stock-analyzer-skill 1.1.0

package/scripts/data/__init__.py

@@ -0,0 +1,208 @@
+"""
+统一数据获取 API。
+
+用法:
+    from data import get_quote, get_kline, get_finance
+
+    quote = get_quote("sh600989")
+    quotes = get_quotes(["sh600989", "sz000807"])
+    bars = get_kline("sh600989", scale=240, datalen=30)
+    records = get_finance("SH600989")
+"""
+import threading
+from typing import Optional
+
+from .types import Quote, KlineBar, FinanceRecord
+from .config import get_config
+from . import cache
+import sys
+from pathlib import Path
+sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
+
+
+def _get_common_helpers():
+    """延迟导入 common，避免 common.py ↔ data/__init__.py 循环导入。"""
+    from common import to_float, to_int
+    return to_float, to_int
+
+# 延迟导入 fetchers（避免循环导入），线程安全
+_fetchers_lock = threading.Lock()
+_fetchers_loaded = False
+_quote_manager = None
+_kline_manager = None
+_finance_manager = None
+
+
+def _load_fetchers():
+    global _fetchers_loaded, _quote_manager, _kline_manager, _finance_manager
+    if _fetchers_loaded:
+        return
+    with _fetchers_lock:
+        if _fetchers_loaded:
+            return
+        from fetchers import get_quote_manager, get_kline_manager, get_finance_manager
+        _quote_manager = get_quote_manager()
+        _kline_manager = get_kline_manager()
+        _finance_manager = get_finance_manager()
+        _fetchers_loaded = True
+
+
+def get_quote(code: str, use_cache: bool = True) -> Optional[Quote]:
+    """获取单只股票行情。"""
+    _load_fetchers()
+    cfg = get_config()
+    key = f"quote_{code}"
+
+    if use_cache:
+        cached = cache.get_json(key, cfg.quote_cache_ttl)
+        if cached:
+            return _dict_to_quote(cached)
+
+    result = _quote_manager.fetch(code)
+    if result is None:
+        return None
+
+    quote = _dict_to_quote(result)
+
+    if use_cache and quote.has_basic_data():
+        cache.set_json(key, quote.to_dict())
+
+    return quote
+
+
+def get_quotes(codes: list, use_cache: bool = True) -> list:
+    """批量获取行情。"""
+    from common import parallel_map
+    cfg = get_config()
+    results = parallel_map(lambda c: get_quote(c, use_cache), codes,
+                           max_workers=cfg.max_workers, timeout=30)
+    return [q for q in results.values() if q is not None]
+
+
+def get_kline(code: str, scale: int = 240, datalen: int = 30,
+              use_cache: bool = True) -> list:
+    """获取 K 线数据。"""
+    _load_fetchers()
+    cfg = get_config()
+    key = f"kline_{code}_{scale}_{datalen}"
+
+    if use_cache:
+        cached = cache.get_json(key, cfg.kline_cache_ttl)
+        if cached:
+            return [_dict_to_kline_bar(bar) for bar in cached]
+
+    records = _kline_manager.fetch(code, scale=scale, datalen=datalen)
+    if not records:
+        return []
+
+    bars = [_dict_to_kline_bar(r) for r in records]
+
+    if use_cache and bars:
+        cache.set_json(key, [b.to_dict() for b in bars])
+
+    return bars
+
+
+def get_finance(code: str, use_cache: bool = True) -> list:
+    """获取财务数据。"""
+    _load_fetchers()
+    cfg = get_config()
+    key = f"finance_{code}"
+
+    if use_cache:
+        cached = cache.get_json(key, cfg.finance_cache_ttl)
+        if cached:
+            records = [_dict_to_finance(r) for r in cached]
+            # 校验缓存有效性：至少有一个非零数据点
+            if any(r.eps != 0 or r.roe != 0 for r in records):
+                return records
+            # 零值缓存视为无效，忽略并重新拉取
+
+    result = _finance_manager.fetch(code)
+    if not result:
+        return []
+
+    records = [_dict_to_finance(r) for r in result]
+
+    if use_cache and records:
+        cache.set_json(key, [r.to_dict() for r in records])
+
+    return records
+
+
+# ---------- 内部转换函数（使用 common.to_float / common.to_int） ----------
+
+
+def _dict_to_quote(d: dict) -> Quote:
+    to_float, to_int = _get_common_helpers()
+    return Quote(
+        code=d.get("code", ""),
+        name=d.get("name", ""),
+        price=to_float(d.get("price")),
+        prev_close=to_float(d.get("prev_close")),
+        open=to_float(d.get("open")),
+        high=to_float(d.get("high")),
+        low=to_float(d.get("low")),
+        change_pct=to_float(d.get("change_pct")),
+        change_amt=to_float(d.get("change_amt")),
+        volume=to_int(d.get("volume")),
+        amount=to_float(d.get("amount")),
+        turnover=to_float(d.get("turnover")),
+        pe=to_float(d.get("pe")),
+        pb=to_float(d.get("pb")),
+        total_cap=to_float(d.get("total_cap")),
+        circulating_cap=to_float(d.get("circulating_cap")),
+        source=d.get("source", ""),
+    )
+
+
+def _dict_to_kline_bar(d: dict) -> KlineBar:
+    to_float, to_int = _get_common_helpers()
+    return KlineBar(
+        day=d.get("day", ""),
+        open=to_float(d.get("open")),
+        high=to_float(d.get("high")),
+        low=to_float(d.get("low")),
+        close=to_float(d.get("close")),
+        volume=to_int(d.get("volume")),
+        amount=to_float(d.get("amount")),
+        pct_chg=to_float(d.get("pct_chg")),
+        source=d.get("source", ""),
+    )
+
+
+def _dict_to_finance(d: dict) -> FinanceRecord:
+    """将 fetcher 返回的 dict 转为 FinanceRecord，支持东财原始字段名映射。"""
+    to_float, _ = _get_common_helpers()
+    FIELD_MAP = {
+        "report_date": ["REPORT_DATE", "REPORTDATETIME", "NOTICE_DATE"],
+        "eps": ["EPSJB"],
+        "roe": ["ROEJQ"],
+        "revenue_yoy": ["TOTALOPERATEREVETZ"],
+        "net_profit_yoy": ["PARENTNETPROFITTZ"],
+        "gross_margin": ["XSMLL"],
+        "net_margin": ["XSJLL"],
+        "debt_ratio": ["ZCFZL"],
+        "bps": ["BPS"],
+        "ocf_per_share": ["MGJYXJJE"],
+    }
+
+    def _find(candidates, default=""):
+        for k in candidates:
+            if k in d and d[k] not in (None, "", "-"):
+                return d[k]
+        return default
+
+    return FinanceRecord(
+        report_date=str(_find(FIELD_MAP["report_date"]))[:10],
+        eps=to_float(_find(FIELD_MAP["eps"])),
+        roe=to_float(_find(FIELD_MAP["roe"])),
+        revenue_yoy=to_float(_find(FIELD_MAP["revenue_yoy"])),
+        net_profit_yoy=to_float(_find(FIELD_MAP["net_profit_yoy"])),
+        gross_margin=to_float(_find(FIELD_MAP["gross_margin"])),
+        net_margin=to_float(_find(FIELD_MAP["net_margin"])),
+        debt_ratio=to_float(_find(FIELD_MAP["debt_ratio"])),
+        bps=to_float(_find(FIELD_MAP["bps"])),
+        ocf_per_share=to_float(_find(FIELD_MAP["ocf_per_share"])),
+        source=d.get("source", ""),
+    )

package/scripts/data/cache.py

@@ -0,0 +1,99 @@
+"""统一缓存管理。"""
+import hashlib
+import json
+import os
+import tempfile
+import time
+from pathlib import Path
+from typing import Optional
+
+_DEFAULT_CACHE_DIR = Path(__file__).resolve().parent.parent.parent / ".cache"
+CACHE_DIR = Path(os.getenv("STOCK_CACHE_DIR", str(_DEFAULT_CACHE_DIR)))
+
+
+def _ensure_dir():
+    CACHE_DIR.mkdir(exist_ok=True)
+
+
+def get(key: str, ttl_seconds: int) -> Optional[bytes]:
+    """读取缓存，TTL 超时返回 None。"""
+    _ensure_dir()
+    f = CACHE_DIR / f"{key}.cache"
+    if not f.exists():
+        return None
+    if time.time() - f.stat().st_mtime > ttl_seconds:
+        f.unlink(missing_ok=True)
+        return None
+    return f.read_bytes()
+
+
+def set(key: str, data: bytes):
+    """写入缓存（原子写入：先写临时文件，再 rename）。"""
+    _ensure_dir()
+    f = CACHE_DIR / f"{key}.cache"
+    fd, tmp_path = tempfile.mkstemp(dir=CACHE_DIR, suffix=".tmp")
+    try:
+        os.write(fd, data)
+        os.close(fd)
+        fd = -1
+        os.replace(tmp_path, f)
+    except Exception:
+        if fd >= 0:
+            try:
+                os.close(fd)
+            except OSError:
+                pass
+        Path(tmp_path).unlink(missing_ok=True)
+        raise
+
+
+def get_json(key: str, ttl_seconds: int):
+    """读取 JSON 缓存。"""
+    raw = get(key, ttl_seconds)
+    if raw is None:
+        return None
+    try:
+        return json.loads(raw)
+    except json.JSONDecodeError:
+        return None
+
+
+def set_json(key: str, data):
+    """写入 JSON 缓存。"""
+    set(key, json.dumps(data, ensure_ascii=False).encode())
+
+
+def clear(prefix: str = ""):
+    """清除指定前缀或全部缓存。"""
+    if not CACHE_DIR.exists():
+        return
+    for f in CACHE_DIR.glob("*.cache"):
+        if not prefix or f.stem.startswith(prefix):
+            f.unlink()
+
+
+def cleanup(prefix: str = None, max_age_seconds: int = 86400):
+    """清理过期缓存。prefix 为空时清理所有过期文件。返回清理数量。"""
+    _ensure_dir()
+    cleaned = 0
+    for f in CACHE_DIR.glob("*.cache"):
+        if prefix and not f.name.startswith(prefix):
+            continue
+        if time.time() - f.stat().st_mtime > max_age_seconds:
+            f.unlink(missing_ok=True)
+            cleaned += 1
+    return cleaned
+
+
+def cache_key(url: str) -> str:
+    """用 URL 的 SHA256 生成缓存键。"""
+    return hashlib.sha256(url.encode()).hexdigest()[:32]
+
+
+def cache_key_for_stock(prefix: str, code: str, **params) -> str:
+    """生成股票相关的缓存键，支持按代码清除。
+    格式: {prefix}_{code}_{param_hash}
+    """
+    param_str = "_".join(f"{k}={v}" for k, v in sorted(params.items()))
+    param_hash = hashlib.md5(param_str.encode()).hexdigest()[:8] if param_str else ""
+    return f"{prefix}_{code}_{param_hash}".rstrip("_")

package/scripts/data/config.py

@@ -0,0 +1,49 @@
+"""统一配置管理。"""
+import os
+from dataclasses import dataclass
+
+
+@dataclass
+class DataConfig:
+    """数据层配置。"""
+    # 缓存 TTL (秒)
+    quote_cache_ttl: int = 900       # 15 分钟
+    kline_cache_ttl: int = 21600     # 6 小时
+    finance_cache_ttl: int = 21600   # 6 小时
+    ann_cache_ttl: int = 1800        # 30 分钟
+
+    # 熔断器
+    circuit_failure_threshold: int = 5
+    circuit_recovery_timeout: int = 60
+
+    # 并发
+    max_workers: int = 8
+
+    @classmethod
+    def from_env(cls) -> "DataConfig":
+        """从环境变量加载配置。"""
+        cfg = cls()
+        cfg.quote_cache_ttl = int(os.getenv("DATA_QUOTE_TTL", cfg.quote_cache_ttl))
+        cfg.kline_cache_ttl = int(os.getenv("DATA_KLINE_TTL", cfg.kline_cache_ttl))
+        cfg.finance_cache_ttl = int(os.getenv("DATA_FINANCE_TTL", cfg.finance_cache_ttl))
+        cfg.circuit_failure_threshold = int(os.getenv("DATA_CIRCUIT_THRESHOLD", cfg.circuit_failure_threshold))
+        cfg.circuit_recovery_timeout = int(os.getenv("DATA_CIRCUIT_TIMEOUT", cfg.circuit_recovery_timeout))
+        cfg.max_workers = int(os.getenv("DATA_MAX_WORKERS", cfg.max_workers))
+        return cfg
+
+
+_config = None
+
+
+def get_config() -> DataConfig:
+    """获取全局配置单例。"""
+    global _config
+    if _config is None:
+        _config = DataConfig.from_env()
+    return _config
+
+
+def reset_config():
+    """重置配置单例（用于测试或动态重载）。"""
+    global _config
+    _config = None

package/scripts/data/industry_thresholds.json

@@ -0,0 +1,199 @@
+{
+  "_description": "行业差异化阈值表。按申万一级行业分组，定义各行业的基本面和估值阈值。",
+  "_usage": "screener.py 和 technical/scoring.py 按行业选择阈值，避免一刀切误判。",
+  "金融": {
+    "roe_min": 10,
+    "roe_excellent": 15,
+    "gross_margin_min": 0,
+    "debt_ratio_max": 90,
+    "pe_undervalued": 8,
+    "pe_reasonable": 12,
+    "pe_expensive": 18,
+    "peg_undervalued": 0.8,
+    "peg_reasonable": 1.5,
+    "vol_threshold": 0.015,
+    "notes": "银行/保险 ROE 天然 10-15%，负债率 90%+ 属正常，用资本充足率替代负债率",
+    "profit_growth_excellent": 15,
+    "revenue_growth_excellent": 10
+  },
+  "银行": {
+    "roe_min": 8,
+    "roe_excellent": 12,
+    "gross_margin_min": 0,
+    "debt_ratio_max": 95,
+    "pe_undervalued": 5,
+    "pe_reasonable": 8,
+    "pe_expensive": 12,
+    "peg_undervalued": 0.5,
+    "peg_reasonable": 1.0,
+    "vol_threshold": 0.012,
+    "notes": "银行 ROE 天然偏低（8-12%），PE 极低（5-8倍），高负债率属正常",
+    "profit_growth_excellent": 12,
+    "revenue_growth_excellent": 8
+  },
+  "券商": {
+    "roe_min": 8,
+    "roe_excellent": 18,
+    "gross_margin_min": 0,
+    "debt_ratio_max": 85,
+    "pe_undervalued": 10,
+    "pe_reasonable": 18,
+    "pe_expensive": 28,
+    "peg_undervalued": 0.8,
+    "peg_reasonable": 1.5,
+    "vol_threshold": 0.025,
+    "notes": "券商 ROE 波动大（牛市高熊市低），PE 中等，周期性强",
+    "profit_growth_excellent": 25,
+    "revenue_growth_excellent": 20
+  },
+  "消费": {
+    "roe_min": 15,
+    "roe_excellent": 25,
+    "gross_margin_min": 30,
+    "debt_ratio_max": 50,
+    "pe_undervalued": 15,
+    "pe_reasonable": 25,
+    "pe_expensive": 40,
+    "peg_undervalued": 0.8,
+    "peg_reasonable": 1.5,
+    "vol_threshold": 0.02,
+    "notes": "白酒/食品 ROE 天然高，毛利率 60%+ 常见，估值享受溢价",
+    "profit_growth_excellent": 25,
+    "revenue_growth_excellent": 20
+  },
+  "科技": {
+    "roe_min": 10,
+    "roe_excellent": 20,
+    "gross_margin_min": 20,
+    "debt_ratio_max": 60,
+    "pe_undervalued": 20,
+    "pe_reasonable": 40,
+    "pe_expensive": 60,
+    "peg_undervalued": 0.8,
+    "peg_reasonable": 2.0,
+    "vol_threshold": 0.03,
+    "notes": "科技股 PE 30-50 常态，高研发投入导致短期 ROE 偏低，不能用传统标准",
+    "profit_growth_excellent": 50,
+    "revenue_growth_excellent": 35
+  },
+  "半导体": {
+    "roe_min": 8,
+    "roe_excellent": 15,
+    "gross_margin_min": 25,
+    "debt_ratio_max": 55,
+    "pe_undervalued": 25,
+    "pe_reasonable": 50,
+    "pe_expensive": 80,
+    "peg_undervalued": 0.8,
+    "peg_reasonable": 2.0,
+    "vol_threshold": 0.035,
+    "notes": "半导体重资产，ROE 天然偏低（8-15%），PE 容忍度高（25-80），研发投入大",
+    "profit_growth_excellent": 60,
+    "revenue_growth_excellent": 40
+  },
+  "软件": {
+    "roe_min": 12,
+    "roe_excellent": 22,
+    "gross_margin_min": 50,
+    "debt_ratio_max": 45,
+    "pe_undervalued": 30,
+    "pe_reasonable": 60,
+    "pe_expensive": 100,
+    "peg_undervalued": 1.0,
+    "peg_reasonable": 2.5,
+    "vol_threshold": 0.032,
+    "notes": "软件轻资产高毛利（50%+），ROE 偏高，PE 估值更激进（30-100）",
+    "profit_growth_excellent": 45,
+    "revenue_growth_excellent": 30
+  },
+  "周期": {
+    "roe_min": 8,
+    "roe_excellent": 15,
+    "gross_margin_min": 10,
+    "debt_ratio_max": 65,
+    "pe_undervalued": 8,
+    "pe_reasonable": 15,
+    "pe_expensive": 25,
+    "peg_undervalued": 0.5,
+    "peg_reasonable": 1.0,
+    "vol_threshold": 0.025,
+    "notes": "周期股用 3 年均值 ROE，单年高增速往往是周期顶部信号",
+    "profit_growth_excellent": 30,
+    "revenue_growth_excellent": 25
+  },
+  "医药": {
+    "roe_min": 12,
+    "roe_excellent": 20,
+    "gross_margin_min": 40,
+    "debt_ratio_max": 45,
+    "pe_undervalued": 15,
+    "pe_reasonable": 30,
+    "pe_expensive": 50,
+    "peg_undervalued": 0.8,
+    "peg_reasonable": 1.8,
+    "vol_threshold": 0.028,
+    "notes": "创新药高 PE 合理，仿制药/中药估值较低，需区分细分赛道",
+    "profit_growth_excellent": 35,
+    "revenue_growth_excellent": 25
+  },
+  "制造": {
+    "roe_min": 12,
+    "roe_excellent": 18,
+    "gross_margin_min": 15,
+    "debt_ratio_max": 60,
+    "pe_undervalued": 12,
+    "pe_reasonable": 22,
+    "pe_expensive": 35,
+    "peg_undervalued": 0.8,
+    "peg_reasonable": 1.5,
+    "vol_threshold": 0.025,
+    "notes": "制造业分化大，高端制造（新能源/半导体）估值高于传统制造",
+    "profit_growth_excellent": 35,
+    "revenue_growth_excellent": 25
+  },
+  "地产": {
+    "roe_min": 8,
+    "roe_excellent": 15,
+    "gross_margin_min": 15,
+    "debt_ratio_max": 80,
+    "pe_undervalued": 5,
+    "pe_reasonable": 10,
+    "pe_expensive": 15,
+    "peg_undervalued": 0.5,
+    "peg_reasonable": 1.0,
+    "vol_threshold": 0.028,
+    "notes": "地产用净负债率<100% 替代资产负债率，PE 天然低",
+    "profit_growth_excellent": 15,
+    "revenue_growth_excellent": 10
+  },
+  "能源": {
+    "roe_min": 10,
+    "roe_excellent": 18,
+    "gross_margin_min": 15,
+    "debt_ratio_max": 55,
+    "pe_undervalued": 6,
+    "pe_reasonable": 12,
+    "pe_expensive": 20,
+    "peg_undervalued": 0.5,
+    "peg_reasonable": 1.0,
+    "vol_threshold": 0.022,
+    "notes": "煤炭/石油周期性强，高分红低估值是常态",
+    "profit_growth_excellent": 20,
+    "revenue_growth_excellent": 15
+  },
+  "默认": {
+    "roe_min": 12,
+    "roe_excellent": 18,
+    "gross_margin_min": 20,
+    "debt_ratio_max": 60,
+    "pe_undervalued": 15,
+    "pe_reasonable": 25,
+    "pe_expensive": 40,
+    "peg_undervalued": 0.8,
+    "peg_reasonable": 1.5,
+    "vol_threshold": 0.025,
+    "notes": "未匹配行业的默认阈值",
+    "profit_growth_excellent": 40,
+    "revenue_growth_excellent": 30
+  }
+}

package/scripts/data/portfolio_example.json

@@ -0,0 +1,14 @@
+{
+  "version": 1,
+  "note": "示例持仓配置。复制此文件到 portfolio.json 并修改 codes 字段为你的实际持仓。",
+  "codes": [
+    "sh600989",
+    "sz000807",
+    "sh518880",
+    "sh603993",
+    "sz002415",
+    "sh603799",
+    "sh601211",
+    "sz300438"
+  ]
+}

package/scripts/data/sector_etf.csv

@@ -0,0 +1,14 @@
+code,name,category,bk_code
+sh510050,上证50ETF,宽基,
+sh510300,沪深300ETF,宽基,
+sh510500,中证500ETF,宽基,
+sh512010,医药ETF,行业,BK0465
+sh512480,半导体ETF,行业,BK0917
+sh512690,白酒ETF,行业,BK0896
+sh512800,银行ETF,行业,BK0475
+sh513120,港股创新药ETF,跨境,
+sh518880,黄金ETF,商品,
+sh515030,新能源车ETF,行业,BK0900
+sh512660,军工ETF,行业,BK0490
+sh515790,光伏ETF,行业,BK0588
+sh516160,新能源ETF,行业,BK0493

package/scripts/data/sector_mapping.json

@@ -0,0 +1,64 @@
+{
+  "_meta": {
+    "description": "板块名 → 东财BK代码映射，供 refresh_pool.py 使用",
+    "updated": "2026-06-05",
+    "api": "https://push2.eastmoney.com/api/qt/clist/get?fs=b:{BK_CODE}"
+  },
+  "金融": {
+    "bk_codes": ["BK0475", "BK0449"],
+    "note": "银行+证券"
+  },
+  "消费": {
+    "bk_codes": ["BK0896", "BK0438"],
+    "note": "白酒+食品饮料"
+  },
+  "医药": {
+    "bk_codes": ["BK0465", "BK0899"],
+    "note": "医药生物+创新药"
+  },
+  "资源": {
+    "bk_codes": ["BK0519", "BK1649"],
+    "note": "稀缺资源+油气"
+  },
+  "电力": {
+    "bk_codes": ["BK1024", "BK0432"],
+    "note": "绿色电力+电力"
+  },
+  "石化": {
+    "bk_codes": ["BK0431"],
+    "note": "石油石化"
+  },
+  "高股息": {
+    "bk_codes": [],
+    "note": "跨板块主题，从其他板块筛选PE<20且ROE>8%的标的",
+    "filter": "dividend"
+  },
+  "科技": {
+    "bk_codes": ["BK0448", "BK0482"],
+    "note": "消费电子+计算机"
+  },
+  "机器人": {
+    "bk_codes": ["BK1184", "BK1090"],
+    "note": "人形机器人+机器人概念"
+  },
+  "PCB/AI算力": {
+    "bk_codes": ["BK0877", "BK1134"],
+    "note": "PCB+算力"
+  },
+  "军工": {
+    "bk_codes": ["BK0490"],
+    "note": "军工"
+  },
+  "半导体": {
+    "bk_codes": ["BK0917"],
+    "note": "半导体概念"
+  },
+  "新能源": {
+    "bk_codes": ["BK0900"],
+    "note": "新能源车"
+  },
+  "光伏": {
+    "bk_codes": ["BK0588"],
+    "note": "光伏概念"
+  }
+}