quantcli 0.2.1__tar.gz → 0.2.5__tar.gz

{quantcli-0.2.1/quantcli.egg-info → quantcli-0.2.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: quantcli
-Version: 0.2.1
+Version: 0.2.5
 Summary: 面向AI的多因子量化选股策略挖掘工具，AI Agent 友好 CLI
 Author-email: QuantCLI Team <quantcli@example.com>
 Project-URL: repository, https://github.com/wumu2013/quantcli

{quantcli-0.2.1 → quantcli-0.2.5}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "quantcli"
-version = "0.2.1"
+version = "0.2.5"
 description = "面向AI的多因子量化选股策略挖掘工具，AI Agent 友好 CLI"
 readme = "README.md"
 requires-python = ">=3.9"

{quantcli-0.2.1 → quantcli-0.2.5}/quantcli/cli.py

@@ -107,10 +107,13 @@ def data(ctx):
 @click.option(
     "--source",
     type=click.Choice(["akshare", "baostock", "mixed", "mysql"]),
-    default="mixed",
+    default="mysql",
     help="数据源 (akshare/baostock/mixed/mysql)",
 )
 @click.option("--use-cache/--no-cache", default=True, help="是否使用缓存")
+@click.option(
+    "--cache-raw/--no-cache-raw", default=False, help="是否缓存原始数据到parquet"
+)
 @click.option(
     "--output",
     "-o",
@@ -119,7 +122,7 @@ def data(ctx):
     help="输出文件路径 (如: data.csv)",
 )
 @click.pass_context
-def data_fetch(ctx, symbol, start, end, source, use_cache, output):
+def data_fetch(ctx, symbol, start, end, source, use_cache, cache_raw, output):
     """获取股票日线数据
 
     股票代码:
@@ -131,6 +134,7 @@ def data_fetch(ctx, symbol, start, end, source, use_cache, output):
       quantcli data fetch 000001 --start 2025-01-01 --end 2025-12-31
       quantcli data fetch 600519 --start 2025-01-01 --source mysql
       quantcli data fetch 000001 --start 2025-01-01 -o data.csv
+      quantcli data fetch 000001 --start 2025-01-01 --cache-raw
     """
     if end is None:
         end = today()
@@ -139,7 +143,7 @@ def data_fetch(ctx, symbol, start, end, source, use_cache, output):
         f"Fetching {symbol} data from {format_date(start)} to {format_date(end)}..."
     )
 
-    config = DataConfig(source=source)
+    config = DataConfig(source=source, cache_raw=cache_raw)
     dm = DataManager(config)
 
     try:
@@ -169,7 +173,93 @@ def data_fetch(ctx, symbol, start, end, source, use_cache, output):
             click.echo(f"Saved to {output}")
 
     except Exception as e:
-        click.echo(f"Error: {e}", err=True)
+        error_msg = str(e)
+        if "Access denied" in error_msg or "Connection refused" in error_msg:
+            click.echo(f"MySQL 连接失败: 请检查数据库配置", err=True)
+            click.echo(
+                "提示: 使用 --source akshare 或 --source baostock 可以切换到其他数据源",
+                err=True,
+            )
+        else:
+            click.echo(f"Error: {e}", err=True)
+        sys.exit(1)
+
+
+@data.command("fetch-fundamental")
+@click.argument("symbols", type=str, nargs=-1)
+@click.option(
+    "--date",
+    type=date_type,
+    default=None,
+    help="截止日期 (YYYY-MM-DD，默认今天)",
+)
+@click.option(
+    "--source",
+    type=click.Choice(["mysql", "baostock"]),
+    default="mysql",
+    help="数据源 (mysql/baostock)",
+)
+@click.option(
+    "--output",
+    "-o",
+    type=click.Path(),
+    default=None,
+    help="输出文件路径 (如: data.csv)",
+)
+@click.pass_context
+def data_fetch_fundamental(ctx, symbols, date, source, output):
+    """获取股票基本面数据
+
+    股票代码:
+      000001 - 平安银行
+      600519 - 贵州茅台
+
+    示例:
+      quantcli data fetch-fundamental 000001 600519
+      quantcli data fetch-fundamental 600519 --date 2024-12-31
+      quantcli data fetch-fundamental 000001 --source baostock -o fundamentals.csv
+    """
+    if not symbols:
+        click.echo("Error: At least one symbol is required")
+        sys.exit(1)
+
+    if date is None:
+        date = today()
+
+    click.echo(
+        f"Fetching fundamental data for {len(symbols)} symbols at {format_date(date)}..."
+    )
+
+    config = DataConfig(source=source)
+    dm = DataManager(config)
+
+    try:
+        df = dm.datasource.get_fundamental(list(symbols), date)
+
+        if df.empty:
+            click.echo(f"No fundamental data found for symbols")
+            return
+
+        click.echo(f"Retrieved {len(df)} rows")
+
+        click.echo(df.to_string())
+
+        if output:
+            if output.endswith(".csv"):
+                df.to_csv(output, index=False)
+            elif output.endswith(".parquet"):
+                df.to_parquet(output, index=False)
+            else:
+                df.to_csv(output, index=False)
+            click.echo(f"Saved to {output}")
+
+    except Exception as e:
+        error_msg = str(e)
+        if "Access denied" in error_msg or "Connection refused" in error_msg:
+            click.echo(f"MySQL 连接失败: 请检查数据库配置", err=True)
+            click.echo("提示: 使用 --source baostock 可以切换到其他数据源", err=True)
+        else:
+            click.echo(f"Error: {e}", err=True)
         sys.exit(1)
 
 
@@ -191,7 +281,7 @@ def data_cache(ctx):
 @click.pass_context
 def data_cache_ls(ctx, pattern, sort):
     """列出缓存文件"""
-    config = DataConfig()
+    config = DataConfig(source="mysql")
     dm = DataManager(config)
     sizes = dm.get_cache_size()
 
@@ -229,7 +319,7 @@ def data_cache_ls(ctx, pattern, sort):
 @click.pass_context
 def data_cache_clean(ctx, older_than):
     """清理缓存文件"""
-    config = DataConfig()
+    config = DataConfig(source="mysql")
     dm = DataManager(config)
 
     count = dm.clear_cache(older_than=older_than)
@@ -240,7 +330,7 @@ def data_cache_clean(ctx, older_than):
 @click.pass_context
 def data_health(ctx):
     """检查数据源健康状态"""
-    config = DataConfig()
+    config = DataConfig(source="mysql")
     dm = DataManager(config)
     health = dm.health_check()
 
@@ -305,7 +395,7 @@ def factor_run(ctx, name, expr, symbol, start, end, output):
     click.echo(f"Computing factor '{name}'...")
 
     # 创建引擎
-    config = DataConfig()
+    config = DataConfig(source="mysql")
     dm = DataManager(config)
     engine = FactorEngine(dm)
 
@@ -366,7 +456,7 @@ def factor_eval(ctx, name, symbol, start, end, method):
 
     click.echo(f"Evaluating factor '{name}'...")
 
-    config = DataConfig()
+    config = DataConfig(source="mysql")
     dm = DataManager(config)
     engine = FactorEngine(dm)
 
@@ -731,7 +821,7 @@ def filter_run(
             if daily_conditions:
                 # 只对候选获取日线数据
                 click.echo(f"Fetching daily data for {len(candidates)} candidates...")
-                dm = DataManager(DataConfig(source="mixed"))
+                dm = DataManager(DataConfig(source="mysql"))
 
                 price_data = {}
                 for symbol in candidates:
@@ -829,7 +919,7 @@ def filter_run(
             # 阶段2: 日线
             click.echo(f"\n=== Stage 2: Daily Screening ===")
             click.echo(f"Fetching daily data for {len(candidates)} candidates...")
-            dm = DataManager(DataConfig(source="mixed"))
+            dm = DataManager(DataConfig(source="mysql"))
 
             price_data = {}
             for symbol in candidates:
@@ -885,7 +975,7 @@ def filter_run(
                 return
 
             click.echo(f"Fetching price data for {len(candidates)} candidates...")
-            dm = DataManager(DataConfig(source="mixed"))
+            dm = DataManager(DataConfig(source="mysql"))
 
             price_data = {}
             for symbol in candidates:
@@ -1303,7 +1393,7 @@ def analyze_ic(ctx, expr, name, symbol, start, end, period, window, method):
     click.echo(f"Forward period: {period} days, Window: {window}")
 
     # 获取数据
-    config = DataConfig()
+    config = DataConfig(source="mysql")
     dm = DataManager(config)
     df = dm.get_daily(symbol, start, end)
 
@@ -1422,7 +1512,7 @@ def analyze_batch(ctx, dir, symbol, start, end, period, window, top, output):
     click.echo(f"Analyzing with period={period}d, window={window}...\n")
 
     # 获取数据
-    config = DataConfig()
+    config = DataConfig(source="mysql")
     dm = DataManager(config)
     df = dm.get_daily(symbol, start, end)
 
@@ -1541,7 +1631,7 @@ def config():
 @click.pass_context
 def config_show(ctx):
     """显示当前配置"""
-    config = DataConfig()
+    config = DataConfig(source="mysql")
     click.echo("QuantCLI Configuration:")
     click.echo(f"  data.source: {config.source}")
     click.echo(f"  data.cache_dir: {config.cache_dir}")

{quantcli-0.2.1 → quantcli-0.2.5}/quantcli/core/data.py

@@ -30,6 +30,7 @@ logger = get_logger(__name__)
 class DataConfig:
     """数据管理配置"""
 
+    cache_raw: bool = False
     source: str = "akshare"
     cache_dir: str = "./data"
     parallel: int = 4
@@ -141,8 +142,7 @@ class DataManager:
             logger.warning(f"No data for {symbol}")
             return df
 
-        # 缓存数据
-        if use_cache:
+        if use_cache and self.config.cache_raw:
             self._save_to_cache(cache_file, df, f"stock_daily_{symbol}")
 
         return df
@@ -257,7 +257,7 @@ class DataManager:
 
         df = self.datasource.get_index_daily(symbol, start_date, end_date)
 
-        if use_cache and not df.empty:
+        if use_cache and self.config.cache_raw and not df.empty:
             self._save_to_cache(cache_file, df, f"index_daily_{symbol}")
 
         return df

{quantcli-0.2.1 → quantcli-0.2.5}/quantcli/datasources/mysql.py

@@ -88,9 +88,14 @@ class MySQLDataSource(DataSource):
         self._prefix = config_dict["table_prefix"]
         self._autocommit = autocommit
         self._conn = None
+        self._tables_initialized = False
 
-        # 初始化数据库表
+    def _ensure_tables(self):
+        """确保数据库表已初始化（懒加载）"""
+        if self._tables_initialized:
+            return
         self._init_tables()
+        self._tables_initialized = True
 
     def _get_connection(self):
         """获取数据库连接"""
@@ -208,9 +213,8 @@ class MySQLDataSource(DataSource):
         """批量转换为 MySQL 新格式"""
         return [to_mysql(s) for s in symbols]
 
-    def get_daily(
-        self, symbol: str, start_date, end_date, fields: Optional[List[str]] = None
-    ) -> pd.DataFrame:
+    def get_daily(self, symbol: str, start_date, end_date) -> pd.DataFrame:
+        self._ensure_tables()
         """获取日线数据
 
         Args:
@@ -559,6 +563,7 @@ class MySQLDataSource(DataSource):
     # ==================== 股票列表和日历 ====================
 
     def get_stock_list(self, market: str = "all") -> pd.DataFrame:
+        self._ensure_tables()
         """获取股票列表"""
         conn = self._get_connection()
 
@@ -595,6 +600,7 @@ class MySQLDataSource(DataSource):
             )
 
     def get_trading_calendar(self, exchange: str = "SSE") -> List[date]:
+        self._ensure_tables()
         """获取交易日历"""
         conn = self._get_connection()
 
@@ -619,6 +625,7 @@ class MySQLDataSource(DataSource):
     def get_fundamental(
         self, symbols: List[str], date, indicators: Optional[List[str]] = None
     ) -> pd.DataFrame:
+        self._ensure_tables()
         """获取基本面数据
 
         Args:

{quantcli-0.2.1 → quantcli-0.2.5/quantcli.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: quantcli
-Version: 0.2.1
+Version: 0.2.5
 Summary: 面向AI的多因子量化选股策略挖掘工具，AI Agent 友好 CLI
 Author-email: QuantCLI Team <quantcli@example.com>
 Project-URL: repository, https://github.com/wumu2013/quantcli

{quantcli-0.2.1 → quantcli-0.2.5}/tests/test_data_manager.py

@@ -52,7 +52,7 @@ class TestGetDailyCacheLogic:
         self, temp_cache_dir, mock_datasource
     ):
         """缓存存在且完整 → 直接返回缓存"""
-        config = DataConfig(cache_dir=temp_cache_dir, source="mock")
+        config = DataConfig(cache_dir=temp_cache_dir, source="mock", cache_raw=True)
         dm = DataManager(config)
         dm._datasource = mock_datasource
 
@@ -95,7 +95,7 @@ class TestGetDailyCacheLogic:
 
     def test_cache_incomplete_fetches_full_data(self, temp_cache_dir, mock_datasource):
         """缓存不完整 → 全量拉取"""
-        config = DataConfig(cache_dir=temp_cache_dir, source="mock")
+        config = DataConfig(cache_dir=temp_cache_dir, source="mock", cache_raw=True)
         dm = DataManager(config)
         dm._datasource = mock_datasource
 
@@ -138,7 +138,7 @@ class TestGetDailyCacheLogic:
 
     def test_no_cache_fetches_full_data(self, temp_cache_dir, mock_datasource):
         """缓存不存在 → 全量拉取"""
-        config = DataConfig(cache_dir=temp_cache_dir, source="mock")
+        config = DataConfig(cache_dir=temp_cache_dir, source="mock", cache_raw=True)
         dm = DataManager(config)
         dm._datasource = mock_datasource
 
@@ -165,7 +165,7 @@ class TestGetDailyCacheLogic:
 
     def test_no_calendar_fetches_full_data(self, temp_cache_dir, mock_datasource):
         """交易日历不可用 → 直接拉取，不使用缓存"""
-        config = DataConfig(cache_dir=temp_cache_dir, source="mock")
+        config = DataConfig(cache_dir=temp_cache_dir, source="mock", cache_raw=True)
         dm = DataManager(config)
         dm._datasource = mock_datasource
 
@@ -200,7 +200,7 @@ class TestGetDailyCacheLogic:
         self, temp_cache_dir, mock_datasource
     ):
         """交易日历不可用 + 无缓存 → 全量拉取"""
-        config = DataConfig(cache_dir=temp_cache_dir, source="mock")
+        config = DataConfig(cache_dir=temp_cache_dir, source="mock", cache_raw=True)
         dm = DataManager(config)
         dm._datasource = mock_datasource
 
@@ -222,7 +222,9 @@ class TestGetDailyBatch:
 
     def test_batch_uses_parallel_when_enabled(self, temp_cache_dir, mock_datasource):
         """批量获取启用并行"""
-        config = DataConfig(cache_dir=temp_cache_dir, source="mock", parallel=2)
+        config = DataConfig(
+            cache_dir=temp_cache_dir, source="mock", parallel=2, cache_raw=True
+        )
         dm = DataManager(config)
         dm._datasource = mock_datasource
 
@@ -240,7 +242,7 @@ class TestCacheCleanup:
 
     def test_clear_cache_removes_files(self, temp_cache_dir, mock_datasource):
         """清理缓存删除文件"""
-        config = DataConfig(cache_dir=temp_cache_dir, source="mock")
+        config = DataConfig(cache_dir=temp_cache_dir, source="mock", cache_raw=True)
         dm = DataManager(config)
         dm._datasource = mock_datasource