新增机场高频推送:10分钟级温度监控 + DEB预测快报

为首尔/釜山/东京三大机场城市实现高频温度监控通道:
- 新增 airport_obs_log 表积累观测数据,支持趋势检测
- AMOS/JMA 成功后自动写入观测日志
- 新增 airport_rapid_temp_change 告警规则(20min窗口、0.5°C/10min阈值)
- 10分钟间隔高频子循环 + 30分钟机场快照(含 DEB 预测最高温)
- 最高温锁定后自动跳过,快照仅当地 08:00-20:00 发送

Tested: ruff check 通过
@
This commit is contained in:
2569718930@qq.com
2026-05-12 17:04:17 +08:00
parent ff4c8b0139
commit 64f8ff21ec
6 changed files with 496 additions and 25 deletions
+30 -25
View File
@@ -1,4 +1,4 @@
# 机场高频数据接入 Telegram 市场监控频道方案
# 机场高频数据接入市场监控频道方案
## 背景
@@ -15,12 +15,14 @@
- **循环**: `start_trade_alert_push_loop`,默认每 30 分钟跑一轮
- **覆盖城市**: `TELEGRAM_ALERT_CITIES`(默认全部 51 城)
- **4 条规则**: Ankara Center DEB 命中、动量突变(30min 斜率 > 0.8°C)、预报突破、暖平流
- **门禁**: 严重度/触发数/市场可交易性/冷却期 多层过滤
- **消息**: 中英双语,包含触发类型、实况温度、市场概率分布、AI 建议
- **门禁**: 严重度/触发数/冷却期 多层过滤
- **消息**: 中英双语,包含触发类型、实况温度
### 问题
高频机场数据已就绪,但现有推送系统没有针对性利用。首尔/釜山 1 分钟级 AMOS 和东京 10 分钟级 JMA 可以做更快、更灵敏的市场信号检测
高频机场数据已就绪,但现有推送系统 30 分钟一轮对所有城市一视同仁。首尔/釜山 1 分钟级 AMOS 和东京 10 分钟级 JMA 接近交易高峰期时,温度变化可能比 30 分钟窗口更快,需要更灵敏的监控
同时,DEB 每日最高温预测是结算的核心参照,将机场实时温度与 DEB 预测并排对比,可以快速发现实际温度偏离预测的程度。
---
@@ -28,24 +30,26 @@
### 核心思路
**增强现有系统,不新建**。对高频机场城市使用更低的推送间隔和更灵敏的规则参数
**在现有 30 分钟主循环之上叠加高频通道**,对三大机场城市用更短间隔推送"实时温度 + DEB 预测最高温"快报。不做市场分析、不输出 AI 建议,只报温度数据本身
### 1. 高频机场城市快速通道
在现有 30 分钟主循环之外,为 `{seoul, busan, tokyo}` 单独跑一个 10 分钟间隔的子循环,检查动量突变和预报突破两条最关键的规则,使用更低的触发阈值
在现有 30 分钟主循环之外,为 `{seoul, busan, tokyo}` 单独跑一个 10 分钟间隔的子循环,检查温度动量突变。
**配置(写死在代码中)**:
```python
HIGH_FREQ_AIRPORT_CITIES = {"seoul", "busan", "tokyo"}
HIGH_FREQ_PUSH_INTERVAL_SEC = 600 # 10 分钟
HIGH_FREQ_MOMENTUM_THRESHOLD_C = 0.5 # 比默认 0.8°C 更灵敏
HIGH_FREQ_COOLDOWN_SEC = 7200 # 同一城市冷却 2 小时
```
**逻辑**:
- 主循环 30 分钟照常跑全部城市(不变)
- 每 10 分钟额外跑一轮高频城市子集
- 高频轮次跳过 Ankara DEB 和 Advection 规则
- 高频告警独立冷却期(如 2 小时,比默认 6 小时更短)
- 高频轮次仅检查温度动量突变一条规则
- 高频告警独立冷却期 2 小时
- **最高温已锁定则跳过**:当日最高温已出现且之后持续下降,不再推送
### 2. 机场观测积累与趋势检测
@@ -79,39 +83,40 @@ CREATE INDEX IF NOT EXISTS idx_airport_obs_log_icao_time
|------|-----|------|
| 滑动窗口 | 20 分钟 | 取最近 20 分钟内的观测 |
| 最少样本 | 3 条 | 确保有足够数据点 |
| 触发阈值 | > 0.3°C/10min | 约 1.8°C/小时 |
| 额外条件 | 温度变化方向与市场桶一致 | 避免无关告警 |
| 触发阈值 | > 0.5°C/10min | 高频通道专用,比默认 0.8°C 更灵敏 |
| 锁定跳过 | 最高温已锁定 | 当日最高已过且持续下降,不推送 |
**告警消息示例**:
```
🚨 首尔/仁川 温度急变
跑道中位数 15.2→16.8°C (+1.6°C / 15min)
180°→210° 暖平流增强
市场桶 17°C 概率 62% → 当前边缘 +8.2%
跑道中位数 14.9°C → 16.5°C (+1.6°C / 15min)
220° 14kt 暖平流增强
DEB 预测今日最高 18.2°C | 目前距预测差 1.7°C
```
### 4. 机场实况快照摘要
### 4. 机场实况快照摘要(含 DEB 预测)
每隔 30 分钟,向市场监控频道推送三座机场的当前实况摘要。
每隔 30 分钟,向市场监控频道推送三座机场的"当前温度 + DEB 预测最高温"摘要。
**推送时段**: 仅各城市当地 08:00-20:00(避免半夜噪音)
**锁定跳过**: 如该城市当日最高温已锁定,跳过该城市不推送(三城中仍有未锁定的则正常推送其余)
**消息格式**:
```
🛫 机场实况快照 14:30 CST
🛫 机场实况 14:30 CST
🇰🇷 首尔/仁川 RKSI
跑道 15L/33R: 14.6°C | 15R/33L: 15.2°C
风 220° 14kt | QNH 1015.2 hPa | 能见度 ≥10km
METAR 14:00Z | 跑道中位数 14.9°C
跑道中位数 14.9°C · DEB 预测最高 18.2°C
风 220° 14kt | QNH 1015.2 hPa | 能见度 ≥10km
🇰🇷 釜山/金海 RKPK
跑道 18L/36R: 13.8°C | 风 180° 8kt
METAR 14:00Z
跑道中位数 13.8°C · DEB 预测最高 16.5°C
风 180° 8kt
🇯🇵 东京/羽田 RJTT
JMA AMeDAS 10-min: 12.4°C (14:20 JST)
当前 12.4°C (14:20 JST) · DEB 预测最高 15.1°C
```
---
@@ -124,7 +129,7 @@ JMA AMeDAS 10-min: 12.4°C (14:20 JST)
| 2 | `src/data_collection/amos_station_sources.py` | 成功获取后调用 `append_airport_obs()` 写日志 |
| 3 | `src/data_collection/jma_amedas_sources.py` | 成功获取后调用 `append_airport_obs()` 写日志,可选扩展提取更多字段(风速/气压) |
| 4 | `src/analysis/market_alert_engine.py` | 新增 `airport_rapid_temp_change` 规则 |
| 5 | `src/utils/telegram_push.py` | 高频快速通道子循环、机场快照推送、温度急变告警集成 |
| 5 | `src/utils/telegram_push.py` | 高频快速通道子循环、机场快照推送(含 DEB 预测)、温度急变告警集成 |
---
@@ -133,8 +138,8 @@ JMA AMeDAS 10-min: 12.4°C (14:20 JST)
1. **Phase 1 — DB 层**: `airport_obs_log` 表 + 读写方法
2. **Phase 2 — 采集层**: AMOS/JMA 成功后自动写日志,部署观察 1-2 天确认数据积累正常
3. **Phase 3 — 告警引擎**: `airport_rapid_temp_change` 规则 + 单元测试
4. **Phase 4 — 推送层**: 高频快速通道 + 机场快照,先在测试频道验证消息格式
5. **Phase 5 — 上线**: 切换到正式市场监控频道,观察 3-7 天调整阈值
4. **Phase 4 — 推送层**: 高频快速通道 + 机场快照 + DEB 预测,直接推送市场监控频道
5. **Phase 5 — 调参**: 观察 3-7 天调整阈值
---
+79
View File
@@ -10,6 +10,7 @@ from datetime import datetime, timezone
from typing import Any, Dict, List, Optional, Tuple
from src.analysis.settlement_rounding import apply_city_settlement
from src.database.db_manager import DBManager
def _sf(v: Any) -> Optional[float]:
@@ -812,6 +813,83 @@ def _build_advice_cn(
return "".join(parts) + ""
_AIRPORT_ICAO_MAP = {"seoul": "RKSI", "busan": "RKPK", "tokyo": "RJTT"}
def _calc_airport_rapid_temp_change(
city_weather: Dict[str, Any], temp_symbol: str
) -> Dict[str, Any]:
city = (city_weather.get("name") or "").lower()
if city not in _AIRPORT_ICAO_MAP:
return {
"type": "airport_rapid_temp_change",
"triggered": False,
"reason": "not_airport_city",
}
icao = _AIRPORT_ICAO_MAP[city]
try:
db = DBManager()
obs = db.get_airport_obs_recent(icao, minutes=20)
except Exception:
return {
"type": "airport_rapid_temp_change",
"triggered": False,
"reason": "db_read_error",
}
if len(obs) < 3:
return {
"type": "airport_rapid_temp_change",
"triggered": False,
"reason": f"insufficient_obs ({len(obs)}<3)",
}
first = obs[0]
last = obs[-1]
first_temp = first.get("temp_c")
last_temp = last.get("temp_c")
if first_temp is None or last_temp is None:
return {
"type": "airport_rapid_temp_change",
"triggered": False,
"reason": "missing_temp",
}
try:
t1 = datetime.fromisoformat(str(first.get("created_at") or ""))
t2 = datetime.fromisoformat(str(last.get("created_at") or ""))
delta_min = (t2 - t1).total_seconds() / 60.0
except Exception:
delta_min = len(obs) * 1.5
if delta_min <= 0:
return {
"type": "airport_rapid_temp_change",
"triggered": False,
"reason": "zero_time_delta",
}
delta_temp = last_temp - first_temp
slope_per_10min = delta_temp / delta_min * 10.0
threshold = _to_unit_delta(0.5, temp_symbol)
triggered = abs(slope_per_10min) > threshold
return {
"type": "airport_rapid_temp_change",
"triggered": triggered,
"direction": "up" if slope_per_10min > 0 else "down",
"first_temp": round(first_temp, 2),
"last_temp": round(last_temp, 2),
"delta_temp": round(delta_temp, 2),
"delta_min": round(delta_min, 1),
"slope_per_10min": round(slope_per_10min, 2),
"threshold": round(threshold, 2),
"sample_count": len(obs),
"icao": icao,
}
def _build_telegram_messages(
city_weather: Dict[str, Any],
rules: Dict[str, Dict[str, Any]],
@@ -1179,6 +1257,7 @@ def build_trading_alerts(
"momentum_spike": _calc_momentum_alert(city_weather, temp_symbol),
"forecast_breakthrough": _calc_forecast_breakthrough_alert(city_weather, temp_symbol),
"advection": _calc_advection_alert(city_weather, temp_symbol),
"airport_rapid_temp_change": _calc_airport_rapid_temp_change(city_weather, temp_symbol),
}
triggered = [
+25
View File
@@ -84,6 +84,7 @@ class StartupCoordinator:
def start_all(self) -> RuntimeStatus:
loops = [
self._start_trade_alert_loop(),
self._start_airport_high_freq_loop(),
self._start_dashboard_prewarm_loop(),
self._start_polygon_wallet_loop(),
self._start_polymarket_wallet_activity_loop(),
@@ -182,6 +183,30 @@ class StartupCoordinator:
),
)
def _start_airport_high_freq_loop(self) -> LoopStatus:
enabled = _env_bool("TELEGRAM_AIRPORT_PUSH_ENABLED", True)
chat_ids = get_telegram_chat_ids_from_env()
interval = max(60, _env_int("TELEGRAM_AIRPORT_PUSH_INTERVAL_SEC", 600))
details = {
"mode": "airport-high-freq",
"interval_sec": interval,
"cities": ["seoul", "busan", "tokyo"],
"chat_targets": len(chat_ids),
"cooldown_sec": _env_int("TELEGRAM_AIRPORT_COOLDOWN_SEC", 7200),
}
validation_error = None if chat_ids else "missing_TELEGRAM_CHAT_IDS"
return self._start_with_validation(
key="airport_high_freq_push",
label="机场高频推送",
configured_enabled=enabled,
details=details,
validation_error=validation_error,
starter=lambda: import_module("src.utils.telegram_push").start_high_freq_airport_push_loop(
self.bot,
self.config,
),
)
def _start_dashboard_prewarm_loop(self) -> LoopStatus:
enabled = _env_bool("POLYWEATHER_DASHBOARD_PREWARM_ENABLED", False)
interval = max(30, _env_int("POLYWEATHER_PREWARM_INTERVAL_SEC", 300))
+23
View File
@@ -15,6 +15,7 @@ from src.data_collection.russia_station_sources import RussiaStationSourceMixin
from src.data_collection.nmc_sources import NmcSourceMixin
from src.data_collection.nws_open_meteo_sources import NwsOpenMeteoSourceMixin
from src.data_collection.amos_station_sources import AmosStationSourceMixin
from src.database.db_manager import DBManager
class WeatherDataCollector(OpenMeteoCacheMixin, SettlementSourceMixin, MetarSourceMixin, MgmSourceMixin, JmaAmedasSourceMixin, RussiaStationSourceMixin, NmcSourceMixin, NwsOpenMeteoSourceMixin, AmosStationSourceMixin):
@@ -865,6 +866,17 @@ class WeatherDataCollector(OpenMeteoCacheMixin, SettlementSourceMixin, MetarSour
if "mgm_nearby" not in results:
results["mgm_nearby"] = official_rows
results["nearby_source"] = "jma"
try:
row = official_rows[0] if official_rows else {}
icao = str(row.get("icao") or row.get("istNo") or "RJTT")
DBManager().append_airport_obs(
icao=icao,
city=city_lower,
temp_c=row.get("temp"),
obs_time=str(row.get("obs_time") or datetime.now().isoformat()),
)
except Exception:
logger.exception("airport_obs_log append failed for jma city={}", city_lower)
def _attach_korean_amos_data(
self, results: Dict, city_lower: str, use_fahrenheit: bool
@@ -882,6 +894,17 @@ class WeatherDataCollector(OpenMeteoCacheMixin, SettlementSourceMixin, MetarSour
city_lower, amos_data.get("temp_c"), amos_data.get("temp_source"),
len(amos_data.get("runway_obs", {}).get("runway_pairs", []) or []))
results["amos"] = amos_data
try:
DBManager().append_airport_obs(
icao=amos_data.get("icao") or "",
city=city_lower,
temp_c=amos_data.get("temp_c"),
wind_kt=amos_data.get("wind_kt"),
pressure_hpa=amos_data.get("pressure_hpa"),
obs_time=amos_data.get("observation_time") or datetime.now().isoformat(),
)
except Exception:
logger.exception("airport_obs_log append failed for amos city={}", city_lower)
else:
logger.warning("AMOS: no data returned for city={}", city_lower)
except Exception as exc:
+55
View File
@@ -280,6 +280,21 @@ class DBManager:
conn.execute(
"CREATE INDEX IF NOT EXISTS idx_supabase_bindings_telegram_id ON supabase_bindings(telegram_id)"
)
conn.execute("""
CREATE TABLE IF NOT EXISTS airport_obs_log (
id INTEGER PRIMARY KEY AUTOINCREMENT,
icao TEXT NOT NULL,
city TEXT NOT NULL,
temp_c REAL,
wind_kt REAL,
pressure_hpa REAL,
obs_time TEXT NOT NULL,
created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
)
""")
conn.execute(
"CREATE INDEX IF NOT EXISTS idx_airport_obs_log_icao_time ON airport_obs_log(icao, created_at DESC)"
)
self._ensure_column(conn, "users", "daily_points", "INTEGER DEFAULT 0")
self._ensure_column(conn, "users", "daily_points_date", "TEXT")
self._ensure_column(conn, "users", "weekly_points", "INTEGER DEFAULT 0")
@@ -1825,3 +1840,43 @@ class DBManager:
)
conn.commit()
return True
def append_airport_obs(
self,
*,
icao: str,
city: str,
temp_c: Optional[float] = None,
wind_kt: Optional[float] = None,
pressure_hpa: Optional[float] = None,
obs_time: str,
) -> None:
with self._get_connection() as conn:
conn.execute(
"""
INSERT INTO airport_obs_log (icao, city, temp_c, wind_kt, pressure_hpa, obs_time)
VALUES (?, ?, ?, ?, ?, ?)
""",
(str(icao).strip().upper(), str(city).strip().lower(),
temp_c, wind_kt, pressure_hpa, str(obs_time)),
)
conn.execute(
"DELETE FROM airport_obs_log WHERE created_at < datetime('now', '-2 hours')"
)
conn.commit()
def get_airport_obs_recent(
self, icao: str, minutes: int = 30
) -> list[dict[str, Any]]:
with self._get_connection() as conn:
conn.row_factory = sqlite3.Row
rows = conn.execute(
"""
SELECT icao, city, temp_c, wind_kt, pressure_hpa, obs_time, created_at
FROM airport_obs_log
WHERE icao = ? AND created_at >= datetime('now', ? || ' minutes')
ORDER BY created_at ASC
""",
(str(icao).strip().upper(), str(-int(minutes))),
).fetchall()
return [dict(r) for r in rows]
+284
View File
@@ -686,3 +686,287 @@ def start_trade_alert_push_loop(bot: Any, config: Dict[str, Any]) -> Optional[th
)
thread.start()
return thread
# ── high-freq airport push loop ──
HIGH_FREQ_AIRPORT_CITIES = {"seoul", "busan", "tokyo"}
HIGH_FREQ_AIRPORT_ICAO = {"seoul": "RKSI", "busan": "RKPK", "tokyo": "RJTT"}
HIGH_FREQ_PUSH_INTERVAL_SEC = 600
HIGH_FREQ_MOMENTUM_THRESHOLD_C = 0.5
HIGH_FREQ_COOLDOWN_SEC = 7200
AIRPORT_SNAPSHOT_INTERVAL_SEC = 1800
_AIRPORT_PUSH_STATE_PATH = os.path.join(
os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))),
"data", "airport_push_state.json",
)
def _load_airport_state() -> Dict[str, Any]:
try:
from src.database.runtime_state import STATE_STORAGE_SQLITE, get_state_storage_mode
if get_state_storage_mode() == STATE_STORAGE_SQLITE:
return _telegram_state_repo.load_state()
except Exception:
pass
path = _AIRPORT_PUSH_STATE_PATH
if not os.path.exists(path):
return {"last_by_city": {}}
try:
with open(path, "r", encoding="utf-8") as fh:
data = json.load(fh)
if isinstance(data, dict):
data.setdefault("last_by_city", {})
return data
except Exception:
pass
return {"last_by_city": {}}
def _save_airport_state(state: Dict[str, Any]) -> None:
try:
from src.database.runtime_state import STATE_STORAGE_SQLITE, get_state_storage_mode
if get_state_storage_mode() == STATE_STORAGE_SQLITE:
_telegram_state_repo.save_state(state)
return
except Exception:
pass
path = _AIRPORT_PUSH_STATE_PATH
os.makedirs(os.path.dirname(path), exist_ok=True)
tmp = f"{path}.tmp"
with open(tmp, "w", encoding="utf-8") as fh:
json.dump(state, fh, ensure_ascii=False, indent=2)
os.replace(tmp, path)
def _check_high_locked(city: str) -> bool:
icao = HIGH_FREQ_AIRPORT_ICAO.get(city)
if not icao:
return False
try:
from src.database.db_manager import DBManager
db = DBManager()
obs = db.get_airport_obs_recent(icao, minutes=120)
except Exception:
return False
if len(obs) < 6:
return False
temps = [r.get("temp_c") for r in obs if r.get("temp_c") is not None]
if len(temps) < 6:
return False
peak_idx = temps.index(max(temps))
if peak_idx >= len(temps) - 2:
return False
post_peak = temps[peak_idx:]
for i in range(1, len(post_peak)):
if post_peak[i] > post_peak[i - 1] + 0.05:
return False
return True
def _build_airport_rapid_change_message(
city: str,
rule: Dict[str, Any],
deb_pred: Optional[float],
) -> str:
city_display = city.title()
airport_label = {"seoul": "首尔/仁川", "busan": "釜山/金海", "tokyo": "东京/羽田"}.get(city, city_display)
emoji = "🔥" if rule.get("direction") == "up" else "❄️"
first = rule.get("first_temp", 0)
last = rule.get("last_temp", 0)
delta = rule.get("delta_temp", 0)
delta_min = rule.get("delta_min", 0)
sign = "+" if delta >= 0 else ""
lines = [
f"🚨 {airport_label} 温度急变 {emoji}",
"",
f"跑道温度 {first:.1f}°C → {last:.1f}°C ({sign}{delta:.1f}°C / {delta_min:.0f}min)",
]
wind_kt = rule.get("wind_kt")
if wind_kt is not None:
lines.append(f"{wind_kt:.0f}kt")
if deb_pred is not None:
gap = last - deb_pred
gap_sign = "+" if gap >= 0 else ""
lines.append(f"DEB 预测今日最高 {deb_pred:.1f}°C | 差距 {gap_sign}{gap:.1f}°C")
return "\n".join(lines)
def _build_airport_snapshot_message(snapshots: list[dict[str, Any]], local_time: str = "") -> str:
flag = {"seoul": "🇰🇷", "busan": "🇰🇷", "tokyo": "🇯🇵"}
name = {"seoul": "首尔/仁川 RKSI", "busan": "釜山/金海 RKPK", "tokyo": "东京/羽田 RJTT"}
lines = [f"🛫 机场实况 {local_time}"]
for s in snapshots:
city = s.get("city", "")
temp = s.get("temp_c")
deb = s.get("deb_prediction")
lines.append("")
header = f"{flag.get(city, '')} {name.get(city, city)}"
parts = []
if temp is not None:
parts.append(f"当前 {temp:.1f}°C")
if deb is not None:
parts.append(f"DEB 预测最高 {deb:.1f}°C")
line = header
if parts:
line += "\n" + " · ".join(parts)
if s.get("wind_kt") is not None:
line += f" | 风 {s['wind_kt']:.0f}kt"
if s.get("pressure_hpa") is not None:
line += f" | QNH {s['pressure_hpa']:.1f} hPa"
lines.append(line)
return "\n".join(lines)
def _run_high_freq_airport_cycle(
bot: Any,
config: Dict[str, Any],
chat_ids: List[str],
state: Dict[str, Any],
) -> bool:
state_dirty = False
now_ts = int(time.time())
last_by_city = state.setdefault("last_by_city", {})
snapshots: list[dict[str, Any]] = []
snapshot_due = now_ts - int(state.get("last_snapshot_ts") or 0) >= AIRPORT_SNAPSHOT_INTERVAL_SEC
for city in sorted(HIGH_FREQ_AIRPORT_CITIES):
try:
icao = HIGH_FREQ_AIRPORT_ICAO.get(city, "")
last_city = last_by_city.get(city) or {}
last_city_ts = int(last_city.get("ts") or 0)
try:
from src.database.db_manager import DBManager
db = DBManager()
recent_obs = db.get_airport_obs_recent(icao, minutes=20)
latest_obs = recent_obs[-1] if recent_obs else {}
except Exception:
latest_obs = {}
# Fetch city_weather once and extract DEB
city_weather: Dict[str, Any] = {}
deb_pred: Optional[float] = None
try:
from web.app import _analyze
city_weather = _analyze(city)
deb_raw = (city_weather.get("deb") or {}).get("prediction")
if deb_raw is not None:
deb_pred = float(deb_raw)
except Exception:
pass
# Check high-locked
if _check_high_locked(city):
if snapshot_due:
snapshots.append({
"city": city,
"temp_c": latest_obs.get("temp_c"),
"wind_kt": latest_obs.get("wind_kt"),
"pressure_hpa": latest_obs.get("pressure_hpa"),
"deb_prediction": deb_pred,
})
continue
from src.analysis.market_alert_engine import _calc_airport_rapid_temp_change
rule = _calc_airport_rapid_temp_change(city_weather, "°C")
if snapshot_due:
snapshots.append({
"city": city,
"temp_c": latest_obs.get("temp_c"),
"wind_kt": latest_obs.get("wind_kt"),
"pressure_hpa": latest_obs.get("pressure_hpa"),
"deb_prediction": deb_pred,
})
if not rule.get("triggered"):
continue
if last_city_ts and now_ts - last_city_ts < HIGH_FREQ_COOLDOWN_SEC:
continue
rule["wind_kt"] = latest_obs.get("wind_kt")
message = _build_airport_rapid_change_message(city, rule, deb_pred)
sent = False
for chat_id in chat_ids:
try:
bot.send_message(chat_id, message)
sent = True
except Exception as exc:
logger.warning("airport push failed city={} chat_id={}: {}", city, chat_id, exc)
if sent:
last_by_city[city] = {"ts": now_ts, "trigger": "rapid_temp_change"}
state_dirty = True
logger.info("airport rapid change pushed city={} delta_temp={:.1f}", city, rule.get("delta_temp", 0))
except Exception:
logger.exception("high freq airport cycle failed for city={}", city)
if snapshot_due and snapshots:
from datetime import timedelta
kst_hour = (datetime.utcnow() + timedelta(hours=9)).hour
if 8 <= kst_hour < 20:
local_time = datetime.now().strftime("%H:%M CST")
snap_message = _build_airport_snapshot_message(snapshots, local_time)
for chat_id in chat_ids:
try:
bot.send_message(chat_id, snap_message)
except Exception as exc:
logger.warning("airport snapshot push failed chat_id={}: {}", chat_id, exc)
logger.info("airport snapshot pushed cities={}", len(snapshots))
else:
logger.info("airport snapshot skipped: outside 08:00-20:00 window (KST hour={})", kst_hour)
state["last_snapshot_ts"] = now_ts
state_dirty = True
return state_dirty
def start_high_freq_airport_push_loop(bot: Any, config: Dict[str, Any]) -> Optional[threading.Thread]:
enabled = _env_bool("TELEGRAM_AIRPORT_PUSH_ENABLED", True)
chat_ids = get_telegram_chat_ids_from_env()
if not enabled:
logger.info("airport high-freq push loop disabled")
return None
if not chat_ids:
logger.warning("airport high-freq push loop skipped: TELEGRAM_CHAT_IDS is not set")
return None
interval_sec = max(60, _env_int("TELEGRAM_AIRPORT_PUSH_INTERVAL_SEC", HIGH_FREQ_PUSH_INTERVAL_SEC))
def _runner() -> None:
state = _load_airport_state()
logger.info(
"airport high-freq push loop started cities={} interval={}s chat_targets={}",
len(HIGH_FREQ_AIRPORT_CITIES), interval_sec, len(chat_ids),
)
while True:
cycle_started = time.time()
state = _load_airport_state()
if _run_high_freq_airport_cycle(
bot=bot,
config=config,
chat_ids=chat_ids,
state=state,
):
_save_airport_state(state)
elapsed = time.time() - cycle_started
sleep_sec = max(5, interval_sec - int(elapsed))
time.sleep(sleep_sec)
thread = threading.Thread(
target=_runner,
name="airport-high-freq-pusher",
daemon=True,
)
thread.start()
logger.info("airport high-freq push loop thread started")
return thread