睡眠智能决策系统 (Sleep Intelligence System)一、实际应用场景描述作为一名全栈开发工程师兼技术布道博主我的工作节奏充满挑战- 深夜赶项目截止日期凌晨2点才睡- 早上7点被闹钟叫醒强撑着去上班- 上午10点开始犯困喝第3杯咖啡- 下午3点效率断崖式下跌错误率飙升- 晚上10点又累又睡不着形成恶性循环传统观念告诉我们睡眠越长越好、周末补觉就行但现实是- 睡9小时反而更疲惫像睡晕了- 周末补觉4小时周一更难起床- 明明睡够了8小时白天依然昏沉- 越想多睡越容易陷入越睡越困的怪圈二、引入痛点1. 睡眠越多越好的认知误区- 线性思维陷阱认为睡眠时长和清醒度是简单的正比关系- 忽视个体差异每个人的最佳睡眠时长不同从6-9小时不等- 混淆数量与质量8小时低质量睡眠 6.5小时高质量睡眠- 补觉无效论周末补觉无法真正修复工作日累积的睡眠债2. 赖床习惯的隐性成本- 睡眠惯性(Sleep Inertia)醒来后30-60分钟内认知功能受损- 生物钟紊乱反复按掉闹钟破坏昼夜节律- 时间浪费多睡30分钟需要1小时才能完全清醒- 心理依赖形成需要更多睡眠才能清醒的错误认知3. 现有解决方案的局限- 睡眠App只记录不分析告诉你睡了多久不告诉你为什么累- 一刀切建议成年人需要7-9小时忽略个人差异- 缺乏行动指导知道问题所在不知道具体怎么做三、核心逻辑讲解1. 睡眠-清醒度非线性模型基于睡眠科学建立倒U型曲线关系清醒度 f(睡眠时长) a * sin(b * 睡眠时长 c) d- 过短(6h)严重睡眠剥夺清醒度急剧下降- 最佳区间(6-9h)清醒度达到峰值- 过长(9h)睡眠惯性增强清醒度反降2. 多因子影响模型清醒度受多维度因素影响- 睡眠时长基础决定因素- 入睡时间影响REM和深睡比例- 睡眠阶段分布深睡修复身体REM整理记忆- 个人生物钟类型云雀型(早鸟)vs猫头鹰型(夜猫)- 前一晚睡眠质量入睡速度、夜间觉醒次数3. 智能决策算法采用曲线拟合 强化学习方法- 收集个人睡眠-清醒度数据- 拟合个性化睡眠-清醒度曲线- 通过试错学习找到最优睡眠方案- 动态调整建议适应用户变化4. 反直觉的赖床干预基于行为经济学和习惯养成理论- 用损失厌恶替代获得快乐作为动力- 设计渐进式唤醒方案- 建立早起奖励机制- 利用社会承诺增加执行力四、代码模块化项目结构sleep_intelligence/├── main.py # 主程序入口├── core/│ ├── __init__.py│ ├── sleep_analyzer.py # 睡眠数据分析器│ ├── curve_fitter.py # 曲线拟合引擎│ ├── decision_engine.py # 智能决策核心│ └── habit_intervention.py # 习惯干预模块├── models/│ ├── __init__.py│ ├── sleep_data.py # 睡眠数据模型│ ├── circadian.py # 生物钟模型│ └── user_profile.py # 用户画像模型├── utils/│ ├── __init__.py│ ├── data_collector.py # 数据收集器│ ├── visualizer.py # 可视化工具│ └── alarms.py # 智能闹钟├── config.py # 配置文件├── requirements.txt # 依赖包└── README.md # 说明文档核心代码实现config.py - 配置文件睡眠智能决策系统 - 配置文件智能决策要点参数可配置化支持个性化定制from dataclasses import dataclass, fieldfrom typing import Dict, List, Tuple, Optionalfrom enum import Enumimport mathclass Chronotype(Enum):生物钟类型LARK lark # 云雀型早睡早起OWL owl # 猫头鹰型晚睡晚起DOLPHIN dolphin # 海豚型睡眠浅易醒BEAR bear # 熊型随大流占人口40%dataclassclass SleepStageWeights:睡眠阶段权重配置影响清醒度计算的不同阶段权重deep_sleep_weight: float 0.4 # 深睡(慢波睡眠) - 身体修复rem_sleep_weight: float 0.35 # REM睡眠 - 记忆整理light_sleep_weight: float 0.15 # 浅睡 - 过渡阶段awake_during_night: float -0.1 # 夜间觉醒 - 负面影响dataclassclass CircadianParams:生物钟参数配置基于研究数据的默认参数# 褪黑素分泌曲线参数melatonin_onset_hour: float 21.0 # 褪黑素开始分泌时间melatonin_peak_hour: float 2.0 # 褪黑素峰值时间cortisol_onset_hour: float 6.0 # 皮质醇开始上升时间# 体温最低点(通常在深睡期)body_temp_min_hour: float 4.0# 认知功能恢复时间(起床后)cognitive_full_hours: float 1.5 # 完全清醒需要的时间# 社会时差影响(工作日vs周末)social_jetlag_hours: float 1.5dataclassclass CurveFitParams:睡眠-清醒度曲线拟合参数初始默认值将通过机器学习个性化调整# 倒U型曲线参数peak_duration: float 7.5 # 峰值睡眠时长amplitude: float 40.0 # 曲线振幅baseline: float 60.0 # 基线清醒度steepness: float 1.2 # 曲线陡峭度# 非线性修正因子oversleep_penalty: float 3.0 # 超睡惩罚系数undersleep_penalty: float 5.0 # 欠睡惩罚系数# 个体差异因子chronotype_modifier: Dict[Chronotype, float] field(default_factorylambda: {Chronotype.LARK: -0.5, # 云雀型峰值更早更短Chronotype.OWL: 0.8, # 猫头鹰型峰值更晚更长Chronotype.DOLPHIN: -0.3, # 海豚型峰值较短Chronotype.BEAR: 0.0 # 熊型接近平均})dataclassclass InterventionParams:习惯干预参数配置# 渐进式唤醒参数gradual_wake_duration: int 30 # 渐进唤醒时长(分钟)wake_light_intensity_start: int 10 # 起始光照强度(%)wake_light_intensity_end: int 100 # 结束光照强度(%)# 睡眠债计算参数sleep_debt_half_life_hours: float 24 # 睡眠债半衰期max_sleep_debt_hours: float 20 # 最大睡眠债# 奖励机制参数streak_bonus_multiplier: float 1.5 # 连续早起奖励倍数weekly_target_bonus: float 50.0 # 周目标达成奖励# 惩罚机制参数snooze_penalty_points: int 10 # 每次贪睡扣分missed_target_penalty: int 30 # 未达成目标扣分dataclassclass UserProfile:用户画像配置捕获个体差异实现个性化决策name: str Default Userchronotype: Chronotype Chronotype.BEARage: int 28gender: str neutral# 生理参数avg_sleep_latency_minutes: float 15.0 # 平均入睡时间avg_wake_transitions: int 2 # 平均夜间觉醒次数preferred_bedtime_hour: float 23.0 # 偏好入睡时间preferred_waketime_hour: float 7.0 # 偏好起床时间# 目标设置target_wake_time: float 7.0 # 目标起床时间min_acceptable_sleep: float 6.0 # 最小可接受睡眠max_acceptable_sleep: float 9.0 # 最大可接受睡眠# 行为参数current_sleep_debt: float 0.0 # 当前睡眠债intervention_stage: int 1 # 干预阶段(1-5)motivation_level: float 0.5 # 动机水平(0-1)# 历史数据sleep_history: List[Dict] field(default_factorylist)feedback_history: List[Dict] field(default_factorylist)# 预定义睡眠环境建议SLEEP_ENVIRONMENT_TIPS {duration_optimization: [保持固定的睡眠时间即使周末也尽量一致,在睡前90分钟避免蓝光(手机/电脑),卧室温度保持在18-22°C,使用遮光窗帘确保完全黑暗],chronotype_adaptation: {lark: [利用早晨高效时段处理重要工作,避免晚上9点后摄入咖啡因,将社交活动安排在上午],owl: [争取弹性工作时间晚1-2小时开始工作,将重要任务安排在下午/晚上,使用强光疗法帮助晨间清醒],dolphin: [建立严格的睡前放松仪式,避免下午3点后摄入咖啡因,考虑分两次睡眠(午睡20-30分钟)],bear: [跟随自然光周期调整作息,保持规律运动但避免睡前3小时,晚餐后2小时再睡觉]},wake_optimization: [将闹钟放在必须下床才能关掉的位置,起床后立即接触自然光或强光源,喝一杯温水启动新陈代谢,进行5分钟轻度运动(伸展/快走)]}# 睡眠-清醒度曲线特征点(用于初始拟合)INITIAL_CURVE_DATA {4.0: 25, # 4小时严重缺觉5.0: 40, # 5小时明显困倦6.0: 65, # 6小时基本清醒6.5: 75, # 6.5小时较清醒7.0: 82, # 7小时很清醒7.5: 85, # 7.5小时峰值8.0: 83, # 8小时略降8.5: 78, # 8.5小时开始下降9.0: 70, # 9小时明显下降9.5: 60, # 9.5小时困倦10.0: 50 # 10小时睡晕了}models/sleep_data.py - 睡眠数据模型睡眠数据模型模块智能决策要点将复杂的睡眠过程抽象为可量化的数据对象from dataclasses import dataclass, fieldfrom datetime import datetime, timedeltafrom typing import List, Optional, Dictimport uuidimport mathfrom enum import Enumimport jsonfrom config import Chronotype, SleepStageWeights, CircadianParamsclass SleepQuality(Enum):睡眠质量评级EXCELLENT excellent # 90-100分GOOD good # 75-89分FAIR fair # 60-74分POOR poor # 40-59分VERY_POOR very_poor # 0-39分dataclassclass SleepStage:单次睡眠阶段记录属性:stage_type: 阶段类型 (N1/N2/N3/REM/awake)start_time: 开始时间end_time: 结束时间duration_minutes: 持续时间(分钟)stage_type: str # N1(浅睡1), N2(浅睡2), N3(深睡), REM, awakestart_time: datetimeend_time: datetimeduration_minutes: float field(initFalse)def __post_init__(self):self.duration_minutes (self.end_time - self.start_time).total_seconds() / 60propertydef duration_hours(self) - float:return self.duration_minutes / 60def to_dict(self) - Dict:return {stage_type: self.stage_type,start_time: self.start_time.strftime(%Y-%m-%d %H:%M:%S),end_time: self.end_time.strftime(%Y-%m-%d %H:%M:%S),duration_minutes: round(self.duration_minutes, 1)}dataclassclass SleepSession:单次睡眠会话记录属性:session_id: 会话唯一标识date: 睡眠日期(以入睡日期为准)bedtime: 上床时间wake_time: 起床时间sleep_latency: 入睡潜伏期(分钟)stages: 睡眠阶段列表interruptions: 夜间觉醒次数total_sleep_time: 总睡眠时间(小时)sleep_efficiency: 睡眠效率(%)sleep_quality_score: 睡眠质量评分(0-100)session_id: str field(default_factorylambda: str(uuid.uuid4())[:8])date: datetime field(default_factorydatetime.now)bedtime: datetime field(default_factorydatetime.now)wake_time: datetime field(default_factorylambda: datetime.now() timedelta(hours7))sleep_latency: float 15.0 # 入睡时间(分钟)stages: List[SleepStage] field(default_factorylist)interruptions: int 0total_sleep_time: float 0.0sleep_efficiency: float 0.0sleep_quality_score: float 0.0def __post_init__(self):if self.total_sleep_time 0.0 and self.stages:self._calculate_totals()def add_stage(self, stage: SleepStage) - None:添加睡眠阶段self.stages.append(stage)self._calculate_totals()def _calculate_totals(self) - None:计算睡眠总时间和效率if not self.stages:return# 计算总睡眠时间(排除入睡潜伏期和夜间觉醒)self.total_sleep_time sum(stage.duration_hours for stage in self.stagesif stage.stage_type in [N1, N2, N3, REM])# 计算在床总时间time_in_bed (self.wake_time - self.bedtime).total_seconds() / 3600# 计算睡眠效率if time_in_bed 0:self.sleep_efficiency (self.total_sleep_time / time_in_bed) * 100# 计算睡眠质量评分self.sleep_quality_score self._calculate_quality_score()def _calculate_quality_score(self) - float:基于多维度计算睡眠质量评分if not self.stages:return 50.0# 各阶段占比total_time self.total_sleep_timeif total_time 0:return 50.0stage_durations {}for stage in self.stages:if stage.stage_type in stage_durations:stage_durations[stage.stage_type] stage.duration_hourselse:stage_durations[stage.stage_type] stage.duration_hours# 理想占比(成人平均值)ideal_proportions {N3: 0.20, # 深睡20%REM: 0.25, # REM睡眠25%N1: 0.05, # N1浅睡5%N2: 0.50 # N2浅睡50%}# 计算偏离度deviation_score 0for stage_type, ideal_prop in ideal_proportions.items():actual_prop stage_durations.get(stage_type, 0) / total_timedeviation_score abs(ideal_prop - actual_prop)# 睡眠效率权重efficiency_component (self.sleep_efficiency / 100) * 30# 入睡潜伏期权重latency_component max(0, (30 - self.sleep_latency) / 30) * 20# 觉醒次数权重interruption_component max(0, (5 - self.interruptions) / 5) * 10# 总分计算base_score (1 - deviation_score / 2) * 40 # 阶段占比得分final_score base_score efficiency_component latency_component interruption_componentreturn round(min(100, max(0, final_score)), 1)propertydef sleep_quality_rating(self) - SleepQuality:获取睡眠质量评级score self.sleep_quality_scoreif score 90:return SleepQuality.EXCELLENTelif score 75:return SleepQuality.GOODelif score 60:return SleepQuality.FAIRelif score 40:return SleepQuality.POORelse:return SleepQuality.VERY_POORpropertydef bedtime_hour(self) - float:获取入睡时间的小时数return self.bedtime.hour self.bedtime.minute / 60propertydef wake_time_hour(self) - float:获取起床时间的小时数return self.wake_time.hour self.wake_time.minute / 60propertydef sleep_duration_hours(self) - float:获取睡眠时长(小时)return self.total_sleep_timepropertydef deep_sleep_ratio(self) - float:深睡占比deep_sleep sum(s.duration_hours for s in self.stages if s.stage_type N3)return deep_sleep / self.total_sleep_time if self.total_sleep_time 0 else 0propertydef rem_sleep_ratio(self) - float:REM睡眠占比rem_sleep sum(s.duration_hours for s in self.stages if s.stage_type REM)return rem_sleep / self.total_sleep_time if self.total_sleep_time 0 else 0def to_dict(self) - Dict:return {session_id: self.session_id,date: self.date.strftime(%Y-%m-%d),bedtime: self.bedtime.strftime(%Y-%m-%d %H:%M:%S),wake_time: self.wake_time.strftime(%Y-%m-%d %H:%M:%S),bedtime_hour: round(self.bedtime_hour, 2),wake_time_hour: round(self.wake_time_hour, 2),sleep_duration_hours: round(self.sleep_duration_hours, 2),sleep_latency: round(self.sleep_latency, 1),interruptions: self.interruptions,sleep_efficiency: round(self.sleep_efficiency, 1),sleep_quality_score: self.sleep_quality_score,sleep_quality_rating: self.sleep_quality_rating.value,deep_sleep_ratio: round(self.deep_sleep_ratio, 3),rem_sleep_ratio: round(self.rem_sleep_ratio, 3),stages: [s.to_dict() for s in self.stages]}classmethoddef from_dict(cls, data: Dict) - SleepSession:从字典创建睡眠会话session cls(session_iddata.get(session_id, str(uuid.uuid4())[:8]),datedatetime.strptime(data[date], %Y-%m-%d),bedtimedatetime.strptime(data[bedtime], %Y-%m-%d %H:%M:%S),wake_timedatetime.strptime(data[wake_time], %Y-%m-%d %H:%M:%S),sleep_latencydata.get(sleep_latency, 15.0),interruptionsdata.get(interruptions, 0))# 重建睡眠阶段for stage_data in data.get(stages, []):stage SleepStage(stage_typestage_data[stage_type],start_timedatetime.strptime(stage_data[start_time], %Y-%m-%d %H:%M:%S),end_timedatetime.strptime(stage_data[end_time], %Y-%m-%d %H:%M:%S))session.add_stage(stage)return sessiondataclassclass WakefulnessRecord:清醒度记录属性:record_id: 记录唯一标识timestamp: 记录时间sleep_session_id: 对应的睡眠会话IDtime_since_wake: 距起床时间(小时)alertness_level: 清醒度水平(0-100)energy_level: 能量水平(0-100)mood: 情绪状态(-50到50)cognitive_performance: 认知表现(0-100)physical_performance: 身体表现(0-100)notes: 备注record_id: str field(default_factorylambda: str(uuid.uuid4())[:8])timestamp: datetime field(default_factorydatetime.now)sleep_session_id: str time_since_wake: float 0.0alertness_level: float 50.0energy_level: float 50.0mood: float 0.0cognitive_performance: float 50.0physical_performance: float 50.0notes: str def to_dict(self) - Dict:return {record_id: self.record_id,timestamp: self.timestamp.strftime(%Y-%m-%d %H:%M:%S),sleep_session_id: self.sleep_session_id,time_since_wake: round(self.time_since_wake, 2),alertness_level: round(self.alertness_level, 1),energy_level: round(self.energy_level, 1),mood: round(self.mood, 1),cognitive_performance: round(self.cognitive_performance, 1),physical_performance: round(self.physical_performance, 1),notes: self.notes}classmethoddef from_dict(cls, data: Dict) - WakefulnessRecord:return cls(**{k: (datetime.strptime(v, %Y-%m-%d %H:%M:%S) if k in [timestamp] and isinstance(v, str) else v)for k, v in data.items()})dataclassclass DailyWakefulnessData:每日清醒度数据汇总属性:date: 日期sleep_session: 当天的睡眠会话wakefulness_records: 清醒度记录列表average_alertness: 平均清醒度peak_alertness: 峰值清醒度peak_alertness_time: 峰值清醒度时间time_to_peak: 达到峰值时间sleep_inertia_duration: 睡眠惯性持续时间date: datetimesleep_session: SleepSessionwakefulness_records: List[WakefulnessRecord] field(default_factorylist)def __post_init__(self):if self.wakefulness_records:self._calculate_aggregates()def add_wakefulness_record(self, record: WakefulnessRecord) - None:添加清醒度记录self.wakefulness_records.append(record)self._calculate_aggregates()def _calculate_aggregates(self) - None:计算聚合指标if not self.wakefulness_records:return# 平均清醒度self.average_alertness sum(r.alertness_level for r in self.wakefulness_records) / \len(self.wakefulness_records)# 峰值清醒度max_record max(self.wakefulness_records, keylambda r: r.alertness_level)self.peak_alertness max_record.alertness_levelself.peak_alertness_time max_record.time_since_wake# 达到峰值时间(首次达到90%峰值的时刻)threshold self.peak_alertness * 0.9for record in sorted(self.wakefulness_records, keylambda r: r.time_since_wake):if record.alertness_level threshold:self.time_to_peak record.time_since_wakebreakelse:self.time_to_peak 2.0 # 默认2小时# 睡眠惯性持续时间(清醒度80%的持续时间)inertia_records [r for r in self.wakefulness_records if r.alertness_level 80]if inertia_records:self.sleep_inertia_duration max(r.time_since_wake for r in inertia_records)else:self.sleep_inertia_duration 0.5def to_dict(self) - Dict:return {date: self.date.strftime(%Y-%m-%d),sleep_session_id: self.sleep_session.session_id,sleep_duration: round(self.sleep_session.sleep_duration_hours, 2),wake_time: self.sleep_session.wake_time_hour,average_alertness: round(self.average_alertness, 1),peak_alertness: round(self.peak_alertness, 1),peak_alertness_time: round(self.peak_alertness_time, 2),time_to_peak: round(self.time_to_peak, 2),sleep_inertia_duration: round(self.sleep_inertia_duration, 2),records_count: len(self.wakefulness_records)}class SleepDatabase:睡眠数据库管理类智能决策要点结构化存储和管理时间序列睡眠数据def __init__(self, storage_path: str sleep_data.json):self.storage_path storage_pathself.sleep_sessions: Dict[str, SleepSession] {}self.wakefulness_records: Dict[str, WakefulnessRecord] {}self.daily_data: Dict[str, DailyWakefulnessData] {}self.load_data()def load_data(self) - None:加载历史数据try:if os.path.exists(self.storage_path):with open(self.storage_path, r, encodingutf-8) as f:data json.load(f)# 加载睡眠会话for session_data in data.get(sleep_sessions, []):session SleepSession.from_dict(session_data)self.sleep_sessions[session.session_id] session# 加载清醒度记录for record_data in data.get(wakefulness_records, []):record WakefulnessRecord.fr利用AI解决实际问题如果你觉得这个工具好用欢迎关注长安牧笛
颠覆“睡眠越长越好”，拟合睡眠时长与清醒曲线，颠覆赖床习惯，找到最高精力睡眠方案。
睡眠智能决策系统 (Sleep Intelligence System)一、实际应用场景描述作为一名全栈开发工程师兼技术布道博主我的工作节奏充满挑战- 深夜赶项目截止日期凌晨2点才睡- 早上7点被闹钟叫醒强撑着去上班- 上午10点开始犯困喝第3杯咖啡- 下午3点效率断崖式下跌错误率飙升- 晚上10点又累又睡不着形成恶性循环传统观念告诉我们睡眠越长越好、周末补觉就行但现实是- 睡9小时反而更疲惫像睡晕了- 周末补觉4小时周一更难起床- 明明睡够了8小时白天依然昏沉- 越想多睡越容易陷入越睡越困的怪圈二、引入痛点1. 睡眠越多越好的认知误区- 线性思维陷阱认为睡眠时长和清醒度是简单的正比关系- 忽视个体差异每个人的最佳睡眠时长不同从6-9小时不等- 混淆数量与质量8小时低质量睡眠 6.5小时高质量睡眠- 补觉无效论周末补觉无法真正修复工作日累积的睡眠债2. 赖床习惯的隐性成本- 睡眠惯性(Sleep Inertia)醒来后30-60分钟内认知功能受损- 生物钟紊乱反复按掉闹钟破坏昼夜节律- 时间浪费多睡30分钟需要1小时才能完全清醒- 心理依赖形成需要更多睡眠才能清醒的错误认知3. 现有解决方案的局限- 睡眠App只记录不分析告诉你睡了多久不告诉你为什么累- 一刀切建议成年人需要7-9小时忽略个人差异- 缺乏行动指导知道问题所在不知道具体怎么做三、核心逻辑讲解1. 睡眠-清醒度非线性模型基于睡眠科学建立倒U型曲线关系清醒度 f(睡眠时长) a * sin(b * 睡眠时长 c) d- 过短(6h)严重睡眠剥夺清醒度急剧下降- 最佳区间(6-9h)清醒度达到峰值- 过长(9h)睡眠惯性增强清醒度反降2. 多因子影响模型清醒度受多维度因素影响- 睡眠时长基础决定因素- 入睡时间影响REM和深睡比例- 睡眠阶段分布深睡修复身体REM整理记忆- 个人生物钟类型云雀型(早鸟)vs猫头鹰型(夜猫)- 前一晚睡眠质量入睡速度、夜间觉醒次数3. 智能决策算法采用曲线拟合 强化学习方法- 收集个人睡眠-清醒度数据- 拟合个性化睡眠-清醒度曲线- 通过试错学习找到最优睡眠方案- 动态调整建议适应用户变化4. 反直觉的赖床干预基于行为经济学和习惯养成理论- 用损失厌恶替代获得快乐作为动力- 设计渐进式唤醒方案- 建立早起奖励机制- 利用社会承诺增加执行力四、代码模块化项目结构sleep_intelligence/├── main.py # 主程序入口├── core/│ ├── __init__.py│ ├── sleep_analyzer.py # 睡眠数据分析器│ ├── curve_fitter.py # 曲线拟合引擎│ ├── decision_engine.py # 智能决策核心│ └── habit_intervention.py # 习惯干预模块├── models/│ ├── __init__.py│ ├── sleep_data.py # 睡眠数据模型│ ├── circadian.py # 生物钟模型│ └── user_profile.py # 用户画像模型├── utils/│ ├── __init__.py│ ├── data_collector.py # 数据收集器│ ├── visualizer.py # 可视化工具│ └── alarms.py # 智能闹钟├── config.py # 配置文件├── requirements.txt # 依赖包└── README.md # 说明文档核心代码实现config.py - 配置文件睡眠智能决策系统 - 配置文件智能决策要点参数可配置化支持个性化定制from dataclasses import dataclass, fieldfrom typing import Dict, List, Tuple, Optionalfrom enum import Enumimport mathclass Chronotype(Enum):生物钟类型LARK lark # 云雀型早睡早起OWL owl # 猫头鹰型晚睡晚起DOLPHIN dolphin # 海豚型睡眠浅易醒BEAR bear # 熊型随大流占人口40%dataclassclass SleepStageWeights:睡眠阶段权重配置影响清醒度计算的不同阶段权重deep_sleep_weight: float 0.4 # 深睡(慢波睡眠) - 身体修复rem_sleep_weight: float 0.35 # REM睡眠 - 记忆整理light_sleep_weight: float 0.15 # 浅睡 - 过渡阶段awake_during_night: float -0.1 # 夜间觉醒 - 负面影响dataclassclass CircadianParams:生物钟参数配置基于研究数据的默认参数# 褪黑素分泌曲线参数melatonin_onset_hour: float 21.0 # 褪黑素开始分泌时间melatonin_peak_hour: float 2.0 # 褪黑素峰值时间cortisol_onset_hour: float 6.0 # 皮质醇开始上升时间# 体温最低点(通常在深睡期)body_temp_min_hour: float 4.0# 认知功能恢复时间(起床后)cognitive_full_hours: float 1.5 # 完全清醒需要的时间# 社会时差影响(工作日vs周末)social_jetlag_hours: float 1.5dataclassclass CurveFitParams:睡眠-清醒度曲线拟合参数初始默认值将通过机器学习个性化调整# 倒U型曲线参数peak_duration: float 7.5 # 峰值睡眠时长amplitude: float 40.0 # 曲线振幅baseline: float 60.0 # 基线清醒度steepness: float 1.2 # 曲线陡峭度# 非线性修正因子oversleep_penalty: float 3.0 # 超睡惩罚系数undersleep_penalty: float 5.0 # 欠睡惩罚系数# 个体差异因子chronotype_modifier: Dict[Chronotype, float] field(default_factorylambda: {Chronotype.LARK: -0.5, # 云雀型峰值更早更短Chronotype.OWL: 0.8, # 猫头鹰型峰值更晚更长Chronotype.DOLPHIN: -0.3, # 海豚型峰值较短Chronotype.BEAR: 0.0 # 熊型接近平均})dataclassclass InterventionParams:习惯干预参数配置# 渐进式唤醒参数gradual_wake_duration: int 30 # 渐进唤醒时长(分钟)wake_light_intensity_start: int 10 # 起始光照强度(%)wake_light_intensity_end: int 100 # 结束光照强度(%)# 睡眠债计算参数sleep_debt_half_life_hours: float 24 # 睡眠债半衰期max_sleep_debt_hours: float 20 # 最大睡眠债# 奖励机制参数streak_bonus_multiplier: float 1.5 # 连续早起奖励倍数weekly_target_bonus: float 50.0 # 周目标达成奖励# 惩罚机制参数snooze_penalty_points: int 10 # 每次贪睡扣分missed_target_penalty: int 30 # 未达成目标扣分dataclassclass UserProfile:用户画像配置捕获个体差异实现个性化决策name: str Default Userchronotype: Chronotype Chronotype.BEARage: int 28gender: str neutral# 生理参数avg_sleep_latency_minutes: float 15.0 # 平均入睡时间avg_wake_transitions: int 2 # 平均夜间觉醒次数preferred_bedtime_hour: float 23.0 # 偏好入睡时间preferred_waketime_hour: float 7.0 # 偏好起床时间# 目标设置target_wake_time: float 7.0 # 目标起床时间min_acceptable_sleep: float 6.0 # 最小可接受睡眠max_acceptable_sleep: float 9.0 # 最大可接受睡眠# 行为参数current_sleep_debt: float 0.0 # 当前睡眠债intervention_stage: int 1 # 干预阶段(1-5)motivation_level: float 0.5 # 动机水平(0-1)# 历史数据sleep_history: List[Dict] field(default_factorylist)feedback_history: List[Dict] field(default_factorylist)# 预定义睡眠环境建议SLEEP_ENVIRONMENT_TIPS {duration_optimization: [保持固定的睡眠时间即使周末也尽量一致,在睡前90分钟避免蓝光(手机/电脑),卧室温度保持在18-22°C,使用遮光窗帘确保完全黑暗],chronotype_adaptation: {lark: [利用早晨高效时段处理重要工作,避免晚上9点后摄入咖啡因,将社交活动安排在上午],owl: [争取弹性工作时间晚1-2小时开始工作,将重要任务安排在下午/晚上,使用强光疗法帮助晨间清醒],dolphin: [建立严格的睡前放松仪式,避免下午3点后摄入咖啡因,考虑分两次睡眠(午睡20-30分钟)],bear: [跟随自然光周期调整作息,保持规律运动但避免睡前3小时,晚餐后2小时再睡觉]},wake_optimization: [将闹钟放在必须下床才能关掉的位置,起床后立即接触自然光或强光源,喝一杯温水启动新陈代谢,进行5分钟轻度运动(伸展/快走)]}# 睡眠-清醒度曲线特征点(用于初始拟合)INITIAL_CURVE_DATA {4.0: 25, # 4小时严重缺觉5.0: 40, # 5小时明显困倦6.0: 65, # 6小时基本清醒6.5: 75, # 6.5小时较清醒7.0: 82, # 7小时很清醒7.5: 85, # 7.5小时峰值8.0: 83, # 8小时略降8.5: 78, # 8.5小时开始下降9.0: 70, # 9小时明显下降9.5: 60, # 9.5小时困倦10.0: 50 # 10小时睡晕了}models/sleep_data.py - 睡眠数据模型睡眠数据模型模块智能决策要点将复杂的睡眠过程抽象为可量化的数据对象from dataclasses import dataclass, fieldfrom datetime import datetime, timedeltafrom typing import List, Optional, Dictimport uuidimport mathfrom enum import Enumimport jsonfrom config import Chronotype, SleepStageWeights, CircadianParamsclass SleepQuality(Enum):睡眠质量评级EXCELLENT excellent # 90-100分GOOD good # 75-89分FAIR fair # 60-74分POOR poor # 40-59分VERY_POOR very_poor # 0-39分dataclassclass SleepStage:单次睡眠阶段记录属性:stage_type: 阶段类型 (N1/N2/N3/REM/awake)start_time: 开始时间end_time: 结束时间duration_minutes: 持续时间(分钟)stage_type: str # N1(浅睡1), N2(浅睡2), N3(深睡), REM, awakestart_time: datetimeend_time: datetimeduration_minutes: float field(initFalse)def __post_init__(self):self.duration_minutes (self.end_time - self.start_time).total_seconds() / 60propertydef duration_hours(self) - float:return self.duration_minutes / 60def to_dict(self) - Dict:return {stage_type: self.stage_type,start_time: self.start_time.strftime(%Y-%m-%d %H:%M:%S),end_time: self.end_time.strftime(%Y-%m-%d %H:%M:%S),duration_minutes: round(self.duration_minutes, 1)}dataclassclass SleepSession:单次睡眠会话记录属性:session_id: 会话唯一标识date: 睡眠日期(以入睡日期为准)bedtime: 上床时间wake_time: 起床时间sleep_latency: 入睡潜伏期(分钟)stages: 睡眠阶段列表interruptions: 夜间觉醒次数total_sleep_time: 总睡眠时间(小时)sleep_efficiency: 睡眠效率(%)sleep_quality_score: 睡眠质量评分(0-100)session_id: str field(default_factorylambda: str(uuid.uuid4())[:8])date: datetime field(default_factorydatetime.now)bedtime: datetime field(default_factorydatetime.now)wake_time: datetime field(default_factorylambda: datetime.now() timedelta(hours7))sleep_latency: float 15.0 # 入睡时间(分钟)stages: List[SleepStage] field(default_factorylist)interruptions: int 0total_sleep_time: float 0.0sleep_efficiency: float 0.0sleep_quality_score: float 0.0def __post_init__(self):if self.total_sleep_time 0.0 and self.stages:self._calculate_totals()def add_stage(self, stage: SleepStage) - None:添加睡眠阶段self.stages.append(stage)self._calculate_totals()def _calculate_totals(self) - None:计算睡眠总时间和效率if not self.stages:return# 计算总睡眠时间(排除入睡潜伏期和夜间觉醒)self.total_sleep_time sum(stage.duration_hours for stage in self.stagesif stage.stage_type in [N1, N2, N3, REM])# 计算在床总时间time_in_bed (self.wake_time - self.bedtime).total_seconds() / 3600# 计算睡眠效率if time_in_bed 0:self.sleep_efficiency (self.total_sleep_time / time_in_bed) * 100# 计算睡眠质量评分self.sleep_quality_score self._calculate_quality_score()def _calculate_quality_score(self) - float:基于多维度计算睡眠质量评分if not self.stages:return 50.0# 各阶段占比total_time self.total_sleep_timeif total_time 0:return 50.0stage_durations {}for stage in self.stages:if stage.stage_type in stage_durations:stage_durations[stage.stage_type] stage.duration_hourselse:stage_durations[stage.stage_type] stage.duration_hours# 理想占比(成人平均值)ideal_proportions {N3: 0.20, # 深睡20%REM: 0.25, # REM睡眠25%N1: 0.05, # N1浅睡5%N2: 0.50 # N2浅睡50%}# 计算偏离度deviation_score 0for stage_type, ideal_prop in ideal_proportions.items():actual_prop stage_durations.get(stage_type, 0) / total_timedeviation_score abs(ideal_prop - actual_prop)# 睡眠效率权重efficiency_component (self.sleep_efficiency / 100) * 30# 入睡潜伏期权重latency_component max(0, (30 - self.sleep_latency) / 30) * 20# 觉醒次数权重interruption_component max(0, (5 - self.interruptions) / 5) * 10# 总分计算base_score (1 - deviation_score / 2) * 40 # 阶段占比得分final_score base_score efficiency_component latency_component interruption_componentreturn round(min(100, max(0, final_score)), 1)propertydef sleep_quality_rating(self) - SleepQuality:获取睡眠质量评级score self.sleep_quality_scoreif score 90:return SleepQuality.EXCELLENTelif score 75:return SleepQuality.GOODelif score 60:return SleepQuality.FAIRelif score 40:return SleepQuality.POORelse:return SleepQuality.VERY_POORpropertydef bedtime_hour(self) - float:获取入睡时间的小时数return self.bedtime.hour self.bedtime.minute / 60propertydef wake_time_hour(self) - float:获取起床时间的小时数return self.wake_time.hour self.wake_time.minute / 60propertydef sleep_duration_hours(self) - float:获取睡眠时长(小时)return self.total_sleep_timepropertydef deep_sleep_ratio(self) - float:深睡占比deep_sleep sum(s.duration_hours for s in self.stages if s.stage_type N3)return deep_sleep / self.total_sleep_time if self.total_sleep_time 0 else 0propertydef rem_sleep_ratio(self) - float:REM睡眠占比rem_sleep sum(s.duration_hours for s in self.stages if s.stage_type REM)return rem_sleep / self.total_sleep_time if self.total_sleep_time 0 else 0def to_dict(self) - Dict:return {session_id: self.session_id,date: self.date.strftime(%Y-%m-%d),bedtime: self.bedtime.strftime(%Y-%m-%d %H:%M:%S),wake_time: self.wake_time.strftime(%Y-%m-%d %H:%M:%S),bedtime_hour: round(self.bedtime_hour, 2),wake_time_hour: round(self.wake_time_hour, 2),sleep_duration_hours: round(self.sleep_duration_hours, 2),sleep_latency: round(self.sleep_latency, 1),interruptions: self.interruptions,sleep_efficiency: round(self.sleep_efficiency, 1),sleep_quality_score: self.sleep_quality_score,sleep_quality_rating: self.sleep_quality_rating.value,deep_sleep_ratio: round(self.deep_sleep_ratio, 3),rem_sleep_ratio: round(self.rem_sleep_ratio, 3),stages: [s.to_dict() for s in self.stages]}classmethoddef from_dict(cls, data: Dict) - SleepSession:从字典创建睡眠会话session cls(session_iddata.get(session_id, str(uuid.uuid4())[:8]),datedatetime.strptime(data[date], %Y-%m-%d),bedtimedatetime.strptime(data[bedtime], %Y-%m-%d %H:%M:%S),wake_timedatetime.strptime(data[wake_time], %Y-%m-%d %H:%M:%S),sleep_latencydata.get(sleep_latency, 15.0),interruptionsdata.get(interruptions, 0))# 重建睡眠阶段for stage_data in data.get(stages, []):stage SleepStage(stage_typestage_data[stage_type],start_timedatetime.strptime(stage_data[start_time], %Y-%m-%d %H:%M:%S),end_timedatetime.strptime(stage_data[end_time], %Y-%m-%d %H:%M:%S))session.add_stage(stage)return sessiondataclassclass WakefulnessRecord:清醒度记录属性:record_id: 记录唯一标识timestamp: 记录时间sleep_session_id: 对应的睡眠会话IDtime_since_wake: 距起床时间(小时)alertness_level: 清醒度水平(0-100)energy_level: 能量水平(0-100)mood: 情绪状态(-50到50)cognitive_performance: 认知表现(0-100)physical_performance: 身体表现(0-100)notes: 备注record_id: str field(default_factorylambda: str(uuid.uuid4())[:8])timestamp: datetime field(default_factorydatetime.now)sleep_session_id: str time_since_wake: float 0.0alertness_level: float 50.0energy_level: float 50.0mood: float 0.0cognitive_performance: float 50.0physical_performance: float 50.0notes: str def to_dict(self) - Dict:return {record_id: self.record_id,timestamp: self.timestamp.strftime(%Y-%m-%d %H:%M:%S),sleep_session_id: self.sleep_session_id,time_since_wake: round(self.time_since_wake, 2),alertness_level: round(self.alertness_level, 1),energy_level: round(self.energy_level, 1),mood: round(self.mood, 1),cognitive_performance: round(self.cognitive_performance, 1),physical_performance: round(self.physical_performance, 1),notes: self.notes}classmethoddef from_dict(cls, data: Dict) - WakefulnessRecord:return cls(**{k: (datetime.strptime(v, %Y-%m-%d %H:%M:%S) if k in [timestamp] and isinstance(v, str) else v)for k, v in data.items()})dataclassclass DailyWakefulnessData:每日清醒度数据汇总属性:date: 日期sleep_session: 当天的睡眠会话wakefulness_records: 清醒度记录列表average_alertness: 平均清醒度peak_alertness: 峰值清醒度peak_alertness_time: 峰值清醒度时间time_to_peak: 达到峰值时间sleep_inertia_duration: 睡眠惯性持续时间date: datetimesleep_session: SleepSessionwakefulness_records: List[WakefulnessRecord] field(default_factorylist)def __post_init__(self):if self.wakefulness_records:self._calculate_aggregates()def add_wakefulness_record(self, record: WakefulnessRecord) - None:添加清醒度记录self.wakefulness_records.append(record)self._calculate_aggregates()def _calculate_aggregates(self) - None:计算聚合指标if not self.wakefulness_records:return# 平均清醒度self.average_alertness sum(r.alertness_level for r in self.wakefulness_records) / \len(self.wakefulness_records)# 峰值清醒度max_record max(self.wakefulness_records, keylambda r: r.alertness_level)self.peak_alertness max_record.alertness_levelself.peak_alertness_time max_record.time_since_wake# 达到峰值时间(首次达到90%峰值的时刻)threshold self.peak_alertness * 0.9for record in sorted(self.wakefulness_records, keylambda r: r.time_since_wake):if record.alertness_level threshold:self.time_to_peak record.time_since_wakebreakelse:self.time_to_peak 2.0 # 默认2小时# 睡眠惯性持续时间(清醒度80%的持续时间)inertia_records [r for r in self.wakefulness_records if r.alertness_level 80]if inertia_records:self.sleep_inertia_duration max(r.time_since_wake for r in inertia_records)else:self.sleep_inertia_duration 0.5def to_dict(self) - Dict:return {date: self.date.strftime(%Y-%m-%d),sleep_session_id: self.sleep_session.session_id,sleep_duration: round(self.sleep_session.sleep_duration_hours, 2),wake_time: self.sleep_session.wake_time_hour,average_alertness: round(self.average_alertness, 1),peak_alertness: round(self.peak_alertness, 1),peak_alertness_time: round(self.peak_alertness_time, 2),time_to_peak: round(self.time_to_peak, 2),sleep_inertia_duration: round(self.sleep_inertia_duration, 2),records_count: len(self.wakefulness_records)}class SleepDatabase:睡眠数据库管理类智能决策要点结构化存储和管理时间序列睡眠数据def __init__(self, storage_path: str sleep_data.json):self.storage_path storage_pathself.sleep_sessions: Dict[str, SleepSession] {}self.wakefulness_records: Dict[str, WakefulnessRecord] {}self.daily_data: Dict[str, DailyWakefulnessData] {}self.load_data()def load_data(self) - None:加载历史数据try:if os.path.exists(self.storage_path):with open(self.storage_path, r, encodingutf-8) as f:data json.load(f)# 加载睡眠会话for session_data in data.get(sleep_sessions, []):session SleepSession.from_dict(session_data)self.sleep_sessions[session.session_id] session# 加载清醒度记录for record_data in data.get(wakefulness_records, []):record WakefulnessRecord.fr利用AI解决实际问题如果你觉得这个工具好用欢迎关注长安牧笛
