feat: 拆分 wecom_shown_keys 与 movement_baseline 历史层

Co-authored-by: Cursor <cursoragent@cursor.com>

daily/board_history.py

@@ -0,0 +1,83 @@
+"""企微展示历史：读写 data.wecom_shown_keys，与 movement_baseline 严格分离。"""
+
+from __future__ import annotations
+
+import json
+import logging
+from datetime import datetime, timedelta
+from pathlib import Path
+from typing import Any
+
+from daily.config import OUTPUT_DIR, board_dedup_days
+from daily.delta import RECENT_BOARD_KEYS, skill_id
+
+logger = logging.getLogger(__name__)
+
+BOARD_KEYS = RECENT_BOARD_KEYS
+
+
+def extract_shown_keys(board: str, items: list[dict[str, Any]]) -> list[str]:
+    """从最终展示 items 抽取稳定 identity key。"""
+    keys: list[str] = []
+    seen: set[str] = set()
+    for item in items:
+        if board.startswith("skills_"):
+            key = skill_id(item)
+        else:
+            key = str(item.get("repo") or "")
+        if not key or key in seen:
+            continue
+        seen.add(key)
+        keys.append(key)
+    return keys
+
+
+def load_recent_shown_keys(
+    date_str: str,
+    *,
+    lookback_days: int | None = None,
+) -> dict[str, set[str]]:
+    """近 N 日 data.wecom_shown_keys 并集（不含当日）。缺省或读失败视为空集。"""
+    empty = {board: set() for board in BOARD_KEYS}
+    try:
+        dt = datetime.strptime(date_str, "%Y-%m-%d")
+    except ValueError:
+        return empty
+
+    days = lookback_days if lookback_days is not None else board_dedup_days()
+    out: dict[str, set[str]] = {board: set() for board in BOARD_KEYS}
+
+    for day_offset in range(1, days + 1):
+        prev_date = (dt - timedelta(days=day_offset)).strftime("%Y-%m-%d")
+        path = OUTPUT_DIR / f"{prev_date}.data.json"
+        if not path.exists():
+            continue
+        try:
+            payload = json.loads(path.read_text(encoding="utf-8"))
+        except (OSError, ValueError) as exc:
+            logger.warning("读取 wecom_shown_keys %s 失败：%s", path, exc)
+            continue
+        data = payload.get("data")
+        if not isinstance(data, dict):
+            continue
+        shown = data.get("wecom_shown_keys")
+        if not isinstance(shown, dict):
+            continue
+        for board in BOARD_KEYS:
+            keys = shown.get(board) or []
+            if not isinstance(keys, list):
+                continue
+            out[board].update(str(k) for k in keys if k)
+    return out
+
+
+def merge_wecom_shown_into_data(
+    data: dict[str, Any],
+    shown: dict[str, list[str]],
+) -> dict[str, Any]:
+    """写入 wecom_shown_keys，不修改 movement_baseline。"""
+    merged = dict(data)
+    merged["wecom_shown_keys"] = {
+        board: list(keys) for board, keys in shown.items()
+    }
+    return merged

daily/config.py

@@ -30,6 +30,24 @@ def wecom_skill_desc_limit() -> int:
     return env_int("DAILY_WECOM_SKILL_DESC_LIMIT", 56)
 
 
+def wecom_news_desc_limit() -> int:
+    """企微新闻摘要建议字数；在句读/词边界截断，不加省略号。"""
+    return max(24, env_int("DAILY_WECOM_NEWS_DESC_LIMIT", 72))
+
+
+def wecom_ai_news_tech_limit() -> int:
+    """research 模式下技术类时讯条数（叠加在 DAILY_WECOM_AI_NEWS 之上）。"""
+    return max(0, env_int("DAILY_WECOM_AI_NEWS_TECH", 5))
+
+
+def wecom_pad_pool_size(display_limit: int) -> int:
+    """Delta 补榜候选池大小（展示条数之上多取，避免去重后凑不满）。"""
+    explicit = env_int("DAILY_WECOM_PAD_POOL", -1)
+    if explicit > 0:
+        return explicit
+    return max(display_limit * 5, 50)
+
+
 def full_desc_limit() -> int:
     """完整版早报摘要长度；0 表示不截断。"""
     return env_int("DAILY_FULL_DESC_LIMIT", 0)
@@ -99,5 +117,53 @@ def delta_baseline_fallback() -> str:
     return raw if raw in {"full", "empty"} else "full"
 
 
+def wecom_delta_pad() -> bool:
+    """Delta 模式下新入榜优先，不足时用当日 Top 榜补满；补榜排除近 N 天 baseline 已出现条目。"""
+    return env_bool("DAILY_WECOM_DELTA_PAD", True)
+
+
+def delta_pad_lookback_days() -> int:
+    """补榜时排除近 N 天 baseline 已出现过的条目（默认与异动对比窗口一致）。"""
+    fallback = env_int("DAILY_DELTA_LOOKBACK_DAYS", 7)
+    return max(1, env_int("DAILY_DELTA_PAD_LOOKBACK_DAYS", fallback))
+
+
 def force_push() -> bool:
     return env_bool("DAILY_FORCE_PUSH", False)
+
+
+def schedule_timezone_name() -> str:
+    return (env("DAILY_SCHEDULE_TZ") or "Asia/Shanghai").strip() or "Asia/Shanghai"
+
+
+def schedule_generate_at() -> str:
+    return (env("DAILY_SCHEDULE_GENERATE_AT") or "08:50").strip() or "08:50"
+
+
+def schedule_push_at() -> str:
+    return (env("DAILY_SCHEDULE_PUSH_AT") or "09:00").strip() or "09:00"
+
+
+def board_dedup_days() -> int:
+    return max(1, env_int("DAILY_BOARD_DEDUP_DAYS", 7))
+
+
+def board_pool_size() -> int:
+    fallback = env_int("DAILY_WECOM_SKILL_POOL", 50)
+    return max(1, env_int("DAILY_BOARD_POOL_SIZE", max(50, fallback)))
+
+
+def featured_dedup_days() -> int:
+    return max(1, env_int("DAILY_FEATURED_DEDUP_DAYS", 30))
+
+
+def theme_ban_days() -> int:
+    return max(1, env_int("DAILY_THEME_BAN_DAYS", 7))
+
+
+def narrative_axis_days() -> int:
+    return max(1, env_int("DAILY_NARRATIVE_AXIS_DAYS", 3))
+
+
+def news_backfill_enabled() -> bool:
+    return env_bool("DAILY_NEWS_BACKFILL", False)

daily/delta.py

@@ -15,6 +15,15 @@ logger = logging.getLogger(__name__)
 KeyFn = Callable[[dict[str, Any]], str]
 
 
+RECENT_BOARD_KEYS = (
+    "skills_trending",
+    "skills_hot",
+    "github_trending",
+    "github_emerging",
+    "github_topic",
+)
+
+
 def compare_depth() -> int:
     return env_int("DAILY_DELTA_COMPARE_DEPTH", 15)
 
@@ -39,6 +48,23 @@ def _key_set(items: list[dict[str, Any]], key_fn: KeyFn, *, depth: int) -> set[s
     return {key_fn(item) for item in items[:depth] if key_fn(item)}
 
 
+def load_recent_board_keys(
+    date_str: str,
+    *,
+    lookback_days: int | None = None,
+) -> dict[str, set[str]]:
+    """近 N 天各榜已展示过的 skill id / repo（不含当日，供补榜去重）。
+
+    委托 board_history.load_recent_shown_keys，只读 wecom_shown_keys，
+    不读 movement_baseline。
+    """
+    from daily.board_history import load_recent_shown_keys
+    from daily.config import board_dedup_days
+
+    days = lookback_days if lookback_days is not None else board_dedup_days()
+    return load_recent_shown_keys(date_str, lookback_days=days)
+
+
 def find_previous_data(date_str: str) -> tuple[str, dict[str, Any]] | None:
     """查找最近一份早于 date_str 的 data.json。"""
     try: