GarminHealthLab/backend/services/analysis.py

"""
Analysis service: metric trends + a rule-based recommendation engine.

Replicates the original Node AnalysisService logic. Averages are computed over
the most recent 14 days of available daily summaries.
"""
from services import health
from services import ai as ai_svc
from db import query_all

METRIC_COLUMNS = {
    "steps": "steps",
    "heart_rate": "heart_rate",
    "sleep_duration": "sleep_duration",
    "sleep_quality": "sleep_quality",
    "stress": "stress",
    "calories_burned": "calories_burned",
}


def get_trends(metric, user_id, start=None, end=None):
    column = METRIC_COLUMNS.get(metric, "steps")
    params = [user_id]
    sql = "WHERE user_id = ?"
    if start:
        sql += " AND date >= ?"
        params.append(start)
    if end:
        sql += " AND date <= ?"
        params.append(end)
    rows = query_all(
        f"SELECT date, {column} AS value FROM health_data {sql} "
        f"AND {column} IS NOT NULL ORDER BY date ASC",
        params,
    )
    return [{"date": r["date"], "value": r["value"]} for r in rows]


def get_recommendations(user_id):
    recent = health.get_summary(user_id)
    last14 = recent[-14:]
    recs = []

    if not last14:
        return [
            {
                "id": "no-data",
                "category": "数据",
                "recommendation": "暂无健康数据，请先同步你的 Garmin 设备数据。",
                "priority": "low",
                "basedOn": [],
            }
        ]

    avg = lambda key: sum((r.get(key) or 0) for r in last14) / len(last14)

    avg_steps = avg("steps")
    sleep_rows = [r["sleep"]["duration"] for r in last14 if r.get("sleep")]
    avg_sleep = sum(sleep_rows) / len(sleep_rows) if sleep_rows else 0
    avg_stress = avg("stress")
    avg_rhr = avg("heartRate")
    avg_hrv = avg("heartRateVariability")

    if avg_steps > 0 and avg_steps < 8000:
        recs.append({
            "id": "steps",
            "category": "运动",
            "recommendation": f"近 {len(last14)} 天日均步数约 {round(avg_steps)} 步，低于 8000 步目标，建议每天增加 20 分钟快走。",
            "priority": "medium",
            "basedOn": ["steps"],
        })

    if avg_sleep > 0 and avg_sleep < 7:
        recs.append({
            "id": "sleep",
            "category": "睡眠",
            "recommendation": f"日均睡眠约 {avg_sleep:.1f} 小时，偏少。建议固定就寝时间，目标 7-8 小时。",
            "priority": "high",
            "basedOn": ["sleep_duration"],
        })

    if avg_stress > 0 and avg_stress > 50:
        recs.append({
            "id": "stress",
            "category": "压力",
            "recommendation": f"平均压力指数 {round(avg_stress)} 偏高，建议安排放松活动（冥想/散步）。",
            "priority": "high",
            "basedOn": ["stress"],
        })

    if avg_rhr > 0 and avg_rhr > 65:
        recs.append({
            "id": "rhr",
            "category": "心肺",
            "recommendation": f"静息心率约 {round(avg_rhr)} bpm 偏高，规律有氧运动有助于改善心肺功能。",
            "priority": "medium",
            "basedOn": ["heart_rate"],
        })

    if avg_hrv > 0 and avg_hrv < 40:
        recs.append({
            "id": "hrv",
            "category": "恢复",
            "recommendation": f"心率变异性（HRV）约 {round(avg_hrv)} ms 偏低，注意恢复与休息，避免过度训练。",
            "priority": "low",
            "basedOn": ["heart_rate_variability"],
        })

    if not recs:
        recs.append({
            "id": "good",
            "category": "状态",
            "recommendation": "近期各项指标良好，保持当前作息与运动习惯即可。",
            "priority": "low",
            "basedOn": [],
        })

    order = {"high": 0, "medium": 1, "low": 2}
    recs.sort(key=lambda r: order[r["priority"]])
    return recs


def get_ai_recommendations(user_id, model=None, days=None):
    """LLM-generated recommendations over the user's full history.

    Falls back to the rule engine if every model fails, so the endpoint always
    returns something useful. The `source` field tells the two apart.
    """
    summary = health.get_summary(user_id)
    if not summary:
        return {
            "recommendations": get_recommendations(user_id),
            "meta": {"model": None, "source": "rules", "reason": "无健康数据"},
        }

    activities = health.get_activities(user_id)
    budget = days or ai_svc.default_day_budget()

    try:
        recs, meta = ai_svc.generate(
            summary, activities, preferred_model=model, day_budget=budget
        )
        return {"recommendations": recs, "meta": {**meta, "source": "ai"}}
    except ai_svc.AIError as e:
        return {
            "recommendations": get_recommendations(user_id),
            "meta": {"model": None, "source": "rules", "reason": str(e)},
        }