thermograph/app.py

"""Thermograph API — grade recent local weather against ~45 years of climatology."""
import contextlib
import datetime
import functools
import hashlib
import json
import os
import time

import pandas as pd
from fastapi import APIRouter, FastAPI, HTTPException, Query, Request, Response
from fastapi.middleware.gzip import GZipMiddleware
from fastapi.responses import RedirectResponse
from fastapi.staticfiles import StaticFiles

import audit
import climate
import grading
import grid
import places
import store

FRONTEND_DIR = os.path.join(os.path.dirname(__file__), "..", "frontend")

# Warm the local place-name index (/suggest's typo tolerance) in the
# background; the app boots and serves fine without it — suggestions just fall
# back to the upstream geocoder until it's ready.
places.start_loading()

# Everything (pages, assets, API) is served under this base path so the app can
# live at https://<host>/thermograph/ behind a shared domain. Override with the
# THERMOGRAPH_BASE env var. The frontend uses base-relative URLs, so it follows
# this automatically without hardcoding the prefix.
BASE = "/" + os.environ.get("THERMOGRAPH_BASE", "/thermograph").strip("/")

# Observed values pulled from a daily record row for grading. Includes the
# temperature-scale metrics (tmax/tmin/feels/wind/gust) plus precip; a column may
# be absent on an older cache, so only carry the ones present.
OBS_COLS = ("tmax", "tmin", "precip", "feels", "humid", "wind", "gust")

# Bump when any response payload shape changes (new metrics, renamed fields…).
# The version is part of every derived-store validity token, so one bump
# atomically orphans all pre-upgrade cached payloads instead of letting a stale
# row whose history_end happens to match keep serving the old shape.
PAYLOAD_VER = "p1"


def _obs_from_row(row) -> dict:
    return {k: row[k] for k in OBS_COLS if k in row}


def _weather_fetch_error(e) -> HTTPException:
    """A clean, retryable message for a rate limit; the raw error otherwise."""
    low = str(e).lower()
    reason = (climate._rate_limit_reason(e) or "").lower()
    blob = low + " " + reason
    if climate._is_rate_limit(e) or "429" in low or "rate-limited" in low or "limit" in reason:
        if "daily" in blob or "tomorrow" in blob:   # daily quota exhausted — resets tomorrow
            return HTTPException(
                status_code=503,
                detail="Open-Meteo's daily request limit is exhausted — new locations will work again "
                       "tomorrow. Places you've already viewed still work.",
            )
        return HTTPException(
            status_code=503,
            detail="The weather service is rate-limited right now — please try again in a minute.",
        )
    return HTTPException(status_code=502, detail=f"weather data fetch failed: {e}")


def _grade_rows(history, rows_df) -> list[dict]:
    """Grade each daily row against its own ±7-day climatology window."""
    return [grading.grade_day(history, row["date"], _obs_from_row(row))
            for _, row in rows_df.iterrows()]


def _attach_dry_streaks(graded: list[dict], *precip_frames) -> None:
    """Attach `dsr` (days since last measurable rain) to each graded day, computed
    over a combined, de-duplicated precip series so streaks stay continuous across
    the history / recent / forecast sources."""
    frames = [f[["date", "precip"]] for f in precip_frames if f is not None and not f.empty]
    if not frames:
        return
    combined = (pd.concat(frames)
                .drop_duplicates(subset="date", keep="last")
                .sort_values("date"))
    dsr_map = grading.dry_streaks(combined["date"].values, combined["precip"].values)
    for g in graded:
        g["dsr"] = dsr_map.get(g["date"])


app = FastAPI(title="Thermograph", version="0.2.0")

# Compress every sizeable response (the 2-year calendar JSON shrinks ~6-8×).
# Applies to API JSON and static assets alike; tiny responses are left alone.
app.add_middleware(GZipMiddleware, minimum_size=1024)


@app.middleware("http")
async def revalidate_static(request, call_next):
    """Serve the frontend with no-cache (NOT no-store): browsers may keep a copy
    but must revalidate it on every use. Starlette's FileResponse/StaticFiles
    already emit ETag + Last-Modified, so an unchanged asset costs one conditional
    request answered with an empty 304 instead of a full re-download — page-to-page
    navigation stops re-transferring the JS/CSS/HTML while still picking up every
    deploy immediately (no "my change isn't showing" bugs)."""
    response = await call_next(request)
    path = request.url.path
    pages = (BASE, f"{BASE}/", f"{BASE}/calendar", f"{BASE}/day", f"{BASE}/compare", f"{BASE}/legend")
    if path.endswith((".js", ".css", ".html")) or path in pages:
        response.headers["Cache-Control"] = "no-cache"
    return response


# --- derived-store plumbing --------------------------------------------------
# Every graded payload is cached in SQLite under (kind, cell, key) and validated
# by a token that encodes what it was computed from (see store.py). The freshness
# drivers stay where they were — climate.get_history tops up the archive tail
# hourly and get_recent_forecast refetches hourly — and the tokens are derived
# from what those return, so a cached payload expires exactly when its inputs
# change and never before. The same tokens double as ETags: a client sending
# If-None-Match gets an empty 304 without the payload even being loaded.

class _NullRun:
    """audit.RunAudit stand-in for offline callers (the migrate script)."""

    def set(self, **kw):
        return self

    def phase(self, name):
        return contextlib.nullcontext()


def _hist_end(history) -> str:
    """Last day in the cell's archive record — the freshness token for everything
    derived purely from history. Advances via the hourly tail top-up inside
    climate.get_history, which cached payloads follow automatically."""
    return pd.Timestamp(history["date"].max()).date().isoformat()


def _etag_for(kind: str, cell_id: str, key: str, token: str) -> str:
    """Deterministic weak ETag from a payload's identity + validity token. Weak
    because the same content may be served under different encodings (gzip)."""
    h = hashlib.sha1(f"{kind}:{cell_id}:{key}:{token}".encode()).hexdigest()[:20]
    return f'W/"{h}"'


def _not_modified(request: Request, etag: str) -> bool:
    """Does the client already hold this exact payload? Answerable from the token
    alone — no payload load needed."""
    inm = request.headers.get("if-none-match")
    if not inm:
        return False
    tags = {t.strip() for t in inm.split(",")}
    return "*" in tags or etag in tags or etag.removeprefix("W/") in tags


def _encode(payload: dict) -> bytes:
    """Compact JSON bytes, matching store.put_payload's encoding (allow_nan=False
    mirrors Starlette — a NaN from grading should fail loudly, not reach a client)."""
    return json.dumps(payload, default=str, allow_nan=False, separators=(",", ":")).encode()


def _json_response(body: bytes, etag: str | None = None) -> Response:
    headers = {"ETag": etag} if etag else {}
    return Response(content=body, media_type="application/json", headers=headers)


# --- payload builders --------------------------------------------------------
# Pure "inputs → response dict" functions shared by the per-view endpoints, the
# /cell bundle, and the offline migrate script. HTTP semantics (audit runs, cache
# lookups, etags) stay in the endpoints; `run` is the audit run or None.

def _build_grade(cell, target, days, history, recent, cache_meta, place, run=None, after=14) -> dict:
    """/grade payload: a window of days centered on the target — `days` of history
    before it, the target itself, then up to `after` days after it — plus the
    target day's climatology summary. Days after the target are observed when they
    are already in the past and forecast when they run into the future, so the
    weekly view can frame two weeks of history around the orange target marker and
    trail off into up to 14 days of observations / forecast after it. The forecast
    only reaches ~7 days out, so a recent target naturally yields some observed days
    plus 1-7 forecast days, while an older target fills the whole 14 with real obs."""
    run = run or _NullRun()
    with run.phase("grading"):
        lo = target - pd.Timedelta(days=days)
        hi = target + pd.Timedelta(days=after)
        # Build the window from both sources. The recent+forecast bundle covers the
        # last few weeks plus the forward forecast (the only source for future days);
        # the archive reaches decades back for targets older than that bundle. Prefer
        # the bundle row for any given date, filling the rest from the archive.
        rwin = recent[(recent["date"] >= lo) & (recent["date"] <= hi)]
        hwin = history[(history["date"] >= lo) & (history["date"] <= hi)]
        hwin = hwin[~hwin["date"].isin(set(rwin["date"]))]
        window = pd.concat([rwin, hwin]).sort_values("date")
        graded = _grade_rows(history, window)
        _attach_dry_streaks(graded, history, recent)
        graded.reverse()  # newest first for display
        climo = grading.climatology(history, int(target.dayofyear))

    # Summarize what this run actually covered. A fresh history fetch pulls the
    # full ~45-year archive (run_type "full"); a cache hit only fetched the
    # recent window (run_type "partial").
    years = pd.to_datetime(history["date"]).dt.year
    full = not cache_meta.get("cached", False)
    run.set(
        run_type="full" if full else "partial",
        history_source="fetch" if full else "cache",
        history_rows=int(len(history)),
        history_years=int(years.max() - years.min() + 1),
        history_span=[int(years.min()), int(years.max())],
        cache_age_days=cache_meta.get("cache_age_days"),
        graded_days=len(graded),
        place_found=place is not None,
    )
    return {
        "cell": cell,
        "place": place,
        "target_date": target.date().isoformat(),
        "cache": cache_meta,
        "climatology": climo,
        "recent": graded,
    }


CAL_MAX_SPAN_DAYS = 732  # ~2 years — the most a single calendar request will grade
                         # (the frontend splits longer spans into 2-year chunks)


def _cal_span(history, start, end, months) -> tuple[pd.Timestamp, pd.Timestamp]:
    """Clamp a requested calendar range to the available record (≤ ~2 years).

    Without a start, defaults to whole months back landing on the SAME
    day-of-month as the end, so the two range endpoints line up
    (e.g. 2024-06-29 → 2026-06-29)."""
    first = pd.Timestamp(history["date"].min()).normalize()
    last = pd.Timestamp(history["date"].max()).normalize()
    end_ts = min(pd.Timestamp(end).normalize(), last) if end else last
    if start:
        start_ts = pd.Timestamp(start).normalize()
    else:
        start_ts = (end_ts - pd.DateOffset(months=months)).normalize()
    # Cap the graded span at ~2 years, and keep it inside the available record.
    min_start = end_ts - pd.Timedelta(days=CAL_MAX_SPAN_DAYS - 1)
    start_ts = max(start_ts, min_start, first)
    start_ts = min(start_ts, end_ts)
    return start_ts, end_ts


def _build_calendar(cell, history, start_ts, end_ts, months, place, run=None) -> dict:
    """/calendar payload: every day in [start_ts, end_ts] graded, compact shape."""
    run = run or _NullRun()
    with run.phase("grading"):
        days = grading.grade_range(history, start_ts, end_ts)
    run.set(graded_days=len(days))
    return {
        "api_version": "v2",
        "cell": cell,
        "place": place,
        "range": {"start": start_ts.date().isoformat(), "end": end_ts.date().isoformat()},
        "months": months,
        "days": days,
    }


def _build_day(cell, history, target, place, run=None, recent=None) -> dict:
    """/day payload: the full percentile breakdown for one day. Observed values
    prefer the archive record; a date newer than it comes from the recent window
    (passed pre-loaded by the bundle path, fetched here otherwise)."""
    run = run or _NullRun()
    last = pd.Timestamp(history["date"].max()).normalize()

    obs = None
    hit = history[history["date"] == target]
    if not hit.empty:
        obs = _obs_from_row(hit.iloc[0])
    elif target > last:
        try:
            if recent is None:
                with run.phase("recent"):
                    recent = climate.get_recent_forecast(cell)
            rr = recent[recent["date"] == target]
            if not rr.empty:
                obs = _obs_from_row(rr.iloc[0])
        except Exception:  # noqa: BLE001 - detail page still works from climatology alone
            obs = None

    with run.phase("detail"):
        detail = grading.day_detail(history, target, obs)
    run.set(has_observation=obs is not None)
    return {
        "api_version": "v2",
        "cell": cell,
        "place": place,
        "latest": last.date().isoformat(),
        "detail": detail,
    }


def _build_forecast(cell, days, history, fc, today, place, run=None) -> dict:
    """/forecast payload: the next `days` forecast days graded — same shape as
    /grade so the frontend renders it identically, furthest-out day first."""
    run = run or _NullRun()
    # Strictly future days (tomorrow onward), earliest→latest, capped at `days`.
    future = fc[fc["date"] > today].sort_values("date").head(days)
    with run.phase("grading"):
        graded = _grade_rows(history, future)
        _attach_dry_streaks(graded, history, fc)
        graded.reverse()  # furthest-out first, so the chart reverses to L→R chronological
        climo = grading.climatology(history, int(today.dayofyear))
    run.set(graded_days=len(graded), place_found=place is not None)
    return {
        "api_version": "v2",
        "cell": cell,
        "place": place,
        "target_date": today.date().isoformat(),
        "forecast": True,
        "climatology": climo,
        "recent": graded,
    }


# --- endpoints ----------------------------------------------------------------

def api_geocode(q: str = Query(..., min_length=1)):
    try:
        return {"results": climate.geocode(q)}
    except Exception as e:  # noqa: BLE001 - surface upstream failures to the client
        raise HTTPException(status_code=502, detail=f"geocoding failed: {e}")


_SUGGEST_LIMIT = 5


@functools.lru_cache(maxsize=1024)
def _suggest_upstream(q: str) -> tuple:
    """Open-Meteo lookup for /suggest, memoized per query string — type-ahead
    re-asks the same prefixes constantly (backspacing, retyping). Failures
    raise and are not cached, so a transient upstream error doesn't stick."""
    return tuple(climate.geocode(q, count=_SUGGEST_LIMIT))


def _suggest_score(r: dict, exact: bool) -> int:
    """Prominence rank with an exactness boost — big enough that a real place
    beats a same-size typo match, small enough that a hamlet spelled exactly
    like the typo can't outrank a metropolis one edit away ("Seatle", a
    village, must not beat Seattle)."""
    pop = r.get("population") or 0
    return pop * 4 + 1 if exact else pop


def _suggest_merge(local: list, upstream: tuple, limit: int) -> list:
    """Blend local-index and upstream results into one top-`limit` list.
    Everything the user's spelling matches exactly counts as exact; only the
    local index's typo-tolerant matches take the fuzzy (unboosted) score."""
    scored = [(r, _suggest_score(r, r.get("match") != "fuzzy")) for r in local]
    scored += [(r, _suggest_score(r, True)) for r in upstream]
    scored.sort(key=lambda t: t[1], reverse=True)
    out, seen = [], set()
    for r, _ in scored:
        key = ((r.get("name") or "").casefold(), r.get("admin1") or "",
               r.get("country_code") or "")
        if key not in seen:
            seen.add(key)
            out.append(r)
            if len(out) == limit:
                break
    return out


def api_suggest(q: str = Query(..., min_length=1)):
    """Type-ahead location suggestions: the top 5 places for a (possibly
    typo'd) query prefix. The local GeoNames index answers instantly and
    tolerates a single-letter typo; the upstream geocoder fills remaining slots
    with what the index doesn't know (neighbourhoods, postcodes). When both
    come up empty, one query token is respelled against known place-name tokens
    and retried ("pest seattle" → "west seattle"); `corrected` reports the
    respelling that produced the results."""
    q = q.strip()
    if len(q) < 2:
        return {"results": [], "corrected": None}
    local = places.search(q, _SUGGEST_LIMIT)     # None while the index loads
    results = list(local or [])
    upstream_error = None
    # Consult the upstream geocoder unless the local index already answered
    # convincingly: full slots and a substantial place matching the spelling as
    # typed. Fuzzy hits don't count as convincing — they're guesses, and when
    # the query is an alternate name the index doesn't know ("münchen" prefix-
    # matches only small towns; Munich lives upstream), neither those towns nor
    # a coincidental fuzzy big-city hit may suppress the real answer.
    prominent = max((r.get("population") or 0 for r in results
                     if r.get("match") == "prefix"), default=0)
    if len(results) < _SUGGEST_LIMIT or prominent < 100_000:
        try:
            results = _suggest_merge(results, _suggest_upstream(q), _SUGGEST_LIMIT)
        except Exception as e:  # noqa: BLE001 - local results (if any) still serve
            upstream_error = e
    corrected = None
    if not results and len(q) >= 4:
        for phrase in places.corrections(q):
            hits = places.search(phrase, _SUGGEST_LIMIT) or []
            if not hits:
                try:
                    hits = list(_suggest_upstream(phrase))
                except Exception:  # noqa: BLE001 - a failed probe just means no hits
                    hits = []
            if hits:
                results, corrected = hits[:_SUGGEST_LIMIT], phrase
                break
    if not results and local is None and upstream_error is not None:
        raise HTTPException(status_code=502, detail=f"geocoding failed: {upstream_error}")
    return {"results": results[:_SUGGEST_LIMIT], "corrected": corrected}


def api_place(
    lat: float = Query(..., ge=-90, le=90),
    lon: float = Query(..., ge=-180, le=180),
):
    """Best-effort neighbourhood/city label for a point — the same string the view
    endpoints expose as ``place`` (snapped to the grid cell, reverse-geocoded and
    cached). Resolved on its own so the compare page can show a location's name as
    soon as it's added, before its full series loads."""
    if not grid.in_north_america(lat, lon):
        return {"place": None}
    cell = grid.snap(lat, lon)
    with audit.RunAudit(endpoint="place", lat=round(lat, 4), lon=round(lon, 4),
                        cell_id=cell["id"]) as run:
        with run.phase("reverse_geocode"):
            place = climate.reverse_geocode(cell["center_lat"], cell["center_lon"])
    return {"place": place,
            "cell": {"center_lat": cell["center_lat"], "center_lon": cell["center_lon"]}}


def api_grade(
    request: Request,
    lat: float = Query(..., ge=-90, le=90),
    lon: float = Query(..., ge=-180, le=180),
    date: str | None = Query(None, description="target date YYYY-MM-DD (default today)"),
    days: int = Query(14, ge=1, le=60, description="days of history to grade before the target"),
    after: int = Query(14, ge=0, le=14,
                       description="days to grade after the target (observed, or forecast when future; "
                                   "the forecast reaches ~7 days out so future days cap there)"),
):
    target = (
        pd.Timestamp(date).normalize()
        if date
        else pd.Timestamp(datetime.date.today())
    )
    cell = grid.snap(lat, lon)

    with audit.RunAudit(
        endpoint="grade",
        lat=round(lat, 4),
        lon=round(lon, 4),
        target_date=target.date().isoformat(),
        recent_days=days,
        cell_id=cell["id"],
    ) as run:
        try:
            with run.phase("history"):
                history, cache_meta = climate.get_history(cell)
            with run.phase("recent"):
                recent = climate.get_recent_forecast(cell)
        except Exception as e:  # noqa: BLE001
            raise _weather_fetch_error(e)

        if history.empty:
            raise HTTPException(status_code=404, detail="No historical data for this cell.")

        key = f"{target.date().isoformat()}:{days}:{after}"
        token = f"{PAYLOAD_VER}:{_hist_end(history)}:{climate.recent_stamp(cell['id'])}"
        etag = _etag_for("grade", cell["id"], key, token)
        if _not_modified(request, etag):
            run.set(run_type="cache", not_modified=True)
            return Response(status_code=304, headers={"ETag": etag})
        body = store.get_payload("grade", cell["id"], key, token)
        if body is not None:
            run.set(run_type="cache", history_source="store")
            return _json_response(body, etag)

        with run.phase("reverse_geocode"):
            place = climate.reverse_geocode(cell["center_lat"], cell["center_lon"])
        payload = _build_grade(cell, target, days, history, recent, cache_meta, place, run, after=after)
        return _json_response(store.put_payload("grade", cell["id"], key, token, payload), etag)


def api_calendar(
    request: Request,
    lat: float = Query(..., ge=-90, le=90),
    lon: float = Query(..., ge=-180, le=180),
    start: str | None = Query(None, description="first date YYYY-MM-DD (overrides months)"),
    end: str | None = Query(None, description="last date YYYY-MM-DD (default = latest available)"),
    months: int = Query(24, ge=1, le=24, description="months back to grade when no start given"),
):
    """Grade every day over a date range (default: last 2 years) for the calendar view.

    A custom [start, end] range is supported and capped at ~2 years per request;
    the frontend splits longer spans into successive 2-year chunks. Grades against
    the already-cached history record (which reaches to ~6 days ago), so no extra
    fetch is needed. The graded payload is cached in the derived store keyed by
    the clamped span and validated by the record's end date, so a repeat span is
    a database read — across restarts — until new archive days arrive. Returns a
    compact per-day shape (see grading.grade_range). New in API v2.
    """
    cell = grid.snap(lat, lon)

    with audit.RunAudit(
        endpoint="calendar",
        lat=round(lat, 4),
        lon=round(lon, 4),
        recent_days=months * 31,  # approximate span graded
        cell_id=cell["id"],
    ) as run:
        try:
            with run.phase("history"):
                history, cache_meta = climate.get_history(cell)
        except Exception as e:  # noqa: BLE001
            raise _weather_fetch_error(e)
        if history.empty:
            raise HTTPException(status_code=404, detail="No historical data for this cell.")

        start_ts, end_ts = _cal_span(history, start, end, months)
        key = f"{start_ts.date().isoformat()}:{end_ts.date().isoformat()}:{months}"
        token = f"{PAYLOAD_VER}:{_hist_end(history)}"
        etag = _etag_for("calendar", cell["id"], key, token)
        if _not_modified(request, etag):
            run.set(run_type="cache", not_modified=True)
            return Response(status_code=304, headers={"ETag": etag})
        body = store.get_payload("calendar", cell["id"], key, token)
        if body is not None:
            run.set(run_type="cache", history_source="store")
            return _json_response(body, etag)

        with run.phase("reverse_geocode"):
            place = climate.reverse_geocode(cell["center_lat"], cell["center_lon"])
        payload = _build_calendar(cell, history, start_ts, end_ts, months, place, run)
        full = not cache_meta.get("cached", False)
        run.set(run_type="full" if full else "partial",
                history_source="fetch" if full else "cache")
        # Don't persist a payload whose place failed to resolve (a transient reverse-
        # geocode miss) — otherwise the bare-coordinates fallback would stick for the
        # life of the token. Compare loads several cells at once, so this is where a
        # miss is most likely; leaving it uncached lets the next request retry.
        return _json_response(
            store.put_payload("calendar", cell["id"], key, token, payload,
                              cache=place is not None), etag)


def api_day(
    request: Request,
    lat: float = Query(..., ge=-90, le=90),
    lon: float = Query(..., ge=-180, le=180),
    date: str | None = Query(None, description="target date YYYY-MM-DD (default = latest available)"),
):
    """Full percentile breakdown for a single day: the value at every tier boundary
    in that day-of-year's ±7-day window, plus where the observed values land.

    Powers the single-day detail page. Grades against the cached history record;
    for a date newer than the cache it pulls the recent window for the observation
    (those payloads expire hourly — the recent bundle's own cadence — while fully
    archived days stay valid until the record itself advances). New in API v2.
    """
    cell = grid.snap(lat, lon)

    with audit.RunAudit(
        endpoint="day",
        lat=round(lat, 4),
        lon=round(lon, 4),
        cell_id=cell["id"],
    ) as run:
        with run.phase("history"):
            history, cache_meta = climate.get_history(cell)
        if history.empty:
            raise HTTPException(status_code=404, detail="No historical data for this cell.")

        last = pd.Timestamp(history["date"].max()).normalize()
        target = pd.Timestamp(date).normalize() if date else last
        run.set(target_date=target.date().isoformat())

        key = target.date().isoformat()
        token = f"{PAYLOAD_VER}:{_hist_end(history)}"
        if target > last:
            token += f":h{int(time.time() // 3600)}"   # obs from the hourly recent bundle
        etag = _etag_for("day", cell["id"], key, token)
        if _not_modified(request, etag):
            run.set(run_type="cache", not_modified=True)
            return Response(status_code=304, headers={"ETag": etag})
        body = store.get_payload("day", cell["id"], key, token)
        if body is not None:
            run.set(run_type="cache", history_source="store")
            return _json_response(body, etag)

        with run.phase("reverse_geocode"):
            place = climate.reverse_geocode(cell["center_lat"], cell["center_lon"])
        payload = _build_day(cell, history, target, place, run)
        full = not cache_meta.get("cached", False)
        run.set(run_type="full" if full else "partial",
                history_source="fetch" if full else "cache")
        return _json_response(store.put_payload("day", cell["id"], key, token, payload), etag)


def api_forecast(
    request: Request,
    lat: float = Query(..., ge=-90, le=90),
    lon: float = Query(..., ge=-180, le=180),
    days: int = Query(7, ge=1, le=14, description="how many forecast days ahead to grade"),
):
    """Grade the next `days` of forecast weather against local climatology.

    Same shape as /grade (so the frontend renders it identically), but `recent`
    holds the forward forecast — furthest-out day first. The forecast is fetched
    fresh and cached only ~1 hour (see climate.get_recent_forecast) to track
    updates; the graded payload's validity is tied to that fetch stamp, so it
    expires exactly when a new forecast lands. New in API v2.
    """
    cell = grid.snap(lat, lon)

    with audit.RunAudit(endpoint="forecast", lat=round(lat, 4), lon=round(lon, 4),
                        recent_days=days, cell_id=cell["id"]) as run:
        try:
            with run.phase("history"):
                history, _ = climate.get_history(cell)
            with run.phase("forecast"):
                fc = climate.get_recent_forecast(cell)
        except Exception as e:  # noqa: BLE001
            raise _weather_fetch_error(e)

        if history.empty:
            raise HTTPException(status_code=404, detail="No historical data for this cell.")

        today = pd.Timestamp(datetime.date.today())
        key = f"{today.date().isoformat()}:{days}"
        token = f"{PAYLOAD_VER}:{_hist_end(history)}:{climate.recent_stamp(cell['id'])}"
        etag = _etag_for("forecast", cell["id"], key, token)
        if _not_modified(request, etag):
            run.set(run_type="cache", not_modified=True)
            return Response(status_code=304, headers={"ETag": etag})
        body = store.get_payload("forecast", cell["id"], key, token)
        if body is not None:
            run.set(run_type="cache", history_source="store")
            return _json_response(body, etag)

        with run.phase("reverse_geocode"):
            place = climate.reverse_geocode(cell["center_lat"], cell["center_lon"])
        payload = _build_forecast(cell, days, history, fc, today, place, run)
        run.set(run_type="partial")
        return _json_response(store.put_payload("forecast", cell["id"], key, token, payload), etag)


def api_cell(
    request: Request,
    lat: float = Query(..., ge=-90, le=90),
    lon: float = Query(..., ge=-180, le=180),
    prefetch: int = Query(0, ge=0, le=1,
                          description="1 = warm-only: never fetch weather upstream; 204 for a cold cell"),
):
    """One bundle carrying every view's payload for a cell, so the frontend warms
    all views with a single request instead of four.

    Each slice is the EXACT payload its per-view endpoint returns — built by the
    same builders and cached under the same derived-store keys/tokens — paired
    with the etag that endpoint would emit. The client seeds its per-view cache
    from the slices and later revalidates each view individually with
    If-None-Match, so the bundle and the per-view endpoints stay one cache.

    prefetch=1 is the neighbor-warming mode with a hard guarantee: it never
    spends weather-API quota. A cell with no cached archive answers 204 (no
    body), and only the history-derived slices (calendar + latest-day detail)
    are built — grading recent/forecast days needs the hourly upstream bundle.
    Reverse geocoding makes at most one Nominatim call for a never-labeled cell;
    the client staggers neighbor prefetches to respect that service. New in API v2.
    """
    cell = grid.snap(lat, lon)
    today = pd.Timestamp(datetime.date.today())

    with audit.RunAudit(endpoint="cell", lat=round(lat, 4), lon=round(lon, 4),
                        cell_id=cell["id"], prefetch=bool(prefetch)) as run:
        recent = None
        if prefetch:
            history = climate.load_cached_history(cell)
            if history is None or history.empty:
                run.set(run_type="cold-skip")
                return Response(status_code=204)
            cache_meta = {"cached": True}
        else:
            try:
                with run.phase("history"):
                    history, cache_meta = climate.get_history(cell)
                with run.phase("recent"):
                    recent = climate.get_recent_forecast(cell)
            except Exception as e:  # noqa: BLE001
                raise _weather_fetch_error(e)
            if history.empty:
                raise HTTPException(status_code=404, detail="No historical data for this cell.")

        cid = cell["id"]
        hist_end = _hist_end(history)
        last = pd.Timestamp(history["date"].max()).normalize()

        with run.phase("reverse_geocode"):
            place = climate.reverse_geocode(cell["center_lat"], cell["center_lon"])

        def slice_for(kind: str, key: str, token: str, build) -> dict:
            """The endpoint's derived row (or build + store it), paired with the
            etag that endpoint would emit for the same request."""
            etag = _etag_for(kind, cid, key, token)
            data = store.get_json(kind, cid, key, token)
            if data is None:
                data = build()
                store.put_payload(kind, cid, key, token, data)
            return {"etag": etag, "data": data}

        slices = {}
        hist_token = f"{PAYLOAD_VER}:{hist_end}"
        # Calendar: the default last-24-months span — what the calendar view (and
        # the cross-view prefetch) requests first.
        start_ts, end_ts = _cal_span(history, None, None, 24)
        cal_key = f"{start_ts.date().isoformat()}:{end_ts.date().isoformat()}:24"
        slices["calendar"] = slice_for(
            "calendar", cal_key, hist_token,
            lambda: _build_calendar(cell, history, start_ts, end_ts, 24, place, run))

        if prefetch:
            # Latest archived day: its observation comes from history alone, so
            # it's buildable without the (skipped) recent bundle.
            slices["day"] = slice_for(
                "day", last.date().isoformat(), hist_token,
                lambda: _build_day(cell, history, last, place, run))
        else:
            rf_token = f"{PAYLOAD_VER}:{hist_end}:{climate.recent_stamp(cid)}"
            slices["grade"] = slice_for(
                "grade", f"{today.date().isoformat()}:14:14", rf_token,
                lambda: _build_grade(cell, today, 14, history, recent, cache_meta, place, run, after=14))
            slices["forecast"] = slice_for(
                "forecast", f"{today.date().isoformat()}:7", rf_token,
                lambda: _build_forecast(cell, 7, history, recent, today, place, run))
            day_token = hist_token
            if today > last:
                day_token += f":h{int(time.time() // 3600)}"   # obs from the hourly recent bundle
            slices["day"] = slice_for(
                "day", today.date().isoformat(), day_token,
                lambda: _build_day(cell, history, today, place, run, recent=recent))

        # The bundle's identity is the combination of its slices' identities.
        etag = _etag_for("cell", cid, "bundle" + (":p" if prefetch else ""),
                         "|".join(s["etag"] for s in slices.values()))
        if _not_modified(request, etag):
            run.set(run_type="cache", not_modified=True)
            return Response(status_code=304, headers={"ETag": etag})
        run.set(run_type="bundle", slices=sorted(slices))
        payload = {
            "api_version": "v2",
            "cell": cell,
            "place": place,
            "today": today.date().isoformat(),
            "slices": slices,
        }
        # Not stored as its own derived row — the slices already are.
        return _json_response(_encode(payload), etag)


# --- API versioning --------------------------------------------------------
# Non-backward-compatible additions land in a new version; every version stays
# mounted and served simultaneously. v1 is the original contract (grade/geocode);
# v2 adds the calendar. The unversioned /api/* paths are kept as aliases of v1 so
# existing clients keep working.
v1 = APIRouter(tags=["v1"])
v1.add_api_route("/geocode", api_geocode, methods=["GET"])
v1.add_api_route("/grade", api_grade, methods=["GET"])

v2 = APIRouter(tags=["v2"])
v2.add_api_route("/geocode", api_geocode, methods=["GET"])
v2.add_api_route("/suggest", api_suggest, methods=["GET"])
v2.add_api_route("/place", api_place, methods=["GET"])
v2.add_api_route("/grade", api_grade, methods=["GET"])
v2.add_api_route("/calendar", api_calendar, methods=["GET"])
v2.add_api_route("/day", api_day, methods=["GET"])
v2.add_api_route("/forecast", api_forecast, methods=["GET"])
v2.add_api_route("/cell", api_cell, methods=["GET"])

app.include_router(v1, prefix=f"{BASE}/api")        # legacy unversioned == v1 (backward compatible)
app.include_router(v1, prefix=f"{BASE}/api/v1")
app.include_router(v2, prefix=f"{BASE}/api/v2")


# --- static frontend (served under BASE) -----------------------------------
def _page(name):
    """Route handler for one HTML page. Serves the file with its __ORIGIN__
    placeholders (the link-preview/Open Graph tags) filled in as the request's
    scheme://host + BASE — preview crawlers (Discord, Slack, …) need absolute
    URLs, and the host differs between LAN and prod. The scheme comes from
    X-Forwarded-Proto when a reverse proxy (Caddy) fronts the plain-HTTP
    uvicorn; the proxy passes the original Host header through untouched."""
    path = os.path.join(FRONTEND_DIR, name)

    def route(request: Request):
        with open(path, encoding="utf-8") as f:
            html = f.read()
        proto = request.headers.get("x-forwarded-proto") or request.url.scheme
        host = request.headers.get("host") or request.url.netloc
        html = html.replace("__ORIGIN__", f"{proto}://{host}{BASE}")
        etag = f'W/"{hashlib.sha1(html.encode()).hexdigest()[:20]}"'
        if _not_modified(request, etag):
            return Response(status_code=304, headers={"ETag": etag})
        return Response(html, media_type="text/html", headers={"ETag": etag})
    return route


# The un-slashed base redirects to BASE/ so the frontend's relative asset URLs
# resolve correctly. The app stays scoped to BASE and never claims "/", leaving
# the domain root free for another app (e.g. a portfolio) to own via the proxy.
# Pages also answer HEAD — link-preview crawlers probe with it before fetching.
app.add_api_route(BASE, lambda: RedirectResponse(url=f"{BASE}/"), methods=["GET", "HEAD"], include_in_schema=False)
app.add_api_route(f"{BASE}/", _page("index.html"), methods=["GET", "HEAD"], include_in_schema=False)
app.add_api_route(f"{BASE}/calendar", _page("calendar.html"), methods=["GET", "HEAD"], include_in_schema=False)
app.add_api_route(f"{BASE}/day", _page("day.html"), methods=["GET", "HEAD"], include_in_schema=False)
app.add_api_route(f"{BASE}/compare", _page("compare.html"), methods=["GET", "HEAD"], include_in_schema=False)
app.add_api_route(f"{BASE}/legend", _page("legend.html"), methods=["GET", "HEAD"], include_in_schema=False)

# Everything else under BASE (app.js, style.css, nav.js, …) is a static asset.
# Registered last so the explicit page routes above win.
app.mount(BASE, StaticFiles(directory=FRONTEND_DIR), name="static")