update

backend/api/utils/thirdpart_gold.py

@@ -9,7 +9,6 @@ import math
 import re
 from statistics import fmean
 import time
-from datetime import datetime, timedelta, timezone
 from urllib.parse import urljoin, urlsplit
 from uuid import uuid4
 
@@ -19,6 +18,7 @@ from redis.asyncio.retry import Retry
 from redis.backoff import NoBackoff
 
 from api.types.colg import GOLD_REGION_CODES, ThirdPartGoldCache
+from api.utils.gold_time import refresh_slot
 from config.main import config
 
 logger = logging.getLogger('dcalc.thirdpart_gold')
@@ -36,6 +36,7 @@ HEADERS = {
 # 网站公开 DNF 页面已停用/移除的渠道；不请求 VIP 专属渠道。
 DISABLED_PLATFORMS = {1, 12, 21, 50, 51, 52, 53}
 PAGE_SIZE = 200
+SAMPLE_SLICE = slice(3, 10)
 FRESH_SECONDS = 30 * 60
 RETENTION_SECONDS = 24 * 60 * 60
 # 按每分钟最多8次页面访问处理，理论间隔7.5秒，取8秒留出边界余量。
@@ -44,7 +45,8 @@ REGION_INTERVAL = 8
 LOCK_SECONDS = 15 * 60
 ERROR_COOLDOWN_SECONDS = 5 * 60
 REDIS_TIMEOUT_SECONDS = 1
-CACHE_PREFIX = 'openapi:colg:thirdpart_gold:v1'
+# 统计范围和单位已改变，旧缓存的万金币/元均值不能作为新的元/万金币单价返回。
+CACHE_PREFIX = 'openapi:colg:thirdpart_gold:v2'
 LOCK_KEY = f'{CACHE_PREFIX}:lock'
 COOLDOWN_KEY = f'{CACHE_PREFIX}:cooldown'
 RELEASE_LOCK = """
@@ -126,7 +128,7 @@ def make_payload(channel: dict, seed: dict, user: dict) -> dict:
 
 
 def first_page_average(offers: list[dict]) -> tuple[float, int]:
-    """等权平均每条报价的 1/Price（万金币/元），不是 1/平均单价。"""
+    """匹配公开页面 i.item1 的第4～10条，平均显示单价（元/万金币）。"""
     valid = []
     for offer in offers:
         try:
@@ -135,9 +137,12 @@ def first_page_average(offers: list[dict]) -> tuple[float, int]:
             number = float(offer['Number'])
         except (KeyError, TypeError, ValueError, OverflowError):
             continue
-        if not all(math.isfinite(value) and value > 0 for value in (price, money, number)):
+        if not all(math.isfinite(value) for value in (price, money, number)) or price < 0:
             continue
-        if not math.isfinite(1 / price):
+        # 页面默认展示条件，不额外设置price1/price2价格过滤。
+        if money != -1 and (money * number < 5 or price == 0):
+            continue
+        if money >= 1_000_000 or (offer.get('Seller') or '').startswith('智莹科技'):
             continue
         valid.append({**offer, 'Price': price})
     valid.sort(key=lambda item: (item['Price'], item['PfId'], item.get('BuyUrl') or ''))
@@ -152,9 +157,12 @@ def first_page_average(offers: list[dict]) -> tuple[float, int]:
         page.append(offer)
         if len(page) == PAGE_SIZE:
             break
-    if not page:
+    selected = page[SAMPLE_SLICE]
+    if not selected:
         raise ValueError('YXDR 未返回有效报价')
-    return round(fmean(1 / item['Price'] for item in page), 4), len(page)
+    # 页面 Math.floorDigits(price, 5) 先toFixed(6)再删掉末位。
+    prices = [float(f"{item['Price']:.6f}"[:-1]) for item in selected]
+    return round(fmean(prices), 2), len(prices)
 
 
 async def _read_response(resp) -> str:
@@ -165,7 +173,9 @@ async def _read_response(resp) -> str:
     return text
 
 
-async def fetch_region(session: aiohttp.ClientSession, region: str) -> dict:
+async def fetch_region(session: aiohttp.ClientSession, region: str, *, started_at: float | None = None) -> dict:
+    if started_at is None:
+        started_at = time.time()
     page_url = f'{BASE_URL}/bijiaqi/dnf/youxibi/{region}'
     async with session.get(page_url) as resp:
         page = await _read_response(resp)
@@ -210,10 +220,9 @@ async def fetch_region(session: aiohttp.ClientSession, region: str) -> dict:
         if isinstance(result, BaseException):
             raise result
     price, count = first_page_average([offer for result in results for offer in result])
-    now = datetime.now(timezone(timedelta(hours=8)))
     return {'region': REGIONS[region], 'code': GOLD_REGION_CODES[REGIONS[region]],
             'goldPrice': price, 'sampleCount': count,
-            'updataTime': int(now.timestamp()), 'fetchedAt': now.timestamp()}
+            'updataTime': refresh_slot(started_at), 'fetchedAt': started_at}
 
 
 async def read_cache(redis) -> list[dict]:
@@ -229,8 +238,9 @@ async def read_cache(redis) -> list[dict]:
                 continue
             # 兼容此前缓存的格式化日期及缺少code的记录，无须重新抓取所有跨区。
             result.append(ThirdPartGoldCache.model_validate({
-                **item, 'updataTime': int(item['fetchedAt']),
-                'code': GOLD_REGION_CODES[REGIONS[region]], 'stale': age >= FRESH_SECONDS,
+                **item, 'updataTime': refresh_slot(item['fetchedAt']),
+                'code': GOLD_REGION_CODES[REGIONS[region]],
+                'stale': refresh_slot(item['fetchedAt']) < refresh_slot(),
             }).model_dump())
         except (KeyError, TypeError, ValueError):
             logger.warning('invalid thirdpart gold cache region=%s', region)
@@ -270,6 +280,7 @@ async def refresh_cache(token: str):
             fresh_regions = {item['region'] for item in cached if not item['stale']}
             pending = [region for region, name in REGIONS.items() if name not in fresh_regions]
             successful, last_start = 0, None
+            started_at = time.time()
             timeout = aiohttp.ClientTimeout(total=15)
             async with aiohttp.ClientSession(headers=HEADERS, timeout=timeout,
                                              connector=aiohttp.TCPConnector(limit=5)) as session:
@@ -281,7 +292,7 @@ async def refresh_cache(token: str):
                         return
                     last_start = time.monotonic()
                     try:
-                        item = await fetch_region(session, region)
+                        item = await fetch_region(session, region, started_at=started_at)
                         if await redis.get(LOCK_KEY) not in (token, token.encode()):
                             return
                         await redis.set(f'{CACHE_PREFIX}:{region}', json.dumps(item),