GarminHealthLab/backend/services/garmin.py

"""
Garmin sync service.

Pulls daily summaries + activities through the `garminconnect` library and
upserts them. The library and real Garmin credentials are required to actually
run a sync; without them the endpoint reports a clear error instead of
crashing.

Garmin credentials: the app only stores a scrypt/PBKDF2 *hash* of the Garmin
password (so it cannot be recovered), therefore a live sync needs the
plaintext garminEmail/garminPassword supplied in the request body.

On the library's API — these were verified against garminconnect 0.2.8:
  * get_user_summary(cdate)                  -> one day of daily totals
  * get_sleep_data(cdate)                    -> sleep, NOT part of the summary
  * get_hrv_data(cdate)                      -> HRV, also separate
  * get_activities_by_date(start, end)       -> activities in a date range
  * get_activities(start, limit)             -> PAGINATION, not dates
The last two are easy to confuse: `get_activities` takes an offset and a count,
so passing it a date silently asks for activity number "2026-08-23".
"""
import datetime
import os

from db import execute, query_one
from config import DB_TYPE
from services import health

# How many days back a sync reaches.
DEFAULT_SYNC_DAYS = int(os.environ.get("GARMIN_SYNC_DAYS") or 7)


def _set_sync_status(user_id, status, now, **fields):
    cols = ["user_id", "status", "last_sync_time"] + list(fields.keys())
    placeholders = ", ".join(["?"] * len(cols))
    if DB_TYPE == "mariadb":
        updates = ", ".join(f"{c}=VALUES({c})" for c in cols if c != "user_id")
        sql = (
            f"INSERT INTO sync_status ({', '.join(cols)}) VALUES ({placeholders}) "
            f"ON DUPLICATE KEY UPDATE {updates}"
        )
    else:
        updates = ", ".join(f"{c}=excluded.{c}" for c in cols if c != "user_id")
        sql = (
            f"INSERT INTO sync_status ({', '.join(cols)}) VALUES ({placeholders}) "
            f"ON CONFLICT(user_id) DO UPDATE SET {updates}"
        )
    execute(sql, [user_id, status, now] + list(fields.values()))


def get_sync_status(user_id):
    row = query_one("SELECT * FROM sync_status WHERE user_id = ?", [user_id])
    if not row:
        return {
            "status": "idle",
            "lastSyncTime": None,
            "recordsSynced": 0,
            "lastError": None,
        }
    return {
        "status": row["status"],
        "lastSyncTime": row["last_sync_time"],
        "recordsSynced": row["records_synced"],
        "lastError": row["last_error"],
    }


class MFARequired(RuntimeError):
    """Raised when a password login needs a code this process cannot obtain."""


def _is_cn():
    # Selects Garmin's China service, a separate backend with separate
    # accounts. This project tracks an international account.
    return (os.environ.get("GARMIN_IS_CN") or "").lower() in ("1", "true", "yes")


def _import_garmin():
    try:
        from garminconnect import Garmin
    except ImportError:
        raise RuntimeError(
            "GARMIN_LIB_MISSING: 请先运行 `pip install garminconnect` 以启用同步"
        )
    return Garmin


def load_token(user_id):
    row = query_one("SELECT token FROM garmin_tokens WHERE user_id = ?", [user_id])
    return row["token"] if row else None


def save_token(user_id, token, garmin_email=None):
    cols = ["user_id", "token", "garmin_email", "updated_at"]
    placeholders = ", ".join(["?"] * len(cols))
    if DB_TYPE == "mariadb":
        updates = ", ".join(f"{c}=VALUES({c})" for c in cols if c != "user_id")
        sql = (f"INSERT INTO garmin_tokens ({', '.join(cols)}) VALUES ({placeholders}) "
               f"ON DUPLICATE KEY UPDATE {updates}")
    else:
        updates = ", ".join(f"{c}=excluded.{c}" for c in cols if c != "user_id")
        sql = (f"INSERT INTO garmin_tokens ({', '.join(cols)}) VALUES ({placeholders}) "
               f"ON CONFLICT(user_id) DO UPDATE SET {updates}")
    execute(sql, [user_id, token, garmin_email,
                  datetime.datetime.utcnow().isoformat(timespec="seconds")])


def has_token(user_id):
    return load_token(user_id) is not None


def _connect(creds, user_id=None):
    """Obtain a logged-in Garmin client.

    Prefers stored OAuth tokens: an account with two-factor auth cannot be
    logged into from a web worker, because the library asks for the code on
    stdin and there is none (the failure surfaces as
    "EOFError: EOF when reading a line"). Tokens are minted once by
    `garmin_login.py`, which runs in a terminal where a code can be typed.
    """
    Garmin = _import_garmin()
    client = Garmin(is_cn=_is_cn())

    token = load_token(user_id) if user_id else None
    if token:
        client.garth.loads(token)
        # Populates display_name/unit_system and proves the token still works.
        client.garth.refresh_oauth2()
        client.display_name = client.garth.profile["displayName"]
        return client

    if not creds.get("garminPassword"):
        raise RuntimeError("缺少 Garmin 密码，且未找到已保存的登录令牌")

    client.username = creds["garminEmail"]
    client.password = creds["garminPassword"]
    try:
        client.login()
    except EOFError as e:
        # garth's default MFA prompt calls input(); under gunicorn stdin is
        # closed, so it raises EOFError rather than anything descriptive.
        raise MFARequired(
            "该 Garmin 账号开启了两步验证，无法在服务端直接登录。"
            "请在 NAS 上执行一次 `python garmin_login.py` 完成验证并保存令牌。"
        ) from e
    return client


def _num(*values):
    """First value that is a usable number."""
    for v in values:
        if isinstance(v, (int, float)) and not isinstance(v, bool):
            return v
    return None


def _extract_daily(client, date_str):
    """One day of metrics, assembled from the endpoints that carry them.

    Sleep and HRV are separate endpoints in this library — they are not part
    of the daily summary — so a sync that only read the summary would record
    every night as "no sleep data".
    """
    summary = client.get_user_summary(date_str) or {}

    sleep_seconds = None
    sleep_quality = None
    try:
        sleep = (client.get_sleep_data(date_str) or {}).get("dailySleepDTO") or {}
        sleep_seconds = _num(sleep.get("sleepTimeSeconds"))
        sleep_quality = _num(sleep.get("sleepScores", {}).get("overall", {}).get("value")
                             if isinstance(sleep.get("sleepScores"), dict) else None)
    except Exception:
        pass  # a missing night must not abort the whole day

    hrv = None
    try:
        hrv_body = client.get_hrv_data(date_str) or {}
        summary_block = hrv_body.get("hrvSummary") or {}
        hrv = _num(summary_block.get("lastNightAvg"), summary_block.get("weeklyAvg"))
    except Exception:
        pass

    return {
        "date": date_str,
        "steps": _num(summary.get("totalSteps")),
        "heartRate": _num(summary.get("restingHeartRate"),
                          summary.get("averageHeartRate")),
        "heartRateVariability": hrv,
        "sleepDuration": round(sleep_seconds / 3600, 1) if sleep_seconds else None,
        "sleepQuality": sleep_quality,
        "stress": _num(summary.get("averageStressLevel")),
        "caloriesBurned": _num(summary.get("totalKilocalories")),
    }


def _activity_end(start, duration_seconds):
    if not start or not duration_seconds:
        return start
    for fmt in ("%Y-%m-%dT%H:%M:%S", "%Y-%m-%d %H:%M:%S", "%Y-%m-%dT%H:%M:%S.%f"):
        try:
            dt = datetime.datetime.strptime(start[:26], fmt)
            return (dt + datetime.timedelta(seconds=duration_seconds)).isoformat()
        except ValueError:
            continue
    return start


def _sync_activities(client, user_id, start_date, end_date):
    """Fetch the window's activities in one call and store the new ones."""
    activities = client.get_activities_by_date(start_date, end_date) or []
    stored = 0
    for a in activities:
        start = a.get("startTimeLocal") or a.get("startTime")
        activity_type = (
            (a.get("activityType") or {}).get("typeKey")
            if isinstance(a.get("activityType"), dict)
            else a.get("activityType")
        ) or "unknown"
        duration = _num(a.get("duration"))

        # Garmin activity ids are stable, so re-syncing a window must not
        # duplicate what is already stored.
        garmin_id = a.get("activityId")
        if garmin_id is not None:
            existing = query_one(
                "SELECT id FROM activities WHERE user_id = ? AND id = ?",
                [user_id, str(garmin_id)],
            )
            if existing:
                continue

        health.insert_activity(
            user_id,
            {
                "id": str(garmin_id) if garmin_id is not None else None,
                "activityType": activity_type,
                "startTime": start,
                "endTime": _activity_end(start, duration),
                "duration": duration,
                "distance": _num(a.get("distance")),
                "calories": _num(a.get("calories")),
                "heartRateAverage": _num(a.get("averageHR")),
                "heartRateMax": _num(a.get("maxHR")),
            },
        )
        stored += 1
    return stored


def sync_data(user_id, creds, days=None, client=None):
    """Pull the last `days` days from Garmin Connect into the local database.

    `client` exists so tests can inject a stub instead of reaching Garmin.
    """
    days = days or DEFAULT_SYNC_DAYS
    now = datetime.datetime.utcnow().isoformat(timespec="seconds")
    _set_sync_status(user_id, "syncing", now, records_synced=0)

    try:
        client = client or _connect(creds, user_id)
    except Exception as e:
        message = str(e)
        _set_sync_status(user_id, "error", now, records_synced=0, last_error=message)
        return {
            "status": "error",
            "recordsSynced": 0,
            "message": message,
            "mfaRequired": isinstance(e, MFARequired),
            "lastSyncTime": now,
        }

    today = datetime.date.today()
    start_date = (today - datetime.timedelta(days=days - 1)).isoformat()

    days_synced = 0
    day_errors = []
    for i in range(days):
        date_str = (today - datetime.timedelta(days=i)).isoformat()
        try:
            record = _extract_daily(client, date_str)
        except Exception as e:
            day_errors.append(f"{date_str}: {e}")
            continue
        # A day Garmin has no data for comes back all-None; storing it would
        # create an empty row that the metric endpoints then have to filter.
        if any(record[k] is not None for k in record if k != "date"):
            health.upsert_health_daily(user_id, record)
            days_synced += 1

    activities_synced = 0
    try:
        activities_synced = _sync_activities(
            client, user_id, start_date, today.isoformat()
        )
    except Exception as e:
        day_errors.append(f"activities: {e}")

    # Every single day failing means something systemic (expired session,
    # API change) — reporting that as a clean success would hide it.
    if days_synced == 0 and len(day_errors) >= days:
        message = "; ".join(day_errors[:3])
        _set_sync_status(user_id, "error", now, records_synced=0, last_error=message)
        return {"status": "error", "recordsSynced": 0,
                "message": f"同步失败：{message}", "lastSyncTime": now}

    _set_sync_status(
        user_id, "idle", now, records_synced=days_synced,
        last_error="; ".join(day_errors[:3]) if day_errors else None,
    )
    message = f"同步完成，更新 {days_synced} 天数据、{activities_synced} 条运动记录"
    if day_errors:
        message += f"（{len(day_errors)} 天跳过）"
    return {
        "status": "success",
        "recordsSynced": days_synced,
        "activitiesSynced": activities_synced,
        "message": message,
        "lastSyncTime": now,
    }