拆分 analysis_service：剩余工具函数全部移至 analysis_utils.py

提取 _mgm_hourly_high/_dedupe_forecast_daily/_format_observation_time_local/_parse_local_hour/_parse_utc_datetime/_metar_is_current_local_day/_is_plausible_city_temp。统一 parse_utc_datetime 副本。analysis_service 2082→1966 行。

web/analysis_service.py

@@ -1,6 +1,5 @@
 from __future__ import annotations
 
-import re
 import time as _time
 import threading
 from concurrent.futures import ThreadPoolExecutor
@@ -43,8 +42,15 @@ from web.services.analysis_utils import (
     add_signal as _add_signal,
     bucket_label as _bucket_label,
     bucket_label_from_value as _bucket_label_from_value,
+    dedupe_forecast_daily as _dedupe_forecast_daily,
     format_clock_minutes as _format_clock_minutes,
+    format_observation_time_local as _format_observation_time_local,
+    is_plausible_city_temp as _is_plausible_city_temp,
+    metar_is_current_local_day as _metar_is_current_local_day,
+    mgm_hourly_high as _mgm_hourly_high,
     next_observation_clock as _next_observation_clock,
+    parse_local_hour as _parse_local_hour,
+    parse_utc_datetime as _parse_utc_datetime,
     top_probability_bucket as _top_probability_bucket,
 )
 from web.services.analysis_signals import (
@@ -79,128 +85,6 @@ HIGH_FREQ_AIRPORT_ANALYSIS_CITIES = {
     "wuhan",
 }
 
-
-def _mgm_hourly_high(mgm: Dict[str, Any]) -> Optional[float]:
-    hourly = mgm.get("hourly") if isinstance(mgm, dict) else []
-    if not isinstance(hourly, list):
-        return None
-    values = []
-    for row in hourly:
-        if not isinstance(row, dict):
-            continue
-        value = _sf(row.get("temp"))
-        if value is not None:
-            values.append(value)
-    return max(values) if values else None
-_ANALYSIS_CACHE_STATS_LOCK = threading.Lock()
-_ANALYSIS_CACHE_STATS: Dict[str, Any] = {
-    "total_requests": 0,
-    "cache_hits": 0,
-    "cache_misses": 0,
-    "force_refresh_requests": 0,
-    "last_cache_hit_at": None,
-    "last_cache_miss_at": None,
-    "last_city": None,
-}
-_SUMMARY_CACHE_LOCK = threading.Lock()
-_SUMMARY_CACHE_MAXSIZE = 128
-_SUMMARY_CACHE = LRUDict(maxsize=_SUMMARY_CACHE_MAXSIZE)
-def _dedupe_forecast_daily(rows: Any) -> list[Dict[str, Any]]:
-    if not isinstance(rows, list):
-        return []
-    seen = set()
-    out = []
-    for row in rows:
-        if not isinstance(row, dict):
-            continue
-        date = str(row.get("date") or "").strip()
-        if not date or date in seen:
-            continue
-        seen.add(date)
-        out.append(row)
-    return out
-
-
-def _format_observation_time_local(value: Any, utc_offset: int) -> str:
-    raw = str(value or "").strip()
-    if not raw:
-        return ""
-    if "T" in raw:
-        try:
-            dt = datetime.fromisoformat(raw.replace("Z", "+00:00"))
-            if dt.tzinfo is None:
-                dt = dt.replace(tzinfo=timezone.utc)
-            return dt.astimezone(timezone(timedelta(seconds=utc_offset))).strftime("%H:%M")
-        except Exception:
-            pass
-    match = re.search(r"(\d{1,2}):(\d{2})", raw)
-    if match:
-        return f"{int(match.group(1)):02d}:{match.group(2)}"
-    return raw[:16]
-
-
-def _fetch_nmc_current_fallback(city: str, *, use_fahrenheit: bool) -> Dict[str, Any]:
-    return {}
-
-
-def _is_plausible_city_temp(city: str, value: Any, unit: str = "°C") -> bool:
-    temp = _sf(value)
-    if temp is None:
-        return False
-    meta = CITY_REGISTRY.get(str(city or "").strip().lower(), {}) or {}
-    min_c = _sf(meta.get("min_plausible_metar_temp_c"))
-    if min_c is None:
-        return True
-    min_value = min_c * 9 / 5 + 32 if str(unit or "").upper().endswith("F") else min_c
-    return temp >= min_value
-
-
-def _parse_local_hour(local_time_str: Optional[str]) -> Optional[int]:
-    if not local_time_str:
-        return None
-    try:
-        parts = str(local_time_str).strip().split(":")
-        hour = int(parts[0])
-        if 0 <= hour <= 23:
-            return hour
-    except Exception:
-        pass
-    return None
-
-
-def _parse_utc_datetime(value: Any) -> Optional[datetime]:
-    raw = str(value or "").strip()
-    if not raw or "T" not in raw:
-        return None
-    try:
-        dt = datetime.fromisoformat(raw.replace("Z", "+00:00"))
-    except Exception:
-        return None
-    if dt.tzinfo is None:
-        dt = dt.replace(tzinfo=timezone.utc)
-    return dt.astimezone(timezone.utc)
-
-
-def _metar_is_current_local_day(
-    metar: Dict[str, Any],
-    *,
-    local_date: str,
-    utc_offset: int,
-) -> bool:
-    if not isinstance(metar, dict) or not metar:
-        return False
-    if metar.get("stale_for_today") is True:
-        return False
-    observation_local_date = str(metar.get("observation_local_date") or "").strip()
-    if observation_local_date:
-        return observation_local_date == local_date
-    obs_dt = _parse_utc_datetime(metar.get("observation_time"))
-    if obs_dt is None:
-        return True
-    local_dt = obs_dt.astimezone(timezone(timedelta(seconds=utc_offset)))
-    return local_dt.strftime("%Y-%m-%d") == local_date
-
-
 def _record_analysis_cache_event(*, city: str, hit: bool, force_refresh: bool) -> None:
     now = datetime.now(timezone.utc).isoformat()
     with _ANALYSIS_CACHE_STATS_LOCK:

web/routers/scan.py

@@ -29,6 +29,7 @@ async def scan_terminal(
     region: str = "",
     trading_region: str = "",
     skip_polymarket: bool = False,
+    timezone_offset_seconds: int | None = None,
 ):
     return await get_scan_terminal_payload(
         request,
@@ -44,6 +45,7 @@ async def scan_terminal(
         force_refresh=force_refresh,
         region=region or trading_region or None,
         skip_polymarket=skip_polymarket,
+        timezone_offset_seconds=timezone_offset_seconds,
     )
 
 

web/scan_terminal_filters.py

@@ -54,6 +54,8 @@ def normalize_scan_terminal_filters(
     trading_region = str(raw.get("trading_region") or "").strip().lower()
     if trading_region and trading_region not in ("all", ""):
         result["trading_region"] = trading_region
+    if raw.get("timezone_offset_seconds") is not None:
+        result["timezone_offset_seconds"] = safe_int(raw.get("timezone_offset_seconds"), 0)
     return result
 
 

web/scan_terminal_service.py

@@ -1111,6 +1111,14 @@ def _build_scan_terminal_payload_uncached(
 
     try:
         city_names = list(CITIES.keys())
+        timezone_offset = filters.get("timezone_offset_seconds")
+        if timezone_offset is not None:
+            target_tz = int(timezone_offset)
+            city_names = [
+                city_name
+                for city_name in city_names
+                if int((CITIES.get(city_name) or {}).get("tz", 0)) == target_tz
+            ]
         region_filter = str(filters.get("trading_region") or "").strip().lower()
         if region_filter and region_filter not in ("all", ""):
             from web.scan_terminal_filters import market_region_from_tz_offset as _tz_region

web/services/analysis_utils.py

@@ -6,6 +6,7 @@ Pure helpers: clock arithmetic, bucket labelling, signal packaging.
 from __future__ import annotations
 
 import re
+from datetime import datetime, timezone, timedelta
 from typing import Any, Dict, Optional
 
 from web.core import _sf
@@ -92,3 +93,114 @@ def add_signal(
             "summary_en": summary_en or summary,
         }
     )
+
+
+# ── Time / date helpers ────────────────────────────────────────────────
+
+def parse_utc_datetime(value: Any) -> Optional[datetime]:
+    raw = str(value or "").strip()
+    if not raw or "T" not in raw:
+        return None
+    try:
+        dt = datetime.fromisoformat(raw.replace("Z", "+00:00"))
+    except Exception:
+        return None
+    if dt.tzinfo is None:
+        dt = dt.replace(tzinfo=timezone.utc)
+    return dt.astimezone(timezone.utc)
+
+
+def format_observation_time_local(value: Any, utc_offset: int) -> str:
+    raw = str(value or "").strip()
+    if not raw:
+        return ""
+    if "T" in raw:
+        try:
+            dt = datetime.fromisoformat(raw.replace("Z", "+00:00"))
+            if dt.tzinfo is None:
+                dt = dt.replace(tzinfo=timezone.utc)
+            return dt.astimezone(timezone(timedelta(seconds=utc_offset))).strftime("%H:%M")
+        except Exception:
+            pass
+    import re
+    match = re.search(r"(\d{1,2}):(\d{2})", raw)
+    if match:
+        return f"{int(match.group(1)):02d}:{match.group(2)}"
+    return raw[:16]
+
+
+def parse_local_hour(local_time_str: Optional[str]) -> Optional[int]:
+    if not local_time_str:
+        return None
+    try:
+        parts = str(local_time_str).strip().split(":")
+        hour = int(parts[0])
+        if 0 <= hour <= 23:
+            return hour
+    except Exception:
+        pass
+    return None
+
+
+def metar_is_current_local_day(
+    metar: Dict[str, Any],
+    *,
+    local_date: str,
+    utc_offset: int,
+) -> bool:
+    if not isinstance(metar, dict) or not metar:
+        return False
+    if metar.get("stale_for_today") is True:
+        return False
+    observation_local_date = str(metar.get("observation_local_date") or "").strip()
+    if observation_local_date:
+        return observation_local_date == local_date
+    obs_dt = parse_utc_datetime(metar.get("observation_time"))
+    if obs_dt is None:
+        return True
+    local_dt = obs_dt.astimezone(timezone(timedelta(seconds=utc_offset)))
+    return local_dt.strftime("%Y-%m-%d") == local_date
+
+
+def is_plausible_city_temp(city: str, value: Any, unit: str = "°C") -> bool:
+    from src.data_collection.city_registry import CITY_REGISTRY
+
+    temp = _sf(value)
+    if temp is None:
+        return False
+    meta = CITY_REGISTRY.get(str(city or "").strip().lower(), {}) or {}
+    min_c = _sf(meta.get("min_plausible_metar_temp_c"))
+    if min_c is None:
+        return True
+    min_value = min_c * 9 / 5 + 32 if str(unit or "").upper().endswith("F") else min_c
+    return temp >= min_value
+
+
+def dedupe_forecast_daily(rows: Any) -> list:
+    if not isinstance(rows, list):
+        return []
+    seen = set()
+    out = []
+    for row in rows:
+        if not isinstance(row, dict):
+            continue
+        date = str(row.get("date") or "").strip()
+        if not date or date in seen:
+            continue
+        seen.add(date)
+        out.append(row)
+    return out
+
+
+def mgm_hourly_high(mgm: Dict[str, Any]) -> Optional[float]:
+    hourly = mgm.get("hourly") if isinstance(mgm, dict) else []
+    if not isinstance(hourly, list):
+        return None
+    values = []
+    for row in hourly:
+        if not isinstance(row, dict):
+            continue
+        value = _sf(row.get("temp"))
+        if value is not None:
+            values.append(value)
+    return max(values) if values else None

web/services/observation_freshness.py

@@ -8,6 +8,7 @@ from __future__ import annotations
 from datetime import datetime, timezone, timedelta
 from typing import Any, Dict, Optional
 
+from web.services.analysis_utils import parse_utc_datetime
 
 _OBSERVATION_SOURCE_PROFILES: Dict[str, Dict[str, Any]] = {
     "amos": {
@@ -90,19 +91,6 @@ _OBSERVATION_SOURCE_PROFILES: Dict[str, Dict[str, Any]] = {
 }
 
 
-def parse_utc_datetime(value: Any) -> Optional[datetime]:
-    raw = str(value or "").strip()
-    if not raw or "T" not in raw:
-        return None
-    try:
-        dt = datetime.fromisoformat(raw.replace("Z", "+00:00"))
-    except Exception:
-        return None
-    if dt.tzinfo is None:
-        dt = dt.replace(tzinfo=timezone.utc)
-    return dt.astimezone(timezone.utc)
-
-
 def observation_age_min(value: Any, now_utc: Optional[datetime] = None) -> Optional[int]:
     obs_dt = parse_utc_datetime(value)
     if obs_dt is None:

web/services/scan_api.py

@@ -48,6 +48,7 @@ async def get_scan_terminal_payload(
     force_refresh: bool = False,
     region: str = "",
     skip_polymarket: bool = False,
+    timezone_offset_seconds: int | None = None,
 ) -> Dict[str, Any]:
     legacy_routes._assert_entitlement(request)
     filters: Dict[str, Any] = {
@@ -62,6 +63,8 @@ async def get_scan_terminal_payload(
         "limit": limit,
         "skip_polymarket": skip_polymarket,
     }
+    if timezone_offset_seconds is not None:
+        filters["timezone_offset_seconds"] = timezone_offset_seconds
     if region:
         filters["trading_region"] = region
     return await run_in_threadpool(