app.db"
+    apns_key_path: str = "./AuthKey_XXXXXXXXXX.p8"
+    apns_key_id: str = ""
+    apns_team_id: str = ""
+    apns_bundle_id: str = "com.yourname.options-sidekick"
+    apns_use_sandbox: bool = True
+    recommendation_throttle_minutes: int = 5
+    monitor_interval_seconds: int = 900  # 15 minutes
+
+    class Config:
+        env_file = ".env"
+        env_file_encoding = "utf-8"
+
+
+settings = Settings()

backend/app/database.py

@@ -0,0 +1,27 @@
+from sqlalchemy import create_engine
+from sqlalchemy.orm import sessionmaker, DeclarativeBase
+from app.config import settings
+
+engine = create_engine(
+    settings.database_url,
+    connect_args={"check_same_thread": False} if "sqlite" in settings.database_url else {},
+)
+
+SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine)
+
+
+class Base(DeclarativeBase):
+    pass
+
+
+def get_db():
+    db = SessionLocal()
+    try:
+        yield db
+    finally:
+        db.close()
+
+
+def init_db():
+    from app.models.db_models import Device, StockPosition, OptionPosition, Recommendation, Alert  # noqa: F401
+    Base.metadata.create_all(bind=engine)

backend/app/main.py

@@ -0,0 +1,54 @@
+from contextlib import asynccontextmanager
+from datetime import datetime
+from typing import AsyncGenerator
+
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+
+from app.database import init_db
+from app.scheduler import start_scheduler, stop_scheduler, scheduler
+from app.routers import devices, portfolio, recommendations, positions, signals, alerts
+from app.models.schemas import HealthResponse
+
+
+@asynccontextmanager
+async def lifespan(app: FastAPI) -> AsyncGenerator:
+    # Startup
+    init_db()
+    start_scheduler()
+    yield
+    # Shutdown
+    stop_scheduler()
+
+
+app = FastAPI(
+    title="Options Sidekick",
+    description="Covered call and cash-secured put recommendation engine",
+    version="1.0.0",
+    lifespan=lifespan,
+)
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# Routers
+app.include_router(devices.router, prefix="/api/v1")
+app.include_router(portfolio.router, prefix="/api/v1")
+app.include_router(recommendations.router, prefix="/api/v1")
+app.include_router(positions.router, prefix="/api/v1")
+app.include_router(signals.router, prefix="/api/v1")
+app.include_router(alerts.router, prefix="/api/v1")
+
+
+@app.get("/api/v1/health", response_model=HealthResponse, tags=["health"])
+def health():
+    from app.services.position_monitor import last_run
+    return HealthResponse(
+        status="ok",
+        scheduler_running=scheduler.running,
+        last_run=last_run,
+    )

backend/app/models/db_models.py

@@ -0,0 +1,95 @@
+from datetime import datetime, date
+from sqlalchemy import Integer, String, Float, Boolean, DateTime, Date, ForeignKey, UniqueConstraint
+from sqlalchemy.orm import Mapped, mapped_column, relationship
+from app.database import Base
+
+
+class Device(Base):
+    __tablename__ = "devices"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, index=True)
+    apns_token: Mapped[str] = mapped_column(String, unique=True, index=True)
+    device_name: Mapped[str | None] = mapped_column(String, nullable=True)
+    registered_at: Mapped[datetime] = mapped_column(DateTime, default=datetime.utcnow)
+    last_seen: Mapped[datetime] = mapped_column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
+
+    stock_positions: Mapped[list["StockPosition"]] = relationship("StockPosition", back_populates="device", cascade="all, delete-orphan")
+    option_positions: Mapped[list["OptionPosition"]] = relationship("OptionPosition", back_populates="device", cascade="all, delete-orphan")
+    alerts: Mapped[list["Alert"]] = relationship("Alert", back_populates="device", cascade="all, delete-orphan")
+
+
+class StockPosition(Base):
+    __tablename__ = "stock_positions"
+    __table_args__ = (UniqueConstraint("device_id", "ticker", name="uq_device_ticker"),)
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, index=True)
+    device_id: Mapped[int] = mapped_column(Integer, ForeignKey("devices.id"), nullable=False)
+    ticker: Mapped[str] = mapped_column(String, nullable=False)
+    shares: Mapped[int] = mapped_column(Integer, nullable=False)
+    cost_basis: Mapped[float | None] = mapped_column(Float, nullable=True)
+    created_at: Mapped[datetime] = mapped_column(DateTime, default=datetime.utcnow)
+    updated_at: Mapped[datetime] = mapped_column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
+
+    device: Mapped["Device"] = relationship("Device", back_populates="stock_positions")
+
+
+class OptionPosition(Base):
+    __tablename__ = "option_positions"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, index=True)
+    device_id: Mapped[int] = mapped_column(Integer, ForeignKey("devices.id"), nullable=False)
+    ticker: Mapped[str] = mapped_column(String, nullable=False)
+    strategy: Mapped[str] = mapped_column(String, nullable=False)  # covered_call | cash_secured_put
+    strike: Mapped[float] = mapped_column(Float, nullable=False)
+    expiration: Mapped[date] = mapped_column(Date, nullable=False)
+    premium_received: Mapped[float] = mapped_column(Float, nullable=False)
+    contracts: Mapped[int] = mapped_column(Integer, default=1)
+    status: Mapped[str] = mapped_column(String, default="open")  # open | closed | rolled
+    close_reason: Mapped[str | None] = mapped_column(String, nullable=True)  # expired | bought_back | rolled
+    opened_at: Mapped[datetime] = mapped_column(DateTime, default=datetime.utcnow)
+    closed_at: Mapped[datetime | None] = mapped_column(DateTime, nullable=True)
+    last_signal_hash: Mapped[str | None] = mapped_column(String(16), nullable=True)
+
+    device: Mapped["Device"] = relationship("Device", back_populates="option_positions")
+    alerts: Mapped[list["Alert"]] = relationship("Alert", back_populates="option_position")
+
+
+class Recommendation(Base):
+    __tablename__ = "recommendations"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, index=True)
+    device_id: Mapped[int] = mapped_column(Integer, ForeignKey("devices.id"), nullable=False)
+    ticker: Mapped[str] = mapped_column(String, nullable=False)
+    strategy: Mapped[str] = mapped_column(String, nullable=False)  # covered_call | cash_secured_put
+    time_horizon: Mapped[str] = mapped_column(String, nullable=False)  # 0dte | 1dte | weekly | monthly
+    current_price: Mapped[float] = mapped_column(Float, nullable=False)
+    recommended_strike: Mapped[float] = mapped_column(Float, nullable=False)
+    recommended_expiration: Mapped[date] = mapped_column(Date, nullable=False)
+    estimated_premium: Mapped[float] = mapped_column(Float, nullable=False)
+    delta: Mapped[float] = mapped_column(Float, nullable=False)
+    theta: Mapped[float] = mapped_column(Float, nullable=False)
+    iv_rank: Mapped[float] = mapped_column(Float, nullable=False)
+    signal_strength: Mapped[str] = mapped_column(String, nullable=False)  # strong | moderate | weak
+    earnings_warning: Mapped[bool] = mapped_column(Boolean, default=False)
+    earnings_date: Mapped[date | None] = mapped_column(Date, nullable=True)
+    rationale: Mapped[str] = mapped_column(String, nullable=False)
+    signal_hash: Mapped[str] = mapped_column(String(16), nullable=False)
+    created_at: Mapped[datetime] = mapped_column(DateTime, default=datetime.utcnow)
+
+
+class Alert(Base):
+    __tablename__ = "alerts"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, index=True)
+    device_id: Mapped[int] = mapped_column(Integer, ForeignKey("devices.id"), nullable=False)
+    option_position_id: Mapped[int | None] = mapped_column(Integer, ForeignKey("option_positions.id"), nullable=True)
+    ticker: Mapped[str] = mapped_column(String, nullable=False)
+    alert_type: Mapped[str] = mapped_column(String, nullable=False)  # close_early | roll_out | roll_up_down | earnings_warning | new_rec
+    message: Mapped[str] = mapped_column(String, nullable=False)
+    old_signal_hash: Mapped[str | None] = mapped_column(String(16), nullable=True)
+    new_signal_hash: Mapped[str | None] = mapped_column(String(16), nullable=True)
+    sent_at: Mapped[datetime] = mapped_column(DateTime, default=datetime.utcnow)
+    acknowledged: Mapped[bool] = mapped_column(Boolean, default=False)
+
+    device: Mapped["Device"] = relationship("Device", back_populates="alerts")
+    option_position: Mapped["OptionPosition | None"] = relationship("OptionPosition", back_populates="alerts")

backend/app/models/schemas.py

@@ -0,0 +1,166 @@
+from datetime import datetime, date
+from pydantic import BaseModel, field_validator
+
+
+# ─── Device ───────────────────────────────────────────────────────────────────
+
+class DeviceRegister(BaseModel):
+    apns_token: str
+    device_name: str | None = None
+
+
+class DeviceResponse(BaseModel):
+    id: int
+    apns_token: str
+    device_name: str | None
+    registered_at: datetime
+
+    model_config = {"from_attributes": True}
+
+
+# ─── Stock Portfolio ───────────────────────────────────────────────────────────
+
+class StockPositionCreate(BaseModel):
+    ticker: str
+    shares: int
+    cost_basis: float | None = None
+
+    @field_validator("ticker")
+    @classmethod
+    def uppercase_ticker(cls, v: str) -> str:
+        return v.upper().strip()
+
+    @field_validator("shares")
+    @classmethod
+    def positive_shares(cls, v: int) -> int:
+        if v <= 0:
+            raise ValueError("shares must be positive")
+        return v
+
+
+class StockPositionResponse(BaseModel):
+    id: int
+    ticker: str
+    shares: int
+    cost_basis: float | None
+    created_at: datetime
+
+    model_config = {"from_attributes": True}
+
+
+# ─── Option Position ───────────────────────────────────────────────────────────
+
+class OptionPositionCreate(BaseModel):
+    ticker: str
+    strategy: str  # covered_call | cash_secured_put
+    strike: float
+    expiration: date
+    premium_received: float
+    contracts: int = 1
+
+    @field_validator("ticker")
+    @classmethod
+    def uppercase_ticker(cls, v: str) -> str:
+        return v.upper().strip()
+
+    @field_validator("strategy")
+    @classmethod
+    def valid_strategy(cls, v: str) -> str:
+        if v not in ("covered_call", "cash_secured_put"):
+            raise ValueError("strategy must be 'covered_call' or 'cash_secured_put'")
+        return v
+
+    @field_validator("contracts")
+    @classmethod
+    def positive_contracts(cls, v: int) -> int:
+        if v <= 0:
+            raise ValueError("contracts must be positive")
+        return v
+
+
+class OptionPositionClose(BaseModel):
+    status: str  # closed | rolled
+    close_reason: str | None = None
+
+
+class OptionPositionResponse(BaseModel):
+    id: int
+    ticker: str
+    strategy: str
+    strike: float
+    expiration: date
+    premium_received: float
+    contracts: int
+    status: str
+    close_reason: str | None
+    opened_at: datetime
+    closed_at: datetime | None
+    last_signal_hash: str | None
+
+    model_config = {"from_attributes": True}
+
+
+# ─── Signals ──────────────────────────────────────────────────────────────────
+
+class SignalSnapshot(BaseModel):
+    ticker: str
+    current_price: float
+    iv_rank: float
+    sma_50: float
+    sma_200: float
+    nearest_support: float | None
+    nearest_resistance: float | None
+    trend: str  # uptrend | downtrend | sideways
+    earnings_date: date | None
+    computed_at: datetime
+
+
+# ─── Recommendations ──────────────────────────────────────────────────────────
+
+class RecommendationResponse(BaseModel):
+    id: int
+    ticker: str
+    strategy: str
+    time_horizon: str
+    current_price: float
+    recommended_strike: float
+    recommended_expiration: date
+    estimated_premium: float
+    delta: float
+    theta: float
+    iv_rank: float
+    signal_strength: str
+    earnings_warning: bool
+    earnings_date: date | None
+    rationale: str
+    signal_hash: str
+    created_at: datetime
+
+    model_config = {"from_attributes": True}
+
+
+class RecommendationWithSignals(BaseModel):
+    recommendation: RecommendationResponse
+    signals: SignalSnapshot
+
+
+# ─── Alerts ───────────────────────────────────────────────────────────────────
+
+class AlertResponse(BaseModel):
+    id: int
+    ticker: str
+    option_position_id: int | None
+    alert_type: str
+    message: str
+    sent_at: datetime
+    acknowledged: bool
+
+    model_config = {"from_attributes": True}
+
+
+# ─── Health ───────────────────────────────────────────────────────────────────
+
+class HealthResponse(BaseModel):
+    status: str
+    scheduler_running: bool
+    last_run: datetime | None

backend/app/routers/alerts.py

@@ -0,0 +1,46 @@
+from fastapi import APIRouter, Depends, Header, HTTPException, Query
+from sqlalchemy.orm import Session
+
+from app.database import get_db
+from app.models.db_models import Device, Alert
+from app.models.schemas import AlertResponse
+
+router = APIRouter(prefix="/alerts", tags=["alerts"])
+
+
+def _get_device(x_device_token: str = Header(...), db: Session = Depends(get_db)) -> Device:
+    device = db.query(Device).filter(Device.apns_token == x_device_token).first()
+    if not device:
+        raise HTTPException(status_code=404, detail="Device not registered.")
+    return device
+
+
+@router.get("", response_model=list[AlertResponse])
+def get_alerts(
+    unread_only: bool = Query(False),
+    device: Device = Depends(_get_device),
+    db: Session = Depends(get_db),
+):
+    query = db.query(Alert).filter(Alert.device_id == device.id)
+    if unread_only:
+        query = query.filter(Alert.acknowledged == False)  # noqa: E712
+    return query.order_by(Alert.sent_at.desc()).limit(100).all()
+
+
+@router.patch("/{alert_id}/acknowledge", response_model=AlertResponse)
+def acknowledge_alert(
+    alert_id: int,
+    device: Device = Depends(_get_device),
+    db: Session = Depends(get_db),
+):
+    alert = (
+        db.query(Alert)
+        .filter(Alert.id == alert_id, Alert.device_id == device.id)
+        .first()
+    )
+    if not alert:
+        raise HTTPException(status_code=404, detail="Alert not found.")
+    alert.acknowledged = True
+    db.commit()
+    db.refresh(alert)
+    return alert

backend/app/routers/devices.py

@@ -0,0 +1,29 @@
+from datetime import datetime
+from fastapi import APIRouter, Depends, HTTPException
+from sqlalchemy.orm import Session
+
+from app.database import get_db
+from app.models.db_models import Device
+from app.models.schemas import DeviceRegister, DeviceResponse
+
+router = APIRouter(prefix="/devices", tags=["devices"])
+
+
+@router.post("/register", response_model=DeviceResponse)
+def register_device(body: DeviceRegister, db: Session = Depends(get_db)):
+    """Register or refresh an APNs device token."""
+    device = db.query(Device).filter(Device.apns_token == body.apns_token).first()
+
+    if device:
+        device.device_name = body.device_name or device.device_name
+        device.last_seen = datetime.utcnow()
+    else:
+        device = Device(
+            apns_token=body.apns_token,
+            device_name=body.device_name,
+        )
+        db.add(device)
+
+    db.commit()
+    db.refresh(device)
+    return device

backend/app/routers/portfolio.py

@@ -0,0 +1,64 @@
+from fastapi import APIRouter, Depends, Header, HTTPException
+from sqlalchemy.orm import Session
+
+from app.database import get_db
+from app.models.db_models import Device, StockPosition
+from app.models.schemas import StockPositionCreate, StockPositionResponse
+
+router = APIRouter(prefix="/portfolio", tags=["portfolio"])
+
+
+def _get_device(x_device_token: str = Header(...), db: Session = Depends(get_db)) -> Device:
+    device = db.query(Device).filter(Device.apns_token == x_device_token).first()
+    if not device:
+        raise HTTPException(status_code=404, detail="Device not registered. Call /devices/register first.")
+    return device
+
+
+@router.get("", response_model=list[StockPositionResponse])
+def get_portfolio(device: Device = Depends(_get_device), db: Session = Depends(get_db)):
+    return db.query(StockPosition).filter(StockPosition.device_id == device.id).all()
+
+
+@router.post("", response_model=list[StockPositionResponse])
+def set_portfolio(
+    positions: list[StockPositionCreate],
+    device: Device = Depends(_get_device),
+    db: Session = Depends(get_db),
+):
+    """Full replace — client sends complete list of stock holdings."""
+    # Delete all existing for this device
+    db.query(StockPosition).filter(StockPosition.device_id == device.id).delete()
+
+    new_positions = []
+    for p in positions:
+        sp = StockPosition(
+            device_id=device.id,
+            ticker=p.ticker,
+            shares=p.shares,
+            cost_basis=p.cost_basis,
+        )
+        db.add(sp)
+        new_positions.append(sp)
+
+    db.commit()
+    for sp in new_positions:
+        db.refresh(sp)
+    return new_positions
+
+
+@router.delete("/{ticker}", status_code=204)
+def delete_ticker(
+    ticker: str,
+    device: Device = Depends(_get_device),
+    db: Session = Depends(get_db),
+):
+    ticker = ticker.upper()
+    deleted = (
+        db.query(StockPosition)
+        .filter(StockPosition.device_id == device.id, StockPosition.ticker == ticker)
+        .delete()
+    )
+    if not deleted:
+        raise HTTPException(status_code=404, detail=f"Ticker {ticker} not in portfolio")
+    db.commit()

backend/app/routers/positions.py

@@ -0,0 +1,78 @@
+from datetime import datetime
+from fastapi import APIRouter, Depends, Header, HTTPException, Query
+from sqlalchemy.orm import Session
+
+from app.database import get_db
+from app.models.db_models import Device, OptionPosition
+from app.models.schemas import OptionPositionCreate, OptionPositionClose, OptionPositionResponse
+
+router = APIRouter(prefix="/positions", tags=["positions"])
+
+
+def _get_device(x_device_token: str = Header(...), db: Session = Depends(get_db)) -> Device:
+    device = db.query(Device).filter(Device.apns_token == x_device_token).first()
+    if not device:
+        raise HTTPException(status_code=404, detail="Device not registered.")
+    return device
+
+
+@router.get("", response_model=list[OptionPositionResponse])
+def get_positions(
+    status: str | None = Query(None),
+    device: Device = Depends(_get_device),
+    db: Session = Depends(get_db),
+):
+    query = db.query(OptionPosition).filter(OptionPosition.device_id == device.id)
+    if status:
+        query = query.filter(OptionPosition.status == status)
+    return query.order_by(OptionPosition.opened_at.desc()).all()
+
+
+@router.post("", response_model=OptionPositionResponse, status_code=201)
+def log_position(
+    body: OptionPositionCreate,
+    device: Device = Depends(_get_device),
+    db: Session = Depends(get_db),
+):
+    position = OptionPosition(
+        device_id=device.id,
+        ticker=body.ticker,
+        strategy=body.strategy,
+        strike=body.strike,
+        expiration=body.expiration,
+        premium_received=body.premium_received,
+        contracts=body.contracts,
+        status="open",
+    )
+    db.add(position)
+    db.commit()
+    db.refresh(position)
+    return position
+
+
+@router.patch("/{position_id}", response_model=OptionPositionResponse)
+def close_position(
+    position_id: int,
+    body: OptionPositionClose,
+    device: Device = Depends(_get_device),
+    db: Session = Depends(get_db),
+):
+    position = (
+        db.query(OptionPosition)
+        .filter(OptionPosition.id == position_id, OptionPosition.device_id == device.id)
+        .first()
+    )
+    if not position:
+        raise HTTPException(status_code=404, detail="Position not found.")
+
+    valid_statuses = ("closed", "rolled")
+    if body.status not in valid_statuses:
+        raise HTTPException(status_code=422, detail=f"status must be one of {valid_statuses}")
+
+    position.status = body.status
+    position.close_reason = body.close_reason
+    position.closed_at = datetime.utcnow()
+
+    db.commit()
+    db.refresh(position)
+    return position

backend/app/routers/recommendations.py

@@ -0,0 +1,141 @@
+from datetime import datetime, timedelta
+from fastapi import APIRouter, Depends, Header, HTTPException, Query
+from sqlalchemy.orm import Session
+
+from app.database import get_db
+from app.models.db_models import Device, StockPosition, Recommendation
+from app.models.schemas import RecommendationResponse, RecommendationWithSignals
+from app.services.signal_engine import compute_signals
+from app.services.recommendation_engine import build_recommendation
+from app.config import settings
+
+router = APIRouter(prefix="/recommendations", tags=["recommendations"])
+
+
+def _get_device(x_device_token: str = Header(...), db: Session = Depends(get_db)) -> Device:
+    device = db.query(Device).filter(Device.apns_token == x_device_token).first()
+    if not device:
+        raise HTTPException(status_code=404, detail="Device not registered.")
+    return device
+
+
+@router.get("", response_model=list[RecommendationResponse])
+def get_recommendations(
+    time_horizon: str | None = Query(None),
+    device: Device = Depends(_get_device),
+    db: Session = Depends(get_db),
+):
+    """Return latest cached recommendations for all portfolio tickers."""
+    query = db.query(Recommendation).filter(Recommendation.device_id == device.id)
+    if time_horizon:
+        query = query.filter(Recommendation.time_horizon == time_horizon)
+    return query.order_by(Recommendation.created_at.desc()).all()
+
+
+@router.get("/{ticker}", response_model=RecommendationWithSignals)
+def get_recommendation_for_ticker(
+    ticker: str,
+    time_horizon: str = Query("weekly"),
+    strategy: str = Query("covered_call"),
+    device: Device = Depends(_get_device),
+    db: Session = Depends(get_db),
+):
+    """Return fresh recommendation + signals for a specific ticker."""
+    ticker = ticker.upper()
+
+    snap = compute_signals(ticker)
+    if snap is None:
+        raise HTTPException(status_code=503, detail=f"Could not fetch market data for {ticker}")
+
+    rec = build_recommendation(
+        device_id=device.id,
+        ticker=ticker,
+        strategy=strategy,
+        time_horizon=time_horizon,
+        snapshot=snap,
+    )
+    if rec is None:
+        raise HTTPException(
+            status_code=404,
+            detail=f"No qualifying options found for {ticker} {strategy} {time_horizon}",
+        )
+
+    # Persist recommendation
+    existing = (
+        db.query(Recommendation)
+        .filter(
+            Recommendation.device_id == device.id,
+            Recommendation.ticker == ticker,
+            Recommendation.strategy == strategy,
+            Recommendation.time_horizon == time_horizon,
+        )
+        .first()
+    )
+    if existing:
+        db.delete(existing)
+    db.add(rec)
+    db.commit()
+    db.refresh(rec)
+
+    return RecommendationWithSignals(recommendation=RecommendationResponse.model_validate(rec), signals=snap)
+
+
+@router.post("/refresh", response_model=list[RecommendationResponse])
+def refresh_recommendations(
+    device: Device = Depends(_get_device),
+    db: Session = Depends(get_db),
+):
+    """
+    On-demand recalculation for all portfolio tickers.
+    Throttled: no-ops if last refresh was less than THROTTLE_MINUTES ago.
+    """
+    throttle = timedelta(minutes=settings.recommendation_throttle_minutes)
+    most_recent = (
+        db.query(Recommendation)
+        .filter(Recommendation.device_id == device.id)
+        .order_by(Recommendation.created_at.desc())
+        .first()
+    )
+    if most_recent and (datetime.utcnow() - most_recent.created_at) < throttle:
+        return db.query(Recommendation).filter(Recommendation.device_id == device.id).all()
+
+    stock_positions = db.query(StockPosition).filter(StockPosition.device_id == device.id).all()
+    if not stock_positions:
+        return []
+
+    results = []
+    for sp in stock_positions:
+        snap = compute_signals(sp.ticker)
+        if snap is None:
+            continue
+        for strategy in ("covered_call", "cash_secured_put"):
+            for horizon in ("weekly", "monthly"):
+                rec = build_recommendation(
+                    device_id=device.id,
+                    ticker=sp.ticker,
+                    strategy=strategy,
+                    time_horizon=horizon,
+                    snapshot=snap,
+                )
+                if rec is None:
+                    continue
+
+                existing = (
+                    db.query(Recommendation)
+                    .filter(
+                        Recommendation.device_id == device.id,
+                        Recommendation.ticker == sp.ticker,
+                        Recommendation.strategy == strategy,
+                        Recommendation.time_horizon == horizon,
+                    )
+                    .first()
+                )
+                if existing:
+                    db.delete(existing)
+                db.add(rec)
+                results.append(rec)
+
+    db.commit()
+    for r in results:
+        db.refresh(r)
+    return results

backend/app/routers/signals.py

@@ -0,0 +1,14 @@
+from fastapi import APIRouter, HTTPException
+from app.models.schemas import SignalSnapshot
+from app.services.signal_engine import compute_signals
+
+router = APIRouter(prefix="/signals", tags=["signals"])
+
+
+@router.get("/{ticker}", response_model=SignalSnapshot)
+def get_signals(ticker: str):
+    ticker = ticker.upper()
+    snap = compute_signals(ticker)
+    if snap is None:
+        raise HTTPException(status_code=503, detail=f"Could not fetch market data for {ticker}")
+    return snap

backend/app/scheduler.py

@@ -0,0 +1,38 @@
+"""
+scheduler.py — APScheduler configuration.
+
+The scheduler is started in main.py's lifespan event.
+"""
+
+import asyncio
+import logging
+
+from apscheduler.schedulers.asyncio import AsyncIOScheduler
+from apscheduler.triggers.interval import IntervalTrigger
+
+from app.config import settings
+
+logger = logging.getLogger(__name__)
+
+scheduler = AsyncIOScheduler()
+
+
+def start_scheduler():
+    from app.services.position_monitor import monitor_all_positions
+
+    scheduler.add_job(
+        monitor_all_positions,
+        trigger=IntervalTrigger(seconds=settings.monitor_interval_seconds),
+        id="position_monitor",
+        name="Position Monitor",
+        replace_existing=True,
+        misfire_grace_time=60,
+    )
+    scheduler.start()
+    logger.info(f"Scheduler started — monitor interval: {settings.monitor_interval_seconds}s")
+
+
+def stop_scheduler():
+    if scheduler.running:
+        scheduler.shutdown(wait=False)
+        logger.info("Scheduler stopped")

backend/app/services/apns_service.py

@@ -0,0 +1,124 @@
+"""
+apns_service.py — Send push notifications via APNs HTTP/2.
+
+Uses JWT authentication with a .p8 key.  JWT is cached for 50 minutes
+and auto-renewed before the 60-minute APNs expiry.
+"""
+
+import json
+import logging
+import time
+from datetime import datetime, timedelta
+from typing import Optional
+
+import httpx
+import jwt as pyjwt
+
+from app.config import settings
+
+logger = logging.getLogger(__name__)
+
+_APNS_HOST_PROD = "https://api.push.apple.com"
+_APNS_HOST_SANDBOX = "https://api.sandbox.push.apple.com"
+
+# Cached JWT state
+_jwt_token: Optional[str] = None
+_jwt_issued_at: Optional[float] = None
+_JWT_TTL_SECONDS = 50 * 60  # 50 minutes — renew before Apple's 60-min limit
+
+
+def _get_apns_host() -> str:
+    return _APNS_HOST_SANDBOX if settings.apns_use_sandbox else _APNS_HOST_PROD
+
+
+def _load_private_key() -> str:
+    try:
+        with open(settings.apns_key_path, "r") as f:
+            return f.read()
+    except FileNotFoundError:
+        logger.warning(f"APNs key not found at {settings.apns_key_path} — push disabled")
+        return ""
+
+
+def _get_jwt() -> Optional[str]:
+    global _jwt_token, _jwt_issued_at
+
+    now = time.time()
+    if _jwt_token and _jwt_issued_at and (now - _jwt_issued_at) < _JWT_TTL_SECONDS:
+        return _jwt_token
+
+    private_key = _load_private_key()
+    if not private_key:
+        return None
+
+    if not settings.apns_key_id or not settings.apns_team_id:
+        logger.warning("APNs key ID or team ID not configured — push disabled")
+        return None
+
+    payload = {
+        "iss": settings.apns_team_id,
+        "iat": int(now),
+    }
+    headers = {
+        "alg": "ES256",
+        "kid": settings.apns_key_id,
+    }
+
+    try:
+        token = pyjwt.encode(payload, private_key, algorithm="ES256", headers=headers)
+        _jwt_token = token if isinstance(token, str) else token.decode("utf-8")
+        _jwt_issued_at = now
+        return _jwt_token
+    except Exception as e:
+        logger.error(f"Failed to generate APNs JWT: {e}")
+        return None
+
+
+async def send_push(
+    apns_token: str,
+    title: str,
+    body: str,
+    payload: Optional[dict] = None,
+) -> bool:
+    """
+    Send a push notification to a device.
+    Returns True on success, False on failure.
+    """
+    jwt_token = _get_jwt()
+    if not jwt_token:
+        logger.warning("APNs push skipped — no valid JWT (check .p8 key config)")
+        return False
+
+    apns_payload = {
+        "aps": {
+            "alert": {"title": title, "body": body},
+            "badge": 1,
+            "sound": "default",
+            "category": "POSITION_ALERT",
+        }
+    }
+    if payload:
+        apns_payload.update(payload)
+
+    url = f"{_get_apns_host()}/3/device/{apns_token}"
+    headers = {
+        "authorization": f"bearer {jwt_token}",
+        "apns-topic": settings.apns_bundle_id,
+        "apns-push-type": "alert",
+        "content-type": "application/json",
+    }
+
+    try:
+        async with httpx.AsyncClient(http2=True, timeout=10.0) as client:
+            response = await client.post(url, headers=headers, content=json.dumps(apns_payload))
+
+        if response.status_code == 200:
+            logger.info(f"Push sent to {apns_token[:8]}... — {title}")
+            return True
+        else:
+            logger.error(f"APNs rejected push: {response.status_code} {response.text}")
+            return False
+
+    except Exception as e:
+        logger.error(f"APNs push failed: {e}")
+        return False

backend/app/services/market_data.py

@@ -0,0 +1,198 @@
+"""
+market_data.py — yfinance wrapper with 15-minute in-memory cache.
+
+All other services call through this module.  Never import yfinance directly
+outside of here — keeps caching and error handling centralized.
+"""
+
+import logging
+from datetime import datetime, date, timedelta
+from functools import lru_cache
+from typing import Optional
+import time
+
+import pandas as pd
+import yfinance as yf
+
+logger = logging.getLogger(__name__)
+
+# ─── Simple time-bucketed cache ───────────────────────────────────────────────
+# Key: (ticker, cache_bucket)  where cache_bucket = int(time() // 900)
+_CACHE: dict = {}
+_CACHE_TTL_SECONDS = 900  # 15 minutes
+
+
+def _bucket() -> int:
+    return int(time.time() // _CACHE_TTL_SECONDS)
+
+
+def _cache_get(key: tuple):
+    return _CACHE.get((key, _bucket()))
+
+
+def _cache_set(key: tuple, value):
+    _CACHE[(key, _bucket())] = value
+    # Prune old entries to avoid unbounded growth
+    current = _bucket()
+    stale = [k for k in list(_CACHE) if k[1] < current - 1]
+    for k in stale:
+        del _CACHE[k]
+
+
+# ─── Price history ─────────────────────────────────────────────────────────────
+
+def get_price_history(ticker: str, period: str = "1y") -> Optional[pd.DataFrame]:
+    """
+    Return OHLCV DataFrame for the past year.
+    Columns: Open, High, Low, Close, Volume
+    Index: DatetimeIndex
+    Returns None on failure.
+    """
+    cache_key = ("price_history", ticker, period)
+    cached = _cache_get(cache_key)
+    if cached is not None:
+        return cached
+
+    try:
+        df = yf.download(ticker, period=period, auto_adjust=True, progress=False)
+        if df.empty:
+            logger.warning(f"Empty price history for {ticker}")
+            return None
+        # Flatten multi-level columns if present (yfinance sometimes returns them)
+        if isinstance(df.columns, pd.MultiIndex):
+            df.columns = df.columns.get_level_values(0)
+        _cache_set(cache_key, df)
+        return df
+    except Exception as e:
+        logger.error(f"Failed to fetch price history for {ticker}: {e}")
+        return None
+
+
+# ─── Current price ─────────────────────────────────────────────────────────────
+
+def get_current_price(ticker: str) -> Optional[float]:
+    cache_key = ("current_price", ticker)
+    cached = _cache_get(cache_key)
+    if cached is not None:
+        return cached
+
+    try:
+        t = yf.Ticker(ticker)
+        price = t.fast_info.get("last_price") or t.fast_info.get("previousClose")
+        if price is None:
+            hist = get_price_history(ticker, period="5d")
+            if hist is not None and not hist.empty:
+                price = float(hist["Close"].iloc[-1])
+        if price is not None:
+            price = float(price)
+            _cache_set(cache_key, price)
+        return price
+    except Exception as e:
+        logger.error(f"Failed to fetch current price for {ticker}: {e}")
+        return None
+
+
+# ─── Options chain ─────────────────────────────────────────────────────────────
+
+def get_option_expirations(ticker: str) -> list[str]:
+    """Return list of available expiration date strings (YYYY-MM-DD)."""
+    cache_key = ("expirations", ticker)
+    cached = _cache_get(cache_key)
+    if cached is not None:
+        return cached
+
+    try:
+        t = yf.Ticker(ticker)
+        expirations = list(t.options)
+        _cache_set(cache_key, expirations)
+        return expirations
+    except Exception as e:
+        logger.error(f"Failed to fetch expirations for {ticker}: {e}")
+        return []
+
+
+def get_options_chain(ticker: str, expiration: str) -> Optional[dict]:
+    """
+    Return {'calls': DataFrame, 'puts': DataFrame} for the given expiry.
+    Columns include: strike, lastPrice, bid, ask, volume, openInterest,
+                     impliedVolatility, delta, theta, gamma (where available).
+    Returns None on failure.
+    """
+    cache_key = ("options_chain", ticker, expiration)
+    cached = _cache_get(cache_key)
+    if cached is not None:
+        return cached
+
+    try:
+        t = yf.Ticker(ticker)
+        chain = t.option_chain(expiration)
+        result = {"calls": chain.calls.copy(), "puts": chain.puts.copy()}
+        # yfinance does not always include Greeks — add NaN columns if missing
+        for df in result.values():
+            for col in ("delta", "theta", "gamma"):
+                if col not in df.columns:
+                    df[col] = float("nan")
+        _cache_set(cache_key, result)
+        return result
+    except Exception as e:
+        logger.error(f"Failed to fetch options chain for {ticker} exp={expiration}: {e}")
+        return None
+
+
+def get_nearest_expiry(ticker: str, target_date: date) -> Optional[str]:
+    """Return the closest available expiry on or after target_date."""
+    expirations = get_option_expirations(ticker)
+    if not expirations:
+        return None
+    target_str = str(target_date)
+    future = [e for e in sorted(expirations) if e >= target_str]
+    return future[0] if future else None
+
+
+def get_same_day_expiry(ticker: str) -> Optional[str]:
+    """Return today's expiry string if it exists, else None."""
+    today_str = str(date.today())
+    expirations = get_option_expirations(ticker)
+    return today_str if today_str in expirations else None
+
+
+# ─── Earnings / dividends ──────────────────────────────────────────────────────
+
+def get_earnings_date(ticker: str) -> Optional[date]:
+    """Return the next earnings date or None."""
+    cache_key = ("earnings", ticker)
+    cached = _cache_get(cache_key)
+    if cached is not None:
+        return cached
+
+    try:
+        t = yf.Ticker(ticker)
+        cal = t.calendar
+        if cal is None:
+            return None
+
+        # calendar can be a dict or DataFrame depending on yfinance version
+        if isinstance(cal, dict):
+            earnings_info = cal.get("Earnings Date", [])
+        else:
+            # DataFrame — look for "Earnings Date" row or column
+            try:
+                earnings_info = cal.loc["Earnings Date"].tolist()
+            except Exception:
+                earnings_info = []
+
+        result = None
+        for item in (earnings_info if isinstance(earnings_info, list) else [earnings_info]):
+            try:
+                d = pd.Timestamp(item).date()
+                if d >= date.today():
+                    result = d
+                    break
+            except Exception:
+                continue
+
+        _cache_set(cache_key, result)
+        return result
+    except Exception as e:
+        logger.error(f"Failed to fetch earnings date for {ticker}: {e}")
+        return None

backend/app/services/position_monitor.py

@@ -0,0 +1,331 @@
+"""
+position_monitor.py — 15-minute job that re-evaluates all open option positions
+and fires alerts when signals change materially.
+
+Also refreshes recommendations for all stock positions.
+"""
+
+import logging
+from datetime import datetime, date
+from typing import Optional
+
+from sqlalchemy.orm import Session
+
+from app.database import SessionLocal
+from app.models.db_models import OptionPosition, StockPosition, Recommendation, Alert, Device
+from app.services import market_data as md
+from app.services.signal_engine import (
+    compute_iv_rank,
+    compute_smas,
+    compute_swing_levels,
+    compute_trend,
+    compute_signal_strength,
+    compute_signal_hash,
+)
+from app.services.recommendation_engine import build_recommendation
+from app.services.apns_service import send_push
+from app.models.schemas import SignalSnapshot
+
+logger = logging.getLogger(__name__)
+
+# Tracks the last time this job ran
+last_run: Optional[datetime] = None
+
+
+def _determine_alert_type(
+    position: OptionPosition,
+    current_delta: float,
+    new_signal_hash: str,
+    new_rec: Optional[Recommendation],
+    earnings_warning: bool,
+) -> Optional[str]:
+    """
+    Determine if and what type of alert to fire.
+    Returns alert_type string or None.
+    """
+    # Earnings warning newly triggered
+    if earnings_warning and not _position_had_earnings_warning(position):
+        return "earnings_warning"
+
+    # Deep ITM — delta threshold
+    abs_delta = abs(current_delta)
+    if abs_delta >= 0.45:
+        return "close_early"
+
+    # Profit capture — if premium has decayed significantly
+    # (We don't track current price here, but high delta ITM is a proxy)
+    if abs_delta >= 0.40:
+        return "close_early"
+
+    # Expiry-based recommendation changed
+    if new_rec:
+        days_to_expiry = (position.expiration - date.today()).days
+        if days_to_expiry <= 5 and abs_delta <= 0.10:
+            return "close_early"  # expiring nearly worthless — take it off
+
+        # Roll suggestion: new recommendation is for a further-out expiry
+        if new_rec.recommended_expiration > position.expiration:
+            return "roll_out"
+
+        # Strike meaningfully different (more than 2 strikes, roughly $2-5 depending on underlying)
+        if abs(new_rec.recommended_strike - position.strike) / position.strike > 0.02:
+            return "roll_up_down"
+
+    return None
+
+
+def _position_had_earnings_warning(position: OptionPosition) -> bool:
+    """Best-effort: check if earnings warning was already flagged on last hash."""
+    # We encode earnings_warning in the hash payload so if it was True before
+    # the hash would already reflect it. This is a simple flag check.
+    return False  # Simplified — the hash change will trigger the alert naturally
+
+
+async def monitor_all_positions():
+    """Main scheduler job. Runs every 15 minutes."""
+    global last_run
+    logger.info("Position monitor: starting run")
+
+    db: Session = SessionLocal()
+    try:
+        # 1. Get all open positions grouped by ticker to batch data fetching
+        open_positions: list[OptionPosition] = (
+            db.query(OptionPosition)
+            .filter(OptionPosition.status == "open")
+            .all()
+        )
+
+        tickers_to_check = list({p.ticker for p in open_positions})
+
+        # Also collect all stock positions to refresh recommendations
+        stock_positions: list[StockPosition] = db.query(StockPosition).all()
+        stock_tickers = list({sp.ticker for sp in stock_positions})
+
+        all_tickers = list(set(tickers_to_check + stock_tickers))
+
+        if not all_tickers:
+            logger.info("Position monitor: no tickers to check")
+            last_run = datetime.utcnow()
+            return
+
+        # 2. Pre-fetch market data for all tickers (cached by market_data module)
+        logger.info(f"Position monitor: checking {len(all_tickers)} tickers")
+        signal_snapshots: dict[str, Optional[SignalSnapshot]] = {}
+        for ticker in all_tickers:
+            snap = _compute_snapshot(ticker)
+            signal_snapshots[ticker] = snap
+
+        # 3. Evaluate each open option position
+        for position in open_positions:
+            snap = signal_snapshots.get(position.ticker)
+            if snap is None:
+                continue
+
+            await _evaluate_position(db, position, snap)
+
+        # 4. Refresh recommendations for all stock positions
+        for stock_pos in stock_positions:
+            snap = signal_snapshots.get(stock_pos.ticker)
+            if snap is None:
+                continue
+            _refresh_recommendations(db, stock_pos.device_id, stock_pos.ticker, snap)
+
+        db.commit()
+        last_run = datetime.utcnow()
+        logger.info("Position monitor: run complete")
+
+    except Exception as e:
+        logger.error(f"Position monitor error: {e}", exc_info=True)
+        db.rollback()
+    finally:
+        db.close()
+
+
+def _compute_snapshot(ticker: str) -> Optional[SignalSnapshot]:
+    """Build signal snapshot from market data."""
+    import math
+    df = md.get_price_history(ticker)
+    if df is None:
+        return None
+    current_price = md.get_current_price(ticker)
+    if current_price is None:
+        return None
+
+    iv_rank = compute_iv_rank(df)
+    smas = compute_smas(df)
+    swing = compute_swing_levels(df)
+    trend = compute_trend(current_price, smas["sma_50"], smas["sma_200"])
+    earnings_date = md.get_earnings_date(ticker)
+
+    return SignalSnapshot(
+        ticker=ticker,
+        current_price=current_price,
+        iv_rank=iv_rank,
+        sma_50=smas["sma_50"] if not math.isnan(smas["sma_50"]) else 0.0,
+        sma_200=smas["sma_200"] if not math.isnan(smas["sma_200"]) else 0.0,
+        nearest_support=swing["nearest_support"],
+        nearest_resistance=swing["nearest_resistance"],
+        trend=trend,
+        earnings_date=earnings_date,
+        computed_at=datetime.utcnow(),
+    )
+
+
+async def _evaluate_position(db: Session, position: OptionPosition, snap: SignalSnapshot):
+    """Re-evaluate one open position and fire an alert if signal changed."""
+    # Get current option data for this specific strike/expiry
+    expiry_str = str(position.expiration)
+    chain = md.get_options_chain(position.ticker, expiry_str)
+
+    current_delta = 0.25  # fallback
+    if chain:
+        chain_df = chain["calls"] if position.strategy == "covered_call" else chain["puts"]
+        row = chain_df[chain_df["strike"] == position.strike]
+        if not row.empty and "delta" in row.columns:
+            delta_val = row["delta"].iloc[0]
+            if delta_val == delta_val:  # not NaN
+                current_delta = abs(float(delta_val))
+
+    # Build a fresh new recommendation to compare expiry/strike
+    new_rec = build_recommendation(
+        device_id=position.device_id,
+        ticker=position.ticker,
+        strategy=position.strategy,
+        time_horizon="weekly",  # use weekly for monitoring comparisons
+        snapshot=snap,
+    )
+
+    earnings_warning = bool(snap.earnings_date and snap.earnings_date <= position.expiration)
+
+    new_hash = compute_signal_hash(
+        iv_rank=snap.iv_rank,
+        sma_50=snap.sma_50,
+        sma_200=snap.sma_200,
+        nearest_support=snap.nearest_support,
+        nearest_resistance=snap.nearest_resistance,
+        recommended_strike=new_rec.recommended_strike if new_rec else position.strike,
+        recommended_expiration=new_rec.recommended_expiration if new_rec else position.expiration,
+        earnings_warning=earnings_warning,
+    )
+
+    # No change — skip
+    if new_hash == position.last_signal_hash:
+        return
+
+    # Determine alert type
+    alert_type = _determine_alert_type(position, current_delta, new_hash, new_rec, earnings_warning)
+    if alert_type is None:
+        # Still update the hash even if no actionable alert
+        position.last_signal_hash = new_hash
+        return
+
+    # Build alert message
+    message = _build_alert_message(position, alert_type, current_delta, snap, new_rec)
+
+    # Save alert to DB
+    alert = Alert(
+        device_id=position.device_id,
+        option_position_id=position.id,
+        ticker=position.ticker,
+        alert_type=alert_type,
+        message=message,
+        old_signal_hash=position.last_signal_hash,
+        new_signal_hash=new_hash,
+        sent_at=datetime.utcnow(),
+        acknowledged=False,
+    )
+    db.add(alert)
+
+    # Update position hash
+    position.last_signal_hash = new_hash
+
+    # Send push notification
+    device: Optional[Device] = db.query(Device).filter(Device.id == position.device_id).first()
+    if device:
+        strategy_label = "Covered Call" if position.strategy == "covered_call" else "Cash-Secured Put"
+        await send_push(
+            apns_token=device.apns_token,
+            title=f"{position.ticker} {strategy_label} — Action Needed",
+            body=message,
+            payload={
+                "alert_type": alert_type,
+                "ticker": position.ticker,
+                "position_id": position.id,
+            },
+        )
+
+    logger.info(f"Alert fired: {position.ticker} {alert_type} — {message[:60]}")
+
+
+def _build_alert_message(
+    position: OptionPosition,
+    alert_type: str,
+    current_delta: float,
+    snap: SignalSnapshot,
+    new_rec: Optional[Recommendation],
+) -> str:
+    strike = position.strike
+    expiry = position.expiration
+
+    if alert_type == "close_early":
+        if current_delta >= 0.45:
+            return (
+                f"Delta has risen to {current_delta:.2f} — position is deep ITM. "
+                f"Consider closing early to limit risk on the ${strike:.0f} strike expiring {expiry}."
+            )
+        return (
+            f"Signals suggest closing early on ${strike:.0f} strike expiring {expiry}. "
+            f"Capturing premium now may be optimal."
+        )
+
+    if alert_type == "roll_out" and new_rec:
+        return (
+            f"Consider rolling your ${strike:.0f} strike out to {new_rec.recommended_expiration} "
+            f"at ${new_rec.recommended_strike:.0f} for ${new_rec.estimated_premium:.2f} credit."
+        )
+
+    if alert_type == "roll_up_down" and new_rec:
+        direction = "up" if new_rec.recommended_strike > strike else "down"
+        return (
+            f"Signals favor rolling {direction} from ${strike:.0f} to ${new_rec.recommended_strike:.0f} "
+            f"at {new_rec.recommended_expiration} for ${new_rec.estimated_premium:.2f} credit."
+        )
+
+    if alert_type == "earnings_warning":
+        return (
+            f"⚠️ Earnings date {snap.earnings_date} now falls within your expiry on {expiry}. "
+            f"Consider closing or rolling before earnings."
+        )
+
+    return f"Signal change detected on {position.ticker} ${strike:.0f} strike. Review your position."
+
+
+def _refresh_recommendations(db: Session, device_id: int, ticker: str, snap: SignalSnapshot):
+    """Rebuild and save latest recommendations for a ticker."""
+    for strategy in ("covered_call", "cash_secured_put"):
+        for horizon in ("weekly", "monthly"):
+            rec = build_recommendation(
+                device_id=device_id,
+                ticker=ticker,
+                strategy=strategy,
+                time_horizon=horizon,
+                snapshot=snap,
+            )
+            if rec is None:
+                continue
+
+            # Replace existing recommendation for same device/ticker/strategy/horizon
+            existing = (
+                db.query(Recommendation)
+                .filter(
+                    Recommendation.device_id == device_id,
+                    Recommendation.ticker == ticker,
+                    Recommendation.strategy == strategy,
+                    Recommendation.time_horizon == horizon,
+                )
+                .first()
+            )
+            if existing:
+                db.delete(existing)
+
+            db.add(rec)

backend/app/services/recommendation_engine.py

@@ -0,0 +1,299 @@
+"""
+recommendation_engine.py — Select optimal strike and expiry for a given strategy/horizon.
+
+For each (ticker, strategy, time_horizon) combination:
+  1. Determine the target expiration date
+  2. Fetch the options chain for that expiry
+  3. Filter by delta range (0.20-0.30 for CC/CSP)
+  4. Among qualifying strikes, pick highest mid-price premium
+  5. Build a Recommendation DB row
+"""
+
+import logging
+import math
+from datetime import date, datetime
+from typing import Optional
+
+import pandas as pd
+
+from app.models.db_models import Recommendation
+from app.models.schemas import SignalSnapshot
+from app.services import market_data as md
+from app.services.signal_engine import (
+    compute_iv_rank,
+    compute_smas,
+    compute_swing_levels,
+    compute_trend,
+    compute_signal_strength,
+    compute_signal_hash,
+)
+from app.utils.date_helpers import (
+    next_trading_day,
+    next_friday,
+    nearest_monthly_expiry,
+    within_dte_window_for_0dte,
+)
+
+logger = logging.getLogger(__name__)
+
+# Delta target ranges for short premium selling
+DELTA_MIN = 0.18
+DELTA_MAX = 0.35
+
+
+def _target_expiry(ticker: str, time_horizon: str) -> Optional[str]:
+    """
+    Return the best available expiry string for the given time horizon.
+    Returns None if no suitable expiry exists.
+    """
+    today = date.today()
+
+    if time_horizon == "0dte":
+        if not within_dte_window_for_0dte():
+            return None
+        expiry = md.get_same_day_expiry(ticker)
+        return expiry  # None if today has no options
+
+    if time_horizon == "1dte":
+        target = next_trading_day(today)
+        return md.get_nearest_expiry(ticker, target)
+
+    if time_horizon == "weekly":
+        target = next_friday(today)
+        return md.get_nearest_expiry(ticker, target)
+
+    if time_horizon == "monthly":
+        target = nearest_monthly_expiry(today, target_dte=30)
+        return md.get_nearest_expiry(ticker, target)
+
+    return None
+
+
+def _best_strike(
+    chain_df: pd.DataFrame,
+    strategy: str,
+    current_price: float,
+    nearest_support: Optional[float],
+    nearest_resistance: Optional[float],
+) -> Optional[pd.Series]:
+    """
+    Filter options chain for the best strike.
+
+    Rules:
+      covered_call — calls, OTM (strike > current_price), delta 0.20-0.35
+      cash_secured_put — puts, OTM (strike < current_price), |delta| 0.20-0.35,
+                         strike preferably > nearest_support
+    Returns the best row or None.
+    """
+    df = chain_df.copy()
+
+    # Ensure we have a usable delta column
+    has_delta = "delta" in df.columns and df["delta"].notna().any()
+
+    if strategy == "covered_call":
+        df = df[df["strike"] > current_price]
+        if has_delta:
+            df = df[(df["delta"] >= DELTA_MIN) & (df["delta"] <= DELTA_MAX)]
+        else:
+            # Approximate OTM calls: within 5% above current price
+            df = df[df["strike"] <= current_price * 1.07]
+    elif strategy == "cash_secured_put":
+        df = df[df["strike"] < current_price]
+        if has_delta:
+            df = df[(df["delta"].abs() >= DELTA_MIN) & (df["delta"].abs() <= DELTA_MAX)]
+        else:
+            df = df[df["strike"] >= current_price * 0.93]
+
+        # Prefer strikes above nearest support to avoid selling below key level
+        if nearest_support:
+            above_support = df[df["strike"] >= nearest_support * 0.99]
+            if not above_support.empty:
+                df = above_support
+
+    if df.empty:
+        return None
+
+    # Compute mid-price and pick highest premium
+    df = df.copy()
+    df["mid"] = (df["bid"] + df["ask"]) / 2
+    df = df[df["mid"] > 0]
+
+    if df.empty:
+        return None
+
+    return df.loc[df["mid"].idxmax()]
+
+
+def build_recommendation(
+    device_id: int,
+    ticker: str,
+    strategy: str,
+    time_horizon: str,
+    snapshot: Optional[SignalSnapshot] = None,
+) -> Optional[Recommendation]:
+    """
+    Build a Recommendation ORM object for the given parameters.
+    Returns None if not enough data exists.
+    """
+    # 1. Get signal snapshot (reuse if provided to avoid duplicate yfinance calls)
+    if snapshot is None:
+        df = md.get_price_history(ticker)
+        if df is None:
+            return None
+        current_price = md.get_current_price(ticker)
+        if current_price is None:
+            return None
+        iv_rank = compute_iv_rank(df)
+        smas = compute_smas(df)
+        swing = compute_swing_levels(df)
+        trend = compute_trend(current_price, smas["sma_50"], smas["sma_200"])
+        earnings_date = md.get_earnings_date(ticker)
+        nearest_support = swing["nearest_support"]
+        nearest_resistance = swing["nearest_resistance"]
+        sma_50 = smas["sma_50"]
+        sma_200 = smas["sma_200"]
+    else:
+        current_price = snapshot.current_price
+        iv_rank = snapshot.iv_rank
+        sma_50 = snapshot.sma_50
+        sma_200 = snapshot.sma_200
+        nearest_support = snapshot.nearest_support
+        nearest_resistance = snapshot.nearest_resistance
+        trend = snapshot.trend
+        earnings_date = snapshot.earnings_date
+
+    # 2. Determine target expiry
+    expiry_str = _target_expiry(ticker, time_horizon)
+    if expiry_str is None:
+        logger.debug(f"No expiry available for {ticker} {time_horizon}")
+        return None
+
+    expiry_date = date.fromisoformat(expiry_str)
+
+    # 3. Earnings warning
+    earnings_warning = bool(earnings_date and earnings_date <= expiry_date)
+
+    # 4. Fetch options chain
+    chain = md.get_options_chain(ticker, expiry_str)
+    if chain is None:
+        return None
+
+    chain_df = chain["calls"] if strategy == "covered_call" else chain["puts"]
+
+    # 5. Pick best strike
+    best = _best_strike(chain_df, strategy, current_price, nearest_support, nearest_resistance)
+    if best is None:
+        logger.debug(f"No qualifying strike for {ticker} {strategy} {time_horizon}")
+        return None
+
+    strike = float(best["strike"])
+    mid_price = float((best["bid"] + best["ask"]) / 2)
+    delta = float(best["delta"]) if not math.isnan(best.get("delta", float("nan"))) else _estimate_delta(strike, current_price, strategy)
+    theta = float(best["theta"]) if not math.isnan(best.get("theta", float("nan"))) else 0.0
+
+    # 6. Signal strength
+    signal_strength = compute_signal_strength(
+        iv_rank=iv_rank,
+        trend=trend,
+        strategy=strategy,
+        nearest_support=nearest_support,
+        nearest_resistance=nearest_resistance,
+        recommended_strike=strike,
+        earnings_warning=earnings_warning,
+    )
+
+    # 7. Signal hash
+    sig_hash = compute_signal_hash(
+        iv_rank=iv_rank,
+        sma_50=sma_50,
+        sma_200=sma_200,
+        nearest_support=nearest_support,
+        nearest_resistance=nearest_resistance,
+        recommended_strike=strike,
+        recommended_expiration=expiry_date,
+        earnings_warning=earnings_warning,
+    )
+
+    # 8. Build human-readable rationale
+    rationale = _build_rationale(
+        strategy=strategy,
+        time_horizon=time_horizon,
+        current_price=current_price,
+        strike=strike,
+        iv_rank=iv_rank,
+        trend=trend,
+        earnings_warning=earnings_warning,
+        earnings_date=earnings_date,
+        signal_strength=signal_strength,
+    )
+
+    return Recommendation(
+        device_id=device_id,
+        ticker=ticker,
+        strategy=strategy,
+        time_horizon=time_horizon,
+        current_price=current_price,
+        recommended_strike=strike,
+        recommended_expiration=expiry_date,
+        estimated_premium=mid_price,
+        delta=delta,
+        theta=theta,
+        iv_rank=iv_rank,
+        signal_strength=signal_strength,
+        earnings_warning=earnings_warning,
+        earnings_date=earnings_date,
+        rationale=rationale,
+        signal_hash=sig_hash,
+        created_at=datetime.utcnow(),
+    )
+
+
+def _estimate_delta(strike: float, current_price: float, strategy: str) -> float:
+    """Rough delta estimate when yfinance doesn't provide it."""
+    moneyness = (strike - current_price) / current_price
+    if strategy == "covered_call":
+        return max(0.05, 0.50 - moneyness * 3)
+    else:
+        return max(0.05, 0.50 - abs(moneyness) * 3)
+
+
+def _build_rationale(
+    strategy: str,
+    time_horizon: str,
+    current_price: float,
+    strike: float,
+    iv_rank: float,
+    trend: str,
+    earnings_warning: bool,
+    earnings_date: Optional[date],
+    signal_strength: str,
+) -> str:
+    parts = []
+
+    strategy_label = "Covered Call" if strategy == "covered_call" else "Cash-Secured Put"
+    direction = "above" if strategy == "covered_call" else "below"
+
+    parts.append(
+        f"{signal_strength.capitalize()} {strategy_label} setup. "
+        f"Strike ${strike:.2f} is {direction} current price ${current_price:.2f}."
+    )
+
+    if iv_rank >= 50:
+        parts.append(f"IV rank is elevated at {iv_rank:.0f}% — favorable premium-selling environment.")
+    elif iv_rank >= 30:
+        parts.append(f"IV rank is moderate at {iv_rank:.0f}%.")
+    else:
+        parts.append(f"IV rank is low at {iv_rank:.0f}% — premiums may be thin.")
+
+    trend_map = {"uptrend": "bullish uptrend", "downtrend": "bearish downtrend", "sideways": "sideways range"}
+    parts.append(f"Price is in a {trend_map.get(trend, trend)}.")
+
+    if earnings_warning and earnings_date:
+        parts.append(
+            f"⚠️ Earnings on {earnings_date} fall within this expiry — elevated risk. Consider a shorter expiry."
+        )
+
+    horizon_map = {"0dte": "0DTE", "1dte": "1DTE", "weekly": "weekly", "monthly": "monthly"}
+    parts.append(f"Horizon: {horizon_map.get(time_horizon, time_horizon)}.")
+
+    return " ".join(parts)

backend/app/services/signal_engine.py

@@ -0,0 +1,274 @@
+"""
+signal_engine.py — Compute all signals for a ticker.
+
+Signals:
+  - IV Rank (using rolling 30-day HV as proxy, since yfinance lacks historical IV)
+  - SMA-50, SMA-200
+  - Swing-based support / resistance (20-day window)
+  - Trend direction
+  - Signal strength score
+  - Signal hash (for change detection)
+"""
+
+import hashlib
+import json
+import logging
+import math
+from datetime import date
+from typing import Optional
+
+import numpy as np
+import pandas as pd
+
+from app.services.market_data import get_price_history, get_earnings_date, get_current_price
+from app.models.schemas import SignalSnapshot
+
+logger = logging.getLogger(__name__)
+
+
+# ─── IV Rank ──────────────────────────────────────────────────────────────────
+
+def compute_iv_rank(df: pd.DataFrame) -> float:
+    """
+    Compute IV Rank (0-100) using a 30-day rolling realized volatility
+    as a proxy for implied volatility.
+
+    Returns the current HV's rank within the past 52-week HV range.
+    Clamped to [0, 100].
+    """
+    if df is None or len(df) < 32:
+        return 50.0  # neutral fallback
+
+    closes = df["Close"].squeeze()
+    log_returns = np.log(closes / closes.shift(1)).dropna()
+
+    # 30-day rolling std, annualized
+    hv_series = log_returns.rolling(30).std() * math.sqrt(252) * 100  # as percentage
+    hv_series = hv_series.dropna()
+
+    if len(hv_series) < 2:
+        return 50.0
+
+    window = hv_series.iloc[-252:]  # last 52 weeks
+    hv_low = float(window.min())
+    hv_high = float(window.max())
+    current_hv = float(hv_series.iloc[-1])
+
+    if hv_high == hv_low:
+        return 50.0
+
+    rank = (current_hv - hv_low) / (hv_high - hv_low) * 100
+    return max(0.0, min(100.0, rank))
+
+
+# ─── Moving Averages ──────────────────────────────────────────────────────────
+
+def compute_smas(df: pd.DataFrame) -> dict:
+    """Return {'sma_50': float, 'sma_200': float}. Uses NaN if insufficient data."""
+    closes = df["Close"].squeeze()
+    sma_50 = float(closes.rolling(50).mean().iloc[-1]) if len(closes) >= 50 else float("nan")
+    sma_200 = float(closes.rolling(200).mean().iloc[-1]) if len(closes) >= 200 else float("nan")
+    return {"sma_50": sma_50, "sma_200": sma_200}
+
+
+def compute_trend(current_price: float, sma_50: float, sma_200: float) -> str:
+    """
+    uptrend   — price > sma50 > sma200
+    downtrend — price < sma50 < sma200
+    sideways  — otherwise
+    """
+    if any(math.isnan(v) for v in [sma_50, sma_200]):
+        return "sideways"
+    if current_price > sma_50 > sma_200:
+        return "uptrend"
+    if current_price < sma_50 < sma_200:
+        return "downtrend"
+    return "sideways"
+
+
+# ─── Support / Resistance ─────────────────────────────────────────────────────
+
+def compute_swing_levels(df: pd.DataFrame, lookback: int = 20, neighbors: int = 2) -> dict:
+    """
+    Find swing highs (resistance) and swing lows (support) over the last
+    `lookback` trading days using a `neighbors`-bar pivot rule.
+
+    Returns:
+        {
+          'nearest_support': float | None,
+          'nearest_resistance': float | None,
+          'support_levels': [float],
+          'resistance_levels': [float],
+        }
+    """
+    recent = df.tail(lookback + neighbors * 2)
+    highs = recent["High"].squeeze().values
+    lows = recent["Low"].squeeze().values
+    n = len(highs)
+
+    swing_highs = []
+    swing_lows = []
+
+    for i in range(neighbors, n - neighbors):
+        if all(highs[i] > highs[i - j] for j in range(1, neighbors + 1)) and \
+           all(highs[i] > highs[i + j] for j in range(1, neighbors + 1)):
+            swing_highs.append(float(highs[i]))
+        if all(lows[i] < lows[i - j] for j in range(1, neighbors + 1)) and \
+           all(lows[i] < lows[i + j] for j in range(1, neighbors + 1)):
+            swing_lows.append(float(lows[i]))
+
+    def cluster(levels: list[float], pct: float = 0.005) -> list[float]:
+        """Merge levels within pct% of each other into their mean."""
+        if not levels:
+            return []
+        sorted_levels = sorted(levels)
+        clusters = [[sorted_levels[0]]]
+        for lvl in sorted_levels[1:]:
+            if abs(lvl - clusters[-1][-1]) / clusters[-1][-1] <= pct:
+                clusters[-1].append(lvl)
+            else:
+                clusters.append([lvl])
+        return [sum(c) / len(c) for c in clusters]
+
+    resistance_levels = cluster(swing_highs)
+    support_levels = cluster(swing_lows)
+
+    # Get current price for context
+    current_price = float(df["Close"].squeeze().iloc[-1])
+
+    nearest_resistance = min(
+        (r for r in resistance_levels if r > current_price), default=None
+    )
+    nearest_support = max(
+        (s for s in support_levels if s < current_price), default=None
+    )
+
+    return {
+        "nearest_support": nearest_support,
+        "nearest_resistance": nearest_resistance,
+        "support_levels": support_levels,
+        "resistance_levels": resistance_levels,
+    }
+
+
+# ─── Signal Strength Scoring ──────────────────────────────────────────────────
+
+def compute_signal_strength(
+    iv_rank: float,
+    trend: str,
+    strategy: str,
+    nearest_support: Optional[float],
+    nearest_resistance: Optional[float],
+    recommended_strike: Optional[float],
+    earnings_warning: bool,
+) -> str:
+    """
+    Score the signal quality and return 'strong', 'moderate', or 'weak'.
+
+    Scoring:
+      +2 iv_rank >= 50 (premium-rich environment)
+      +1 iv_rank >= 30
+      +1 trend aligned (uptrend for covered_call, not downtrend for csp)
+      +1 sma alignment bonus (trend == uptrend for CC)
+      +1 strike positioned well vs nearest level
+      -2 earnings_warning (dangerous to hold through earnings)
+    """
+    score = 0
+
+    if iv_rank >= 50:
+        score += 2
+    elif iv_rank >= 30:
+        score += 1
+
+    if strategy == "covered_call":
+        if trend == "uptrend":
+            score += 2  # trend aligned (counts as +1 trend + +1 sma bonus)
+        elif trend == "sideways":
+            score += 1
+        if nearest_resistance and recommended_strike and recommended_strike < nearest_resistance:
+            score += 1
+    elif strategy == "cash_secured_put":
+        if trend != "downtrend":
+            score += 2
+        if trend == "uptrend":
+            score += 1  # extra bonus for strong bullish trend on a bullish strategy
+        if nearest_support and recommended_strike and recommended_strike > nearest_support:
+            score += 1
+
+    if earnings_warning:
+        score -= 2
+
+    if score >= 5:
+        return "strong"
+    elif score >= 3:
+        return "moderate"
+    else:
+        return "weak"
+
+
+# ─── Signal Hash ──────────────────────────────────────────────────────────────
+
+def compute_signal_hash(
+    iv_rank: float,
+    sma_50: float,
+    sma_200: float,
+    nearest_support: Optional[float],
+    nearest_resistance: Optional[float],
+    recommended_strike: Optional[float],
+    recommended_expiration: Optional[date],
+    earnings_warning: bool,
+) -> str:
+    """
+    16-char deterministic hash of rounded signal inputs.
+    Only changes when signals shift meaningfully — not on every tick.
+    """
+    payload = {
+        "ivr": round(iv_rank, 1),
+        "sma50": round(sma_50, 2) if sma_50 and not math.isnan(sma_50) else None,
+        "sma200": round(sma_200, 2) if sma_200 and not math.isnan(sma_200) else None,
+        "support": round(nearest_support, 2) if nearest_support else None,
+        "resistance": round(nearest_resistance, 2) if nearest_resistance else None,
+        "strike": recommended_strike,
+        "expiry": str(recommended_expiration) if recommended_expiration else None,
+        "ew": earnings_warning,
+    }
+    raw = json.dumps(payload, sort_keys=True)
+    return hashlib.sha256(raw.encode()).hexdigest()[:16]
+
+
+# ─── Full Signal Computation ──────────────────────────────────────────────────
+
+def compute_signals(ticker: str) -> Optional[SignalSnapshot]:
+    """
+    Compute and return a full SignalSnapshot for a ticker.
+    Returns None if market data is unavailable.
+    """
+    from datetime import datetime
+
+    df = get_price_history(ticker)
+    if df is None or df.empty:
+        logger.warning(f"No price history for {ticker} — cannot compute signals")
+        return None
+
+    current_price = get_current_price(ticker)
+    if current_price is None:
+        current_price = float(df["Close"].squeeze().iloc[-1])
+
+    iv_rank = compute_iv_rank(df)
+    smas = compute_smas(df)
+    swing = compute_swing_levels(df)
+    trend = compute_trend(current_price, smas["sma_50"], smas["sma_200"])
+    earnings_date = get_earnings_date(ticker)
+
+    return SignalSnapshot(
+        ticker=ticker,
+        current_price=current_price,
+        iv_rank=iv_rank,
+        sma_50=smas["sma_50"] if not math.isnan(smas["sma_50"]) else 0.0,
+        sma_200=smas["sma_200"] if not math.isnan(smas["sma_200"]) else 0.0,
+        nearest_support=swing["nearest_support"],
+        nearest_resistance=swing["nearest_resistance"],
+        trend=trend,
+        earnings_date=earnings_date,
+        computed_at=datetime.utcnow(),
+    )

backend/app/utils/date_helpers.py

@@ -0,0 +1,102 @@
+from datetime import date, timedelta
+import pandas_market_calendars as _mcal  # optional dep — fall back gracefully
+from typing import Optional
+import logging
+
+logger = logging.getLogger(__name__)
+
+_NYSE_CAL = None
+
+
+def _get_nyse():
+    global _NYSE_CAL
+    if _NYSE_CAL is None:
+        try:
+            import pandas_market_calendars as mcal
+            _NYSE_CAL = mcal.get_calendar("NYSE")
+        except ImportError:
+            pass
+    return _NYSE_CAL
+
+
+def is_trading_day(d: date) -> bool:
+    """Return True if d is a NYSE trading day."""
+    cal = _get_nyse()
+    if cal is None:
+        return d.weekday() < 5
+    import pandas as pd
+    schedule = cal.schedule(start_date=str(d), end_date=str(d))
+    return not schedule.empty
+
+
+def next_trading_day(d: date) -> date:
+    """Return the next trading day after d."""
+    candidate = d + timedelta(days=1)
+    while not is_trading_day(candidate):
+        candidate += timedelta(days=1)
+    return candidate
+
+
+def next_friday(from_date: Optional[date] = None) -> date:
+    """Return the next Friday on or after from_date."""
+    d = from_date or date.today()
+    days_ahead = 4 - d.weekday()  # Friday is weekday 4
+    if days_ahead < 0:
+        days_ahead += 7
+    elif days_ahead == 0:
+        pass  # today is Friday
+    return d + timedelta(days=days_ahead)
+
+
+def nearest_monthly_expiry(from_date: Optional[date] = None, target_dte: int = 30) -> date:
+    """Return the standard monthly expiry (third Friday) closest to target_dte from from_date."""
+    d = from_date or date.today()
+    target = d + timedelta(days=target_dte)
+
+    # Find third Friday of that month
+    year, month = target.year, target.month
+    third_friday = _third_friday(year, month)
+
+    # If the third Friday of target month is already past relative to today, advance one month
+    if third_friday <= d:
+        if month == 12:
+            year += 1
+            month = 1
+        else:
+            month += 1
+        third_friday = _third_friday(year, month)
+
+    return third_friday
+
+
+def _third_friday(year: int, month: int) -> date:
+    """Return the third Friday of the given month."""
+    first = date(year, month, 1)
+    # Find first Friday
+    days_to_friday = (4 - first.weekday()) % 7
+    first_friday = first + timedelta(days=days_to_friday)
+    return first_friday + timedelta(weeks=2)
+
+
+def market_is_open_now() -> bool:
+    """Best-effort check: is the US market currently open (9:30–16:00 ET)?"""
+    from datetime import datetime
+    import zoneinfo
+    now_et = datetime.now(tz=zoneinfo.ZoneInfo("America/New_York"))
+    if not is_trading_day(now_et.date()):
+        return False
+    open_time = now_et.replace(hour=9, minute=30, second=0, microsecond=0)
+    close_time = now_et.replace(hour=16, minute=0, second=0, microsecond=0)
+    return open_time <= now_et <= close_time
+
+
+def within_dte_window_for_0dte() -> bool:
+    """True if it's a trading day and between 9:30 AM and 2:00 PM ET."""
+    from datetime import datetime
+    import zoneinfo
+    now_et = datetime.now(tz=zoneinfo.ZoneInfo("America/New_York"))
+    if not is_trading_day(now_et.date()):
+        return False
+    open_time = now_et.replace(hour=9, minute=30, second=0, microsecond=0)
+    cutoff = now_et.replace(hour=14, minute=0, second=0, microsecond=0)
+    return open_time <= now_et <= cutoff

backend/migrations/init_db.py

@@ -0,0 +1,10 @@
+"""Run this once to initialize the SQLite database schema."""
+import sys
+import os
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), ".."))
+
+from app.database import init_db
+
+if __name__ == "__main__":
+    init_db()
+    print("Database initialized.")

backend/requirements.txt

@@ -0,0 +1,16 @@
+fastapi==0.115.6
+uvicorn[standard]==0.34.0
+sqlalchemy==2.0.36
+yfinance==0.2.50
+pandas==2.2.3
+numpy==2.2.1
+apscheduler==3.10.4
+httpx==0.28.1
+PyJWT==2.10.1
+cryptography==44.0.0
+python-dotenv==1.0.1
+pydantic==2.10.4
+pydantic-settings==2.7.0
+pytest==8.3.4
+pytest-asyncio==0.25.0
+httpx[test]

backend/tests/conftest.py

@@ -0,0 +1,4 @@
+"""Pytest configuration — sets asyncio mode."""
+import pytest
+
+pytest_plugins = ["pytest_asyncio"]

backend/tests/test_api.py

@@ -0,0 +1,220 @@
+"""
+Integration tests for the FastAPI API layer.
+Uses TestClient with an in-memory SQLite DB — no live market data.
+"""
+
+from datetime import date, timedelta
+from unittest.mock import MagicMock, patch
+
+import pytest
+from fastapi.testclient import TestClient
+from sqlalchemy import create_engine
+from sqlalchemy.orm import sessionmaker
+
+from app.database import Base, get_db
+from app.main import app
+
+# ─── Test DB setup ─────────────────────────────────────────────────────────────
+
+TEST_DB_URL = "sqlite://"  # in-memory
+
+test_engine = create_engine(TEST_DB_URL, connect_args={"check_same_thread": False})
+TestSessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=test_engine)
+
+
+def override_get_db():
+    db = TestSessionLocal()
+    try:
+        yield db
+    finally:
+        db.close()
+
+
+@pytest.fixture(autouse=True)
+def setup_db():
+    Base.metadata.create_all(bind=test_engine)
+    yield
+    Base.metadata.drop_all(bind=test_engine)
+
+
+app.dependency_overrides[get_db] = override_get_db
+
+# Disable scheduler during tests
+with patch("app.main.start_scheduler"), patch("app.main.stop_scheduler"):
+    client = TestClient(app, raise_server_exceptions=True)
+
+FAKE_TOKEN = "abc123device0000000000000000000000000000000000000000000000000000"
+
+
+# ─── Device registration ───────────────────────────────────────────────────────
+
+def test_register_device():
+    resp = client.post("/api/v1/devices/register", json={"apns_token": FAKE_TOKEN, "device_name": "Test iPhone"})
+    assert resp.status_code == 200
+    data = resp.json()
+    assert data["apns_token"] == FAKE_TOKEN
+    assert "id" in data
+
+
+def test_register_device_idempotent():
+    client.post("/api/v1/devices/register", json={"apns_token": FAKE_TOKEN})
+    resp = client.post("/api/v1/devices/register", json={"apns_token": FAKE_TOKEN})
+    assert resp.status_code == 200
+
+
+# ─── Portfolio ─────────────────────────────────────────────────────────────────
+
+@pytest.fixture
+def registered_device():
+    client.post("/api/v1/devices/register", json={"apns_token": FAKE_TOKEN})
+    return FAKE_TOKEN
+
+
+def test_add_portfolio(registered_device):
+    resp = client.post(
+        "/api/v1/portfolio",
+        json=[{"ticker": "AAPL", "shares": 100}, {"ticker": "MSFT", "shares": 200}],
+        headers={"X-Device-Token": registered_device},
+    )
+    assert resp.status_code == 200
+    tickers = [p["ticker"] for p in resp.json()]
+    assert "AAPL" in tickers
+    assert "MSFT" in tickers
+
+
+def test_get_portfolio_empty(registered_device):
+    resp = client.get("/api/v1/portfolio", headers={"X-Device-Token": registered_device})
+    assert resp.status_code == 200
+    assert resp.json() == []
+
+
+def test_get_portfolio_after_add(registered_device):
+    client.post(
+        "/api/v1/portfolio",
+        json=[{"ticker": "NVDA", "shares": 50}],
+        headers={"X-Device-Token": registered_device},
+    )
+    resp = client.get("/api/v1/portfolio", headers={"X-Device-Token": registered_device})
+    assert resp.status_code == 200
+    assert resp.json()[0]["ticker"] == "NVDA"
+
+
+def test_delete_ticker(registered_device):
+    client.post(
+        "/api/v1/portfolio",
+        json=[{"ticker": "AMD", "shares": 100}],
+        headers={"X-Device-Token": registered_device},
+    )
+    resp = client.delete("/api/v1/portfolio/AMD", headers={"X-Device-Token": registered_device})
+    assert resp.status_code == 204
+    remaining = client.get("/api/v1/portfolio", headers={"X-Device-Token": registered_device}).json()
+    assert all(p["ticker"] != "AMD" for p in remaining)
+
+
+def test_portfolio_unregistered_device():
+    resp = client.get("/api/v1/portfolio", headers={"X-Device-Token": "nonexistent_token"})
+    assert resp.status_code == 404
+
+
+# ─── Option Positions ──────────────────────────────────────────────────────────
+
+def test_log_position(registered_device):
+    expiry = str(date.today() + timedelta(days=14))
+    resp = client.post(
+        "/api/v1/positions",
+        json={
+            "ticker": "AAPL",
+            "strategy": "covered_call",
+            "strike": 195.0,
+            "expiration": expiry,
+            "premium_received": 2.50,
+            "contracts": 1,
+        },
+        headers={"X-Device-Token": registered_device},
+    )
+    assert resp.status_code == 201
+    data = resp.json()
+    assert data["ticker"] == "AAPL"
+    assert data["status"] == "open"
+    assert data["id"] is not None
+
+
+def test_close_position(registered_device):
+    expiry = str(date.today() + timedelta(days=14))
+    pos = client.post(
+        "/api/v1/positions",
+        json={
+            "ticker": "AAPL",
+            "strategy": "covered_call",
+            "strike": 195.0,
+            "expiration": expiry,
+            "premium_received": 2.50,
+        },
+        headers={"X-Device-Token": registered_device},
+    ).json()
+
+    resp = client.patch(
+        f"/api/v1/positions/{pos['id']}",
+        json={"status": "closed", "close_reason": "bought_back"},
+        headers={"X-Device-Token": registered_device},
+    )
+    assert resp.status_code == 200
+    assert resp.json()["status"] == "closed"
+
+
+def test_get_open_positions_filter(registered_device):
+    expiry = str(date.today() + timedelta(days=14))
+    client.post(
+        "/api/v1/positions",
+        json={"ticker": "TSLA", "strategy": "cash_secured_put", "strike": 200.0, "expiration": expiry, "premium_received": 3.0},
+        headers={"X-Device-Token": registered_device},
+    )
+    resp = client.get("/api/v1/positions?status=open", headers={"X-Device-Token": registered_device})
+    assert resp.status_code == 200
+    assert len(resp.json()) >= 1
+    assert all(p["status"] == "open" for p in resp.json())
+
+
+# ─── Alerts ────────────────────────────────────────────────────────────────────
+
+def test_get_alerts_empty(registered_device):
+    resp = client.get("/api/v1/alerts", headers={"X-Device-Token": registered_device})
+    assert resp.status_code == 200
+    assert resp.json() == []
+
+
+def test_acknowledge_alert(registered_device):
+    from app.models.db_models import Alert
+    from datetime import datetime
+
+    db = TestSessionLocal()
+    device_id = client.post("/api/v1/devices/register", json={"apns_token": FAKE_TOKEN}).json()["id"]
+    alert = Alert(
+        device_id=device_id,
+        ticker="AAPL",
+        alert_type="close_early",
+        message="Test alert",
+        sent_at=datetime.utcnow(),
+        acknowledged=False,
+    )
+    db.add(alert)
+    db.commit()
+    db.refresh(alert)
+    alert_id = alert.id
+    db.close()
+
+    resp = client.patch(
+        f"/api/v1/alerts/{alert_id}/acknowledge",
+        headers={"X-Device-Token": FAKE_TOKEN},
+    )
+    assert resp.status_code == 200
+    assert resp.json()["acknowledged"] is True
+
+
+# ─── Health ────────────────────────────────────────────────────────────────────
+
+def test_health():
+    with patch("app.routers"), patch("app.services.position_monitor.last_run", None):
+        resp = client.get("/api/v1/health")
+    assert resp.status_code == 200
+    assert resp.json()["status"] == "ok"

backend/tests/test_signal_engine.py

@@ -0,0 +1,267 @@
+"""Unit tests for signal_engine.py — all run without live market data."""
+
+import math
+from datetime import date, timedelta
+from unittest.mock import patch
+
+import numpy as np
+import pandas as pd
+import pytest
+
+from app.services.signal_engine import (
+    compute_iv_rank,
+    compute_smas,
+    compute_swing_levels,
+    compute_trend,
+    compute_signal_strength,
+    compute_signal_hash,
+)
+
+
+# ─── Helpers ──────────────────────────────────────────────────────────────────
+
+def _make_price_df(closes: list[float]) -> pd.DataFrame:
+    """Create a minimal OHLCV DataFrame from a list of close prices."""
+    n = len(closes)
+    dates = pd.date_range(end="2024-01-31", periods=n, freq="B")
+    arr = np.array(closes)
+    df = pd.DataFrame(
+        {
+            "Open": arr * 0.999,
+            "High": arr * 1.005,
+            "Low": arr * 0.995,
+            "Close": arr,
+            "Volume": np.ones(n) * 1_000_000,
+        },
+        index=dates,
+    )
+    return df
+
+
+def _trending_prices(start: float, end: float, n: int = 252) -> list[float]:
+    return list(np.linspace(start, end, n))
+
+
+# ─── IV Rank ──────────────────────────────────────────────────────────────────
+
+def test_iv_rank_high_returns_above_50():
+    """When current volatility is near the 52-week high, IVR should be > 50."""
+    # Create prices that spike at the end (high recent volatility)
+    stable = [100.0] * 220
+    volatile = list(np.linspace(100, 130, 16)) + list(np.linspace(130, 85, 16))
+    closes = stable + volatile
+    df = _make_price_df(closes)
+    ivr = compute_iv_rank(df)
+    assert ivr > 50, f"Expected ivr > 50, got {ivr:.1f}"
+
+
+def test_iv_rank_low_returns_below_50():
+    """Flat price series → low historical volatility → IVR near 0."""
+    closes = [100.0] * 252
+    df = _make_price_df(closes)
+    ivr = compute_iv_rank(df)
+    assert ivr <= 50, f"Expected ivr <= 50, got {ivr:.1f}"
+
+
+def test_iv_rank_insufficient_data_returns_neutral():
+    """Too few data points should return the 50.0 neutral fallback."""
+    df = _make_price_df([100.0] * 20)
+    ivr = compute_iv_rank(df)
+    assert ivr == 50.0
+
+
+def test_iv_rank_clamped_to_0_100():
+    closes = [100.0] * 252
+    df = _make_price_df(closes)
+    ivr = compute_iv_rank(df)
+    assert 0.0 <= ivr <= 100.0
+
+
+# ─── SMAs ─────────────────────────────────────────────────────────────────────
+
+def test_sma_50_correct_value():
+    closes = list(range(1, 253))  # 1..252
+    df = _make_price_df(closes)
+    smas = compute_smas(df)
+    expected_sma50 = sum(range(203, 253)) / 50  # last 50 values: 203-252
+    assert abs(smas["sma_50"] - expected_sma50) < 0.01
+
+
+def test_sma_200_nan_when_insufficient_data():
+    df = _make_price_df([100.0] * 100)
+    smas = compute_smas(df)
+    assert math.isnan(smas["sma_200"])
+
+
+def test_sma_50_nan_when_insufficient_data():
+    df = _make_price_df([100.0] * 30)
+    smas = compute_smas(df)
+    assert math.isnan(smas["sma_50"])
+
+
+# ─── Trend ────────────────────────────────────────────────────────────────────
+
+def test_trend_uptrend():
+    assert compute_trend(110.0, 105.0, 100.0) == "uptrend"
+
+
+def test_trend_downtrend():
+    assert compute_trend(90.0, 95.0, 100.0) == "downtrend"
+
+
+def test_trend_sideways_mixed():
+    # price > sma50 but sma50 < sma200
+    assert compute_trend(106.0, 105.0, 107.0) == "sideways"
+
+
+def test_trend_sideways_with_nan_smas():
+    assert compute_trend(100.0, float("nan"), float("nan")) == "sideways"
+
+
+# ─── Support / Resistance ─────────────────────────────────────────────────────
+
+def test_swing_levels_detects_swing_high():
+    """A clear peak in the middle of the window should be detected as resistance."""
+    closes = [100.0] * 25
+    highs = [100.0] * 25
+    lows = [98.0] * 25
+    # Create a swing high at index 12
+    highs[12] = 108.0
+    closes[12] = 107.0
+
+    dates = pd.date_range(end="2024-01-31", periods=25, freq="B")
+    df = pd.DataFrame({"Open": closes, "High": highs, "Low": lows, "Close": closes, "Volume": [1e6] * 25}, index=dates)
+
+    result = compute_swing_levels(df, lookback=20)
+    assert result["nearest_resistance"] is not None
+    assert abs(result["nearest_resistance"] - 108.0) < 1.0
+
+
+def test_swing_levels_detects_swing_low():
+    """A clear trough should be detected as support."""
+    closes = [100.0] * 25
+    highs = [102.0] * 25
+    lows = [98.0] * 25
+    lows[12] = 90.0
+    closes[12] = 91.0
+
+    dates = pd.date_range(end="2024-01-31", periods=25, freq="B")
+    df = pd.DataFrame({"Open": closes, "High": highs, "Low": lows, "Close": closes, "Volume": [1e6] * 25}, index=dates)
+
+    result = compute_swing_levels(df, lookback=20)
+    assert result["nearest_support"] is not None
+    assert result["nearest_support"] < 100.0
+
+
+def test_swing_levels_flat_no_crash():
+    """Flat price should return None for support/resistance without crashing."""
+    df = _make_price_df([100.0] * 30)
+    result = compute_swing_levels(df, lookback=20)
+    # No assertion on values — just shouldn't raise
+    assert "nearest_support" in result
+    assert "nearest_resistance" in result
+
+
+# ─── Signal Strength ──────────────────────────────────────────────────────────
+
+def test_signal_strength_strong_covered_call():
+    strength = compute_signal_strength(
+        iv_rank=70.0,
+        trend="uptrend",
+        strategy="covered_call",
+        nearest_support=95.0,
+        nearest_resistance=110.0,
+        recommended_strike=107.0,
+        earnings_warning=False,
+    )
+    assert strength == "strong"
+
+
+def test_signal_strength_earnings_warning_downgrades():
+    strength_no_ew = compute_signal_strength(
+        iv_rank=70.0, trend="uptrend", strategy="covered_call",
+        nearest_support=95.0, nearest_resistance=110.0,
+        recommended_strike=107.0, earnings_warning=False,
+    )
+    strength_ew = compute_signal_strength(
+        iv_rank=70.0, trend="uptrend", strategy="covered_call",
+        nearest_support=95.0, nearest_resistance=110.0,
+        recommended_strike=107.0, earnings_warning=True,
+    )
+    assert strength_ew in ("weak", "moderate")
+    assert strength_no_ew == "strong"
+
+
+def test_signal_strength_low_ivr_gives_weak():
+    strength = compute_signal_strength(
+        iv_rank=10.0,
+        trend="downtrend",
+        strategy="covered_call",
+        nearest_support=None,
+        nearest_resistance=None,
+        recommended_strike=None,
+        earnings_warning=True,
+    )
+    assert strength == "weak"
+
+
+def test_signal_strength_csp_uptrend_bonus():
+    strength = compute_signal_strength(
+        iv_rank=55.0,
+        trend="uptrend",
+        strategy="cash_secured_put",
+        nearest_support=95.0,
+        nearest_resistance=110.0,
+        recommended_strike=97.0,
+        earnings_warning=False,
+    )
+    assert strength in ("strong", "moderate")
+
+
+# ─── Signal Hash ──────────────────────────────────────────────────────────────
+
+def test_signal_hash_stability():
+    """Same inputs must always produce the same hash."""
+    args = dict(
+        iv_rank=45.5,
+        sma_50=150.0,
+        sma_200=145.0,
+        nearest_support=148.0,
+        nearest_resistance=155.0,
+        recommended_strike=153.0,
+        recommended_expiration=date(2024, 3, 15),
+        earnings_warning=False,
+    )
+    h1 = compute_signal_hash(**args)
+    h2 = compute_signal_hash(**args)
+    assert h1 == h2
+    assert len(h1) == 16
+
+
+def test_signal_hash_changes_on_strike_change():
+    base = dict(
+        iv_rank=45.5, sma_50=150.0, sma_200=145.0,
+        nearest_support=148.0, nearest_resistance=155.0,
+        recommended_strike=153.0, recommended_expiration=date(2024, 3, 15),
+        earnings_warning=False,
+    )
+    h1 = compute_signal_hash(**base)
+    base["recommended_strike"] = 155.0
+    h2 = compute_signal_hash(**base)
+    assert h1 != h2
+
+
+def test_signal_hash_stable_on_noise():
+    """Tiny IV rank change (< 0.1) should produce same hash due to rounding."""
+    base = dict(
+        iv_rank=45.500,
+        sma_50=150.01, sma_200=145.009,
+        nearest_support=148.001, nearest_resistance=155.002,
+        recommended_strike=153.0, recommended_expiration=date(2024, 3, 15),
+        earnings_warning=False,
+    )
+    tweaked = dict(base)
+    tweaked["iv_rank"] = 45.503  # < 0.1 change → rounds to same 45.5
+    h1 = compute_signal_hash(**base)
+    h2 = compute_signal_hash(**tweaked)
+    assert h1 == h2