rates/backend/api/analysis.py

"""
Rate Analysis API — advance purchase curves, DOW analysis, rate timelines, strategy summary
"""
from fastapi import APIRouter, Depends, Query
from sqlalchemy import text
from typing import Optional
from datetime import date, timedelta

from database import AsyncSessionLocal
from auth import get_current_user, require_cap

router = APIRouter()


async def get_db():
    async with AsyncSessionLocal() as db:
        yield db


# ─── Hotels available for analysis ───────────────────────────────────────────

@router.get("/hotels")
async def list_analysis_hotels(
    tier: Optional[str] = Query(None, description="Filter by tier: own|competitor|market"),
    user=Depends(get_current_user)
):
    require_cap(user, "rate_analysis")
    async with AsyncSessionLocal() as db:
        where = "is_active = true"
        params = {}
        if tier:
            where += " AND tier = :tier"
            params["tier"] = tier
        result = await db.execute(
            text(f"""
                SELECT h.id, h.name, h.tier, h.star_rating, h.review_score,
                       h.booking_com_url,
                       COUNT(DISTINCT r.rate_date) AS scraped_dates,
                       MAX(r.scraped_at)           AS last_scraped
                FROM booking_com_hotels h
                LEFT JOIN booking_com_rates r ON r.hotel_id = h.id
                WHERE {where}
                GROUP BY h.id, h.name, h.tier, h.star_rating, h.review_score, h.booking_com_url
                ORDER BY h.tier, h.display_order, h.name
            """),
            params
        )
        rows = result.mappings().all()
        return [dict(r) for r in rows]


# ─── Full analysis for one hotel ─────────────────────────────────────────────

@router.get("/hotel/{hotel_id}")
async def analyse_hotel(
    hotel_id: int,
    from_date: date = Query(default_factory=lambda: date.today()),
    to_date: date = Query(default_factory=lambda: date.today() + timedelta(days=89)),
    user=Depends(get_current_user)
):
    require_cap(user, "rate_analysis")
    async with AsyncSessionLocal() as db:
        # Hotel info
        hotel_row = await db.execute(
            text("SELECT id, name, tier, star_rating, booking_com_url FROM booking_com_hotels WHERE id = :id"),
            {"id": hotel_id}
        )
        hotel = hotel_row.mappings().fetchone()
        if not hotel:
            from fastapi import HTTPException
            raise HTTPException(status_code=404, detail="Hotel not found")

        # Advance purchase curve
        apc_result = await db.execute(
            text("""
                SELECT
                    CASE
                        WHEN (rate_date - scraped_at::date) <= 7  THEN '0-7d'
                        WHEN (rate_date - scraped_at::date) <= 29 THEN '8-29d'
                        WHEN (rate_date - scraped_at::date) <= 89 THEN '30-89d'
                        ELSE '90+d'
                    END AS lead_bucket,
                    ROUND(AVG(rate_gross)::numeric, 2) AS avg_rate,
                    COUNT(*)                           AS sample_count
                FROM booking_com_rates
                WHERE hotel_id   = :hid
                  AND rate_date  BETWEEN :from_date AND :to_date
                  AND availability_status = 'available'
                  AND rate_gross IS NOT NULL
                GROUP BY lead_bucket
                ORDER BY
                    CASE lead_bucket
                        WHEN '0-7d'   THEN 1
                        WHEN '8-29d'  THEN 2
                        WHEN '30-89d' THEN 3
                        ELSE 4
                    END
            """),
            {"hid": hotel_id, "from_date": from_date, "to_date": to_date}
        )
        advance_purchase_curve = [dict(r) for r in apc_result.mappings().all()]

        # Day-of-week averages (latest scrape per date)
        dow_result = await db.execute(
            text("""
                SELECT
                    EXTRACT(DOW FROM rate_date)::int        AS dow,
                    TO_CHAR(rate_date, 'Dy')                AS dow_label,
                    ROUND(AVG(rate_gross)::numeric, 2)      AS avg_rate,
                    COUNT(DISTINCT rate_date)               AS date_count
                FROM (
                    -- Latest row per date first, THEN filter: a 'not_listed'
                    -- latest row must drop the date, not resurface an older rate
                    SELECT DISTINCT ON (rate_date)
                        rate_date, rate_gross, availability_status
                    FROM booking_com_rates
                    WHERE hotel_id  = :hid
                      AND rate_date BETWEEN :from_date AND :to_date
                    ORDER BY rate_date, scraped_at DESC
                ) latest
                WHERE availability_status = 'available'
                  AND rate_gross IS NOT NULL
                GROUP BY dow, dow_label
                ORDER BY dow
            """),
            {"hid": hotel_id, "from_date": from_date, "to_date": to_date}
        )
        dow_analysis = [dict(r) for r in dow_result.mappings().all()]

        # Sold-out pattern by day-of-week
        sold_out_result = await db.execute(
            text("""
                SELECT
                    EXTRACT(DOW FROM rate_date)::int  AS dow,
                    TO_CHAR(rate_date, 'Dy')          AS dow_label,
                    COUNT(DISTINCT rate_date)         AS total_dates,
                    COUNT(DISTINCT rate_date) FILTER (
                        WHERE availability_status = 'sold_out'
                    )                                 AS sold_out_dates
                FROM (
                    SELECT DISTINCT ON (rate_date)
                        rate_date, availability_status
                    FROM booking_com_rates
                    WHERE hotel_id  = :hid
                      AND rate_date BETWEEN :from_date AND :to_date
                    ORDER BY rate_date, scraped_at DESC
                ) latest
                GROUP BY dow, dow_label
                ORDER BY dow
            """),
            {"hid": hotel_id, "from_date": from_date, "to_date": to_date}
        )
        sold_out_pattern = [dict(r) for r in sold_out_result.mappings().all()]

        # Strategy summary
        strategy = _compute_strategy(advance_purchase_curve, dow_analysis, sold_out_pattern)

        return {
            "hotel": dict(hotel),
            "date_range": {"from": str(from_date), "to": str(to_date)},
            "advance_purchase_curve": advance_purchase_curve,
            "dow_analysis": dow_analysis,
            "sold_out_pattern": sold_out_pattern,
            "strategy_summary": strategy,
        }


# ─── Rate timeline for a single date ─────────────────────────────────────────

@router.get("/hotel/{hotel_id}/timeline")
async def rate_timeline(
    hotel_id: int,
    rate_date: date = Query(...),
    user=Depends(get_current_user)
):
    require_cap(user, "rate_analysis")
    async with AsyncSessionLocal() as db:
        result = await db.execute(
            text("""
                SELECT
                    scraped_at,
                    rate_gross,
                    availability_status,
                    rooms_left,
                    room_type,
                    (rate_date - scraped_at::date) AS days_out
                FROM booking_com_rates
                WHERE hotel_id = :hid AND rate_date = :rd
                ORDER BY scraped_at ASC
            """),
            {"hid": hotel_id, "rd": rate_date}
        )
        rows = result.mappings().all()
        return {
            "hotel_id": hotel_id,
            "rate_date": str(rate_date),
            "timeline": [dict(r) for r in rows],
        }


# ─── Own vs Competitor comparison ────────────────────────────────────────────

@router.get("/comparison")
async def rate_comparison(
    from_date: date = Query(default_factory=lambda: date.today()),
    to_date: date = Query(default_factory=lambda: date.today() + timedelta(days=29)),
    competitor_ids: Optional[str] = Query(None, description="Comma-separated hotel IDs; defaults to all active competitors"),
    user=Depends(get_current_user)
):
    """Per-hotel market comparison: avg own vs competitor rate over the range."""
    require_cap(user, "rate_analysis")
    comp_ids = None
    if competitor_ids:
        try:
            comp_ids = [int(x.strip()) for x in competitor_ids.split(",") if x.strip()]
        except ValueError:
            from fastapi import HTTPException
            raise HTTPException(status_code=400, detail="competitor_ids must be comma-separated integers")

    async with AsyncSessionLocal() as db:
        # Own hotel avg rate per date (across included categories)
        own_result = await db.execute(
            text("""
                SELECT rate_date, AVG(rate_gross) AS own_rate
                FROM newbook_current_rates
                WHERE rate_date BETWEEN :from_date AND :to_date
                GROUP BY rate_date
            """),
            {"from_date": from_date, "to_date": to_date}
        )
        own_rates = {r.rate_date: float(r.own_rate) for r in own_result if r.own_rate}

        # Competitor latest rate per date
        hotel_filter = "h.id = ANY(:comp_ids)" if comp_ids else "h.tier = 'competitor'"
        comp_result = await db.execute(
            text(f"""
                SELECT r.hotel_id, h.name AS hotel_name, r.rate_date, r.rate_gross
                FROM (
                    SELECT DISTINCT ON (hotel_id, rate_date)
                        hotel_id, rate_date, rate_gross
                    FROM booking_com_rates
                    WHERE rate_date BETWEEN :from_date AND :to_date
                    ORDER BY hotel_id, rate_date, scraped_at DESC
                ) r
                JOIN booking_com_hotels h ON h.id = r.hotel_id
                WHERE h.is_active = true AND {hotel_filter}
                ORDER BY h.display_order, h.name
            """),
            {"from_date": from_date, "to_date": to_date,
             **({"comp_ids": comp_ids} if comp_ids else {})}
        )

        # Aggregate per hotel, comparing own rates over the same dates
        by_hotel: dict = {}
        for row in comp_result.mappings().all():
            entry = by_hotel.setdefault(row["hotel_id"], {
                "hotel_id": row["hotel_id"],
                "hotel_name": row["hotel_name"],
                "their": [], "ours": [],
            })
            if row["rate_gross"]:
                entry["their"].append(float(row["rate_gross"]))
                if row["rate_date"] in own_rates:
                    entry["ours"].append(own_rates[row["rate_date"]])

        rows = []
        for entry in by_hotel.values():
            their_rate = round(sum(entry["their"]) / len(entry["their"]), 2) if entry["their"] else None
            our_rate = round(sum(entry["ours"]) / len(entry["ours"]), 2) if entry["ours"] else None
            price_index = round(their_rate / our_rate * 100, 1) if their_rate and our_rate else None
            rows.append({
                "hotel_id": entry["hotel_id"],
                "hotel_name": entry["hotel_name"],
                "our_rate": our_rate,
                "their_rate": their_rate,
                "price_index": price_index,
                "days_checked": len(entry["their"]),
            })
        return rows


# ─── Strategy computation helper ─────────────────────────────────────────────

def _compute_strategy(apc: list, dow: list, sold_out: list) -> dict:
    # Advance discount — compare 0-7d vs 30-89d
    rates_by_bucket = {r["lead_bucket"]: float(r["avg_rate"]) for r in apc if r.get("avg_rate")}
    advance_discount_pct = None
    if "0-7d" in rates_by_bucket and "30-89d" in rates_by_bucket:
        close_in = rates_by_bucket["0-7d"]
        far_out  = rates_by_bucket["30-89d"]
        if far_out > 0:
            # Positive = closes in higher (scarcity premium); negative = discount for advance
            advance_discount_pct = round((close_in - far_out) / far_out * 100, 1)

    # Weekend premium — Fri(5)+Sat(6) vs Mon(1)–Thu(4)
    rates_by_dow = {r["dow"]: float(r["avg_rate"]) for r in dow if r.get("avg_rate")}
    weekend_premium_pct = None
    weekend_rates = [rates_by_dow[d] for d in [5, 6] if d in rates_by_dow]
    weekday_rates = [rates_by_dow[d] for d in [1, 2, 3, 4] if d in rates_by_dow]
    if weekend_rates and weekday_rates:
        avg_wk = sum(weekend_rates) / len(weekend_rates)
        avg_wd = sum(weekday_rates) / len(weekday_rates)
        if avg_wd > 0:
            weekend_premium_pct = round((avg_wk - avg_wd) / avg_wd * 100, 1)

    # Sold-out rate
    total_dates = sum(r["total_dates"] for r in sold_out)
    sold_out_dates = sum(r["sold_out_dates"] for r in sold_out)
    sold_out_rate_pct = round(sold_out_dates / total_dates * 100, 1) if total_dates > 0 else None

    # Strategy label
    label = "Mixed / insufficient data"
    if advance_discount_pct is not None:
        if advance_discount_pct <= -5:
            label = "Advance-booking discounter"
        elif advance_discount_pct >= 8 and (sold_out_rate_pct or 0) >= 10:
            label = "Yield manager (scarcity-driven)"
        elif advance_discount_pct >= 3:
            label = "Flat-rate / hold-firm strategy"
        else:
            label = "Stable pricing"

    return {
        "advance_discount_pct": advance_discount_pct,
        "weekend_premium_pct": weekend_premium_pct,
        "avg_sold_out_rate_pct": sold_out_rate_pct,
        "strategy_label": label,
    }