quantcli 0.1.0__py3-none-any.whl

quantcli/cli.py

@@ -0,0 +1,996 @@
+"""QuantCLI 命令行界面
+
+提供数据获取、因子计算、回测和配置管理命令。
+
+Usage:
+    quantcli --help
+    quantcli data fetch --symbol 600519 --start 2020-01-01
+    quantcli factor run -n momentum -e "(close / delay(close, 20)) - 1"
+    quantcli backtest run -s dual_ma.py --start 2020-01-01
+"""
+
+import sys
+from datetime import date, timedelta
+from typing import Optional
+
+import click
+import pandas as pd
+from tabulate import tabulate
+
+from .core import DataManager, DataConfig, FactorEngine, BacktestEngine, BacktestConfig, Strategy
+from .factors import FactorDefinition
+from .utils import parse_date, format_date, today, get_logger, TimeContext
+
+logger = get_logger(__name__)
+
+
+# =============================================================================
+# 通用选项
+# =============================================================================
+
+def verbose_option(f):
+    """Verbose output option"""
+    def callback(ctx, param, value):
+        ctx.ensure_object(dict)
+        ctx.obj["verbose"] = value
+        if value:
+            import logging
+            logging.basicConfig(level=logging.DEBUG)
+    return click.option("-v", "--verbose", is_flag=True, help="Enable verbose output", callback=callback)(f)
+
+
+def date_type(s: str) -> date:
+    """Click type for date validation"""
+    try:
+        return parse_date(s)
+    except Exception:
+        raise click.BadParameter(f"Invalid date: {s}. Use YYYY-MM-DD format")
+
+
+# =============================================================================
+# 主命令
+# =============================================================================
+
+@click.group()
+@click.version_option(version="0.1.0", prog_name="quantcli")
+@click.option("--verbose", "-v", is_flag=True, help="Enable verbose output", default=False)
+@click.pass_context
+def quantcli(ctx, verbose):
+    """QuantCLI - 量化因子挖掘与回测工具"""
+    ctx.ensure_object(dict)
+    ctx.obj["verbose"] = verbose
+
+    if verbose:
+        import logging
+        logging.basicConfig(level=logging.DEBUG)
+
+    logger.info(f"QuantCLI v0.1.0 started")
+
+
+# =============================================================================
+# Data 命令
+# =============================================================================
+
+@quantcli.group()
+def data():
+    """数据获取与管理"""
+    pass
+
+
+@data.command("fetch")
+@click.argument("symbol", type=str)
+@click.option("--start", type=date_type, required=True, help="Start date (YYYY-MM-DD)")
+@click.option("--end", type=date_type, default=None, help="End date (YYYY-MM-DD)")
+@click.option("--source", type=click.Choice(["akshare", "baostock", "mixed"]), default="mixed", help="Data source")
+@click.option("--use-cache/--no-cache", default=True, help="Use cache")
+@click.option("--output", type=click.Path(), default=None, help="Output file path")
+@click.pass_context
+def data_fetch(ctx, symbol, start, end, source, use_cache, output):
+    """获取股票日线数据"""
+    if end is None:
+        end = today()
+
+    click.echo(f"Fetching {symbol} data from {format_date(start)} to {format_date(end)}...")
+
+    config = DataConfig(source=source)
+    dm = DataManager(config)
+
+    try:
+        df = dm.get_daily(symbol, start, end, use_cache=use_cache)
+
+        if df.empty:
+            click.echo(f"No data found for {symbol}")
+            return
+
+        click.echo(f"Retrieved {len(df)} rows")
+
+        # 数据统计
+        if "close" in df.columns:
+            click.echo(f"Close: {df['close'].min():.2f} - {df['close'].max():.2f}")
+
+        # 输出到文件
+        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:
+        click.echo(f"Error: {e}", err=True)
+        sys.exit(1)
+
+
+@data.group("cache")
+@click.pass_context
+def data_cache():
+    """缓存管理"""
+    pass
+
+
+@data_cache.command("ls")
+@click.pass_context
+def data_cache_ls(ctx):
+    """列出缓存文件"""
+    config = DataConfig()
+    dm = DataManager(config)
+    sizes = dm.get_cache_size()
+
+    if not sizes or "_total" not in sizes:
+        click.echo("No cached data found")
+        return
+
+    rows = []
+    for k, v in sizes.items():
+        if k != "_total":
+            rows.append([k, v])
+
+    if rows:
+        click.echo("Cached files:")
+        click.echo(tabulate(rows, headers=["File", "Size"], tablefmt="simple"))
+    else:
+        click.echo("No cached files found")
+
+    click.echo(f"\nTotal: {sizes.get('_total', '0')}")
+
+
+@data_cache.command("clean")
+@click.option("--older-than", type=int, default=None, help="Remove files older than N days")
+@click.pass_context
+def data_cache_clean(ctx, older_than):
+    """清理缓存文件"""
+    config = DataConfig()
+    dm = DataManager(config)
+
+    count = dm.clear_cache(older_than=older_than)
+    click.echo(f"Cleaned {count} cache files")
+
+
+@data.command("health")
+@click.pass_context
+def data_health(ctx):
+    """检查数据源健康状态"""
+    config = DataConfig()
+    dm = DataManager(config)
+    health = dm.health_check()
+
+    click.echo("Data Manager Health:")
+    for k, v in health.items():
+        if k == "cache":
+            continue
+        click.echo(f"  {k}: {v}")
+
+
+# =============================================================================
+# Factor 命令
+# =============================================================================
+
+@quantcli.group()
+def factor():
+    """因子定义与计算"""
+    pass
+
+
+@factor.command("run")
+@click.option("--name", "-n", required=True, help="Factor name")
+@click.option("--expr", "-e", required=True, help="Factor expression/formula")
+@click.option("--symbol", default="600519", help="Stock symbol for testing")
+@click.option("--start", type=date_type, default="2020-01-01", help="Start date")
+@click.option("--end", type=date_type, default=None, help="End date")
+@click.option("--output", type=click.Path(), default=None, help="Output file path")
+@click.pass_context
+def factor_run(ctx, name, expr, symbol, start, end, output):
+    """运行因子计算"""
+    if end is None:
+        end = today()
+
+    click.echo(f"Computing factor '{name}'...")
+
+    # 创建引擎
+    config = DataConfig()
+    dm = DataManager(config)
+    engine = FactorEngine(dm)
+
+    # 注册因子
+    factor = FactorDefinition(name=name, type="technical", expr=expr)
+    engine.register(factor)
+
+    # 获取数据
+    df = dm.get_daily(symbol, start, end)
+    if df.empty:
+        click.echo(f"No data for {symbol}")
+        return
+
+    # 添加 returns 列
+    if "close" in df.columns:
+        df = df.copy()
+        df["returns"] = df["close"].pct_change()
+
+    # 计算因子
+    try:
+        result = engine.compute(name, df)
+
+        click.echo(f"Factor computed: {len(result)} values")
+        if not result.empty:
+            click.echo(f"  Mean: {result.mean():.4f}")
+            click.echo(f"  Std: {result.std():.4f}")
+            click.echo(f"  Min: {result.min():.4f}")
+            click.echo(f"  Max: {result.max():.4f}")
+
+        # 保存结果
+        if output:
+            result_df = result.reset_index()
+            result_df.columns = ["date", name]
+            result_df.to_csv(output, index=False)
+            click.echo(f"Saved to {output}")
+
+    except Exception as e:
+        click.echo(f"Error computing factor: {e}", err=True)
+        sys.exit(1)
+
+
+@factor.command("eval")
+@click.argument("name", type=str)
+@click.option("--symbol", default="600519", help="Stock symbol")
+@click.option("--start", type=date_type, default="2020-01-01", help="Start date")
+@click.option("--end", type=date_type, default=None, help="End date")
+@click.option("--method", type=click.Choice(["ic", "quantile", "full"]), default="ic", help="Evaluation method")
+@click.pass_context
+def factor_eval(ctx, name, symbol, start, end, method):
+    """评估因子表现"""
+    if end is None:
+        end = today()
+
+    click.echo(f"Evaluating factor '{name}'...")
+
+    config = DataConfig()
+    dm = DataManager(config)
+    engine = FactorEngine(dm)
+
+    df = dm.get_daily(symbol, start, end)
+    if df.empty:
+        click.echo(f"No data for {symbol}")
+        return
+
+    # 添加 returns 列
+    df = df.copy()
+    df["returns"] = df["close"].pct_change()
+
+    try:
+        if method == "ic":
+            result = engine.evaluate_ic(name, df)
+            click.echo(f"\nIC Analysis for {name}:")
+            if "ic_stats" in result:
+                stats = result["ic_stats"]
+                click.echo(f"  IC Mean: {stats.get('ic_mean', 0):.4f}")
+                click.echo(f"  IC Std: {stats.get('ic_std', 0):.4f}")
+                click.echo(f"  IC IR: {stats.get('ic_ir', 0):.4f}")
+
+        elif method == "quantile":
+            result = engine.evaluate_quantiles(name, df)
+            click.echo(f"\nQuantile Analysis for {name}:")
+            click.echo(f"  Groups: {result.get('groups', 10)}")
+            click.echo(f"  Long-Short Return: {result.get('long_short_return', 0):.4f}")
+
+        elif method == "full":
+            result = engine.evaluate_full(name, df)
+            click.echo(f"\nFull Evaluation for {name}:")
+            click.echo(f"  IC Mean: {result.ic_mean:.4f}")
+            click.echo(f"  IC IR: {result.ic_ir:.4f}")
+            click.echo(f"  Sample Size: {result.sample_size}")
+
+    except Exception as e:
+        click.echo(f"Error evaluating factor: {e}", err=True)
+        sys.exit(1)
+
+
+@factor.command("list")
+@click.pass_context
+def factor_list(ctx):
+    """列出已注册的因子"""
+    engine = FactorEngine()
+    factors = engine.registry.list_all()
+
+    if not factors:
+        click.echo("No factors registered")
+        return
+
+    click.echo("Registered factors:")
+    for name in factors:
+        factor = engine.registry.get(name)
+        if factor:
+            click.echo(f"  - {name}: {factor.formula[:50]}...")
+
+
+@factor.command("run-file")
+@click.option("--file", "-f", required=True, type=click.Path(exists=True), help="Factor config YAML file")
+@click.option("--symbol", default="600519", help="Stock symbol for testing")
+@click.option("--start", type=date_type, default="2020-01-01", help="Start date")
+@click.option("--end", type=date_type, default=None, help="End date")
+@click.option("--output", type=click.Path(), default=None, help="Output file path")
+@click.pass_context
+def factor_run_file(ctx, file, symbol, start, end, output):
+    """从 YAML 配置文件运行因子计算"""
+    from .factors import load_strategy, FactorComputer
+
+    if end is None:
+        end = today()
+
+    click.echo(f"Loading factor config from {file}...")
+
+    try:
+        config = load_strategy(file)
+        click.echo(f"Config: {config.name} (v{config.version})")
+
+        # 创建数据管理器
+        config_data = DataConfig()
+        dm = DataManager(config_data)
+
+        # 获取数据
+        df = dm.get_daily(symbol, start, end)
+        if df.empty:
+            click.echo(f"No data for {symbol}")
+            return
+
+        click.echo(f"Data: {len(df)} rows")
+
+        # 获取权重配置中的因子
+        weights = config.ranking.get("weights", {})
+        from .factors.loader import load_all_factors
+        factors = load_all_factors(weights, file)
+
+        # 计算因子
+        computer = FactorComputer()
+        factor_data = computer.compute_all_factors(factors, {symbol: df}, {}, [symbol])
+
+        # 显示结果
+        click.echo("\nFactor Results:")
+        if factor_data:
+            for row in factor_data:
+                for name, value in row.items():
+                    if name != "symbol":
+                        click.echo(f"  {name}: {value:.4f}" if isinstance(value, (int, float)) else f"  {name}: {value}")
+
+        # 保存结果
+        if output and factor_data:
+            result_df = pd.DataFrame(factor_data)
+            result_df.to_csv(output, index=False)
+            click.echo(f"\nSaved to {output}")
+
+    except Exception as e:
+        click.echo(f"Error: {e}", err=True)
+        sys.exit(1)
+
+
+@factor.command("score")
+@click.option("--file", "-f", required=True, type=click.Path(exists=True), help="Factor config YAML file")
+@click.option("--symbols", default=None, help="Stock symbols (comma-separated)")
+@click.option("--start", type=date_type, default="2020-01-01", help="Start date")
+@click.option("--end", type=date_type, default=None, help="End date")
+@click.option("--top", type=int, default=None, help="Show top N results")
+@click.option("--output", type=click.Path(), default=None, help="Output file path")
+@click.pass_context
+def factor_score(ctx, file, symbols, start, end, top, output):
+    """根据 YAML 配置文件对股票进行评分选股"""
+    from .factors import load_strategy, FactorPipeline
+
+    if end is None:
+        end = today()
+
+    click.echo(f"Loading factor config from {file}...")
+
+    try:
+        config = load_strategy(file)
+        click.echo(f"Config: {config.name} (v{config.version})")
+
+        # 创建数据管理器
+        config_data = DataConfig()
+        dm = DataManager(config_data)
+
+        # 解析股票列表
+        if symbols:
+            symbol_list = [s.strip() for s in symbols.split(",")]
+        else:
+            symbol_list = ["600519"]
+
+        click.echo(f"Fetching data for {len(symbol_list)} symbols...")
+
+        # 获取多只股票数据
+        stock_data = {}
+        for symbol in symbol_list:
+            df = dm.get_daily(symbol, start, end)
+            if not df.empty:
+                stock_data[symbol] = df
+
+        if not stock_data:
+            click.echo("No data found for any symbol")
+            return
+
+        click.echo(f"Got data for {len(stock_data)} symbols")
+
+        # 使用 Pipeline 执行评分
+        pipeline = FactorPipeline(file)
+        results = pipeline.run(symbol_list, end, stock_data, pd.DataFrame(), limit=top)
+
+        if results.empty:
+            click.echo("No results")
+            return
+
+        # 显示结果
+        click.echo("\n=== Scoring Results ===")
+        click.echo(results.to_string(index=False))
+
+        # 保存结果
+        if output:
+            results.to_csv(output, index=False)
+            click.echo(f"\nSaved to {output}")
+
+    except Exception as e:
+        click.echo(f"Error: {e}", err=True)
+        import traceback
+        if ctx.obj.get("verbose"):
+            traceback.print_exc()
+        sys.exit(1)
+
+
+# =============================================================================
+# Filter 命令
+# =============================================================================
+
+@quantcli.group()
+def filter():
+    """多阶段因子筛选"""
+    pass
+
+
+@filter.command("run")
+@click.option("--file", "-f", required=True, type=click.Path(exists=True), help="Strategy config YAML file")
+@click.option("--symbols", default=None, help="Stock symbols (comma-separated, default: all)")
+@click.option("--start", type=date_type, default=None, help="Start date for price data (default: 60 days ago)")
+@click.option("--end", type=date_type, default=None, help="End date for price data")
+@click.option("--as-of", type=date_type, default=None, help="Time baseline date (for filtering as of this date)")
+@click.option("--fundamental-date", type=date_type, default=None, help="Date for fundamental data")
+@click.option("--top", type=int, default=None, help="Show top N results")
+@click.option("--output", type=click.Path(), default=None, help="Output file path")
+@click.option("--intraday/--no-intraday", default=True, help="Enable intraday data fetching")
+@click.pass_context
+def filter_run(ctx, file, symbols, start, end, as_of, fundamental_date, top, output, intraday):
+    """运行多阶段因子筛选（基本面筛选 + 日线筛选 + 权重排序）"""
+    from .factors.pipeline import FactorPipeline
+    from .datasources import create_datasource
+
+    # 设置时间基线
+    if as_of is not None:
+        TimeContext.set_date(as_of)
+        click.echo(f"Time baseline set to: {format_date(as_of)}")
+
+    if end is None:
+        end = today()
+    if start is None:
+        start = end - timedelta(days=60)  # 默认最近60天
+    if fundamental_date is None:
+        fundamental_date = end
+
+    click.echo(f"Loading strategy config from {file}...")
+
+    try:
+        pipeline = FactorPipeline(file)
+        click.echo(f"Strategy: {pipeline.config.name} (v{pipeline.config.version})")
+
+        # 检查是否为多阶段配置
+        screening_config = pipeline.config.screening
+        is_multi_stage = bool(screening_config.get("fundamental_conditions") or
+                              screening_config.get("daily_conditions"))
+        intraday_config = getattr(pipeline.config, 'intraday', {})
+        has_intraday = bool(intraday_config.get("weights", {}))
+
+        # 解析股票列表
+        if symbols:
+            symbol_list = [s.strip() for s in symbols.split(",")]
+        else:
+            # 获取全部股票
+            click.echo("Fetching all stock symbols...")
+            source = create_datasource("mixed")
+            stocks_df = source.get_stock_list()
+            symbol_list = stocks_df["symbol"].tolist()
+            click.echo(f"Found {len(symbol_list)} stocks")
+
+        if not symbol_list:
+            click.echo("No symbols to filter")
+            return
+
+        # ==================== 多阶段流程 ====================
+        if is_multi_stage and intraday and has_intraday:
+            # 完整多阶段流程：基本面 → 日线 → 分钟
+            click.echo("Running multi-stage filtering...")
+
+            # 阶段1: 基本面筛选
+            click.echo(f"\n=== Stage 1: Fundamental Screening ===")
+            click.echo(f"Fetching fundamental data for {len(symbol_list)} symbols...")
+            source = create_datasource("mixed")
+            fundamental_data = source.get_fundamental(symbol_list, fundamental_date)
+            click.echo(f"Got fundamental data for {len(fundamental_data)} symbols")
+
+            # 基本面条件筛选
+            fundamental_conditions = screening_config.get("fundamental_conditions", [])
+            candidates = pipeline.screening_only(symbol_list, fundamental_data) if fundamental_conditions else symbol_list
+            if fundamental_conditions:
+                click.echo(f"After fundamental screening: {len(candidates)}")
+
+            if not candidates:
+                click.echo("No candidates after fundamental screening")
+                return
+
+            # 阶段2: 日线筛选
+            click.echo(f"\n=== Stage 2: Daily Screening ===")
+            daily_conditions = screening_config.get("daily_conditions", [])
+            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"))
+
+                price_data = {}
+                for symbol in candidates:
+                    try:
+                        df = dm.get_daily(symbol, start, end)
+                        if not df.empty:
+                            df = df.copy()
+                            df["returns"] = df["close"].pct_change()
+                            price_data[symbol] = df
+                    except Exception as e:
+                        logger.warning(f"Failed to get daily data for {symbol}: {e}")
+                        continue
+
+                click.echo(f"Got daily data for {len(price_data)} symbols")
+
+                # 执行日线条件筛选
+                if price_data and daily_conditions:
+                    # 收集价格数据
+                    candidate_price_data = []
+                    for symbol in candidates:
+                        if symbol in price_data:
+                            df = price_data[symbol].copy()
+                            df["symbol"] = symbol
+                            candidate_price_data.append(df)
+
+                    if candidate_price_data:
+                        price_df = pd.concat(candidate_price_data, ignore_index=True)
+                        latest_date = price_df["date"].max()
+                        latest_df = price_df[price_df["date"] == latest_date].copy()
+
+                        passed = pipeline._evaluate_screening(daily_conditions, latest_df)
+                        candidates = [
+                            s for s in candidates
+                            if s in latest_df["symbol"].values and
+                               passed.get(latest_df[latest_df["symbol"] == s].index[0], True)
+                        ]
+                        click.echo(f"After daily screening: {len(candidates)}")
+            else:
+                price_data = {}
+
+            if not candidates:
+                click.echo("No candidates after daily screening")
+                return
+
+            # 阶段3: 分钟级排名
+            click.echo(f"\n=== Stage 3: Intraday Ranking ===")
+            click.echo(f"Fetching intraday data for {len(candidates)} candidates...")
+            source = create_datasource("mixed")
+
+            intraday_data = {}
+            for symbol in candidates:
+                try:
+                    df = source.get_intraday(symbol, start, end)
+                    if not df.empty:
+                        intraday_data[symbol] = df
+                except Exception as e:
+                    logger.warning(f"Failed to get intraday data for {symbol}: {e}")
+                    continue
+
+            click.echo(f"Got intraday data for {len(intraday_data)} symbols")
+
+            # 执行分钟级排名
+            results = pipeline.run_multi_stage(
+                symbols=candidates,
+                date=fundamental_date,
+                price_data=price_data,
+                intraday_data=intraday_data,
+                fundamental_data=fundamental_data,
+                limit=top
+            )
+
+        elif is_multi_stage:
+            # 简化的多阶段：基本面 → 日线
+            click.echo("Running simplified multi-stage filtering...")
+
+            # 阶段1: 基本面
+            click.echo(f"\n=== Stage 1: Fundamental Screening ===")
+            click.echo(f"Fetching fundamental data for {len(symbol_list)} symbols...")
+            source = create_datasource("mixed")
+            fundamental_data = source.get_fundamental(symbol_list, fundamental_date)
+            click.echo(f"Got fundamental data for {len(fundamental_data)} symbols")
+
+            candidates = pipeline.screening_only(symbol_list, fundamental_data)
+            click.echo(f"After fundamental screening: {len(candidates)}")
+
+            if not candidates:
+                click.echo("No candidates after fundamental screening")
+                return
+
+            # 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"))
+
+            price_data = {}
+            for symbol in candidates:
+                try:
+                    df = dm.get_daily(symbol, start, end)
+                    if not df.empty:
+                        df = df.copy()
+                        df["returns"] = df["close"].pct_change()
+                        price_data[symbol] = df
+                except Exception as e:
+                    logger.warning(f"Failed to get daily data for {symbol}: {e}")
+                    continue
+
+            click.echo(f"Got daily data for {len(price_data)} symbols")
+
+            # 获取分钟数据（用于混合 ranking）
+            click.echo(f"Fetching intraday data for {len(candidates)} candidates...")
+            intraday_data = {}
+            for symbol in candidates:
+                try:
+                    df = source.get_intraday(symbol, start, end)
+                    if not df.empty:
+                        intraday_data[symbol] = df
+                except Exception as e:
+                    logger.warning(f"Failed to get intraday data for {symbol}: {e}")
+                    continue
+
+            click.echo(f"Got intraday data for {len(intraday_data)} symbols")
+
+            results = pipeline.run_multi_stage(
+                symbols=candidates,
+                date=fundamental_date,
+                price_data=price_data,
+                intraday_data=intraday_data,
+                fundamental_data=fundamental_data,
+                limit=top
+            )
+
+        else:
+            # 原有单阶段流程
+            click.echo("Running single-stage filtering...")
+
+            click.echo(f"Fetching fundamental data for {len(symbol_list)} symbols...")
+            source = create_datasource("mixed")
+            fundamental_data = source.get_fundamental(symbol_list, fundamental_date)
+            click.echo(f"Got fundamental data for {len(fundamental_data)} symbols")
+
+            candidates = pipeline.screening_only(symbol_list, fundamental_data)
+            click.echo(f"Candidates after screening: {len(candidates)}")
+
+            if not candidates:
+                click.echo("No candidates after screening")
+                return
+
+            click.echo(f"Fetching price data for {len(candidates)} candidates...")
+            dm = DataManager(DataConfig(source="mixed"))
+
+            price_data = {}
+            for symbol in candidates:
+                try:
+                    df = dm.get_daily(symbol, start, end)
+                    if not df.empty:
+                        df = df.copy()
+                        df["returns"] = df["close"].pct_change()
+                        price_data[symbol] = df
+                except Exception as e:
+                    logger.warning(f"Failed to get data for {symbol}: {e}")
+                    continue
+
+            click.echo(f"Got price data for {len(price_data)} symbols")
+
+            results = pipeline.run(
+                symbols=candidates,
+                date=fundamental_date,
+                price_data=price_data,
+                fundamental_data=fundamental_data,
+                limit=top
+            )
+
+        if results.empty:
+            click.echo("No results")
+            return
+
+        # 显示结果
+        click.echo("\n=== Filter Results ===")
+        click.echo(results.to_string(index=False))
+
+        # 保存结果
+        if output:
+            results.to_csv(output, index=False)
+            click.echo(f"\nSaved to {output}")
+
+    except Exception as e:
+        click.echo(f"Error: {e}", err=True)
+        import traceback
+        if ctx.obj.get("verbose"):
+            traceback.print_exc()
+        sys.exit(1)
+    finally:
+        TimeContext.reset()
+
+
+@filter.command("screening")
+@click.option("--file", "-f", required=True, type=click.Path(exists=True), help="Strategy config YAML file")
+@click.option("--symbols", default=None, help="Stock symbols (comma-separated)")
+@click.option("--fundamental-date", type=date_type, default=None, help="Date for fundamental data")
+@click.option("--output", type=click.Path(), default=None, help="Output file path")
+@click.pass_context
+def filter_screening(ctx, file, symbols, fundamental_date, output):
+    """仅执行筛选阶段"""
+    from .factors.pipeline import FactorPipeline
+    from .datasources import create_datasource
+
+    if fundamental_date is None:
+        fundamental_date = today()
+
+    click.echo(f"Loading strategy config from {file}...")
+
+    try:
+        pipeline = FactorPipeline(file)
+        click.echo(f"Strategy: {pipeline.config.name}")
+
+        # 解析股票列表
+        if symbols:
+            symbol_list = [s.strip() for s in symbols.split(",")]
+        else:
+            source = create_datasource("mixed")
+            stocks_df = source.get_stock_list()
+            symbol_list = stocks_df["symbol"].tolist()
+
+        # 获取基本面数据
+        source = create_datasource("mixed")
+        fundamental_data = source.get_fundamental(symbol_list, fundamental_date)
+
+        # 执行筛选
+        candidates = pipeline.screening_only(symbol_list, fundamental_data)
+
+        click.echo(f"\nScreening Results: {len(candidates)} stocks passed")
+        if candidates:
+            click.echo(", ".join(candidates[:50]))
+            if len(candidates) > 50:
+                click.echo(f"... and {len(candidates) - 50} more")
+
+        # 保存结果
+        if output:
+            with open(output, "w") as f:
+                for s in candidates:
+                    f.write(f"{s}\n")
+            click.echo(f"\nSaved to {output}")
+
+    except Exception as e:
+        click.echo(f"Error: {e}", err=True)
+        sys.exit(1)
+
+
+# =============================================================================
+# Backtest 命令
+# =============================================================================
+
+@quantcli.group()
+def backtest():
+    """回测引擎"""
+    pass
+
+
+@backtest.command("run")
+@click.option("--strategy", "-s", required=True, type=str, help="Strategy file or class name")
+@click.option("--symbol", default="600519", help="Stock symbol")
+@click.option("--start", type=date_type, default="2020-01-01", help="Start date")
+@click.option("--end", type=date_type, default=None, help="End date")
+@click.option("--as-of", type=date_type, default=None, help="Time baseline date (for backtesting as of this date)")
+@click.option("--capital", type=float, default=1000000.0, help="Initial capital")
+@click.option("--fee", type=float, default=0.0003, help="Transaction fee rate")
+@click.pass_context
+def backtest_run(ctx, strategy, symbol, start, end, as_of, capital, fee):
+    """运行回测"""
+    # 设置时间基线
+    if as_of is not None:
+        TimeContext.set_date(as_of)
+        click.echo(f"Time baseline set to: {format_date(as_of)}")
+
+    if end is None:
+        end = today()
+
+    click.echo(f"Running backtest for {symbol}...")
+
+    # 加载策略
+    try:
+        strategy_cls = load_strategy(strategy)
+    except Exception as e:
+        click.echo(f"Error loading strategy: {e}", err=True)
+        sys.exit(1)
+
+    # 获取数据
+    config = DataConfig()
+    dm = DataManager(config)
+    df = dm.get_daily(symbol, start, end)
+
+    if df.empty:
+        click.echo(f"No data for {symbol}")
+        return
+
+    click.echo(f"Data: {len(df)} rows")
+
+    # 配置回测
+    bt_config = BacktestConfig(
+        initial_capital=capital,
+        fee=fee,
+        start_date=start,
+        end_date=end
+    )
+    engine = BacktestEngine(bt_config)
+    engine.add_data(symbol, df)
+
+    # 运行回测
+    try:
+        result = engine.run(strategy_cls)
+
+        click.echo(f"\n=== Backtest Results ===")
+        click.echo(f"Total Return: {result.total_return:.2%}")
+        click.echo(f"Annual Return: {result.annual_return:.2%}")
+        click.echo(f"Max Drawdown: {result.max_drawdown:.2%}")
+        click.echo(f"Sharpe Ratio: {result.sharpe:.2f}")
+        click.echo(f"Sortino Ratio: {result.sortino:.2f}")
+        click.echo(f"Win Rate: {result.win_rate:.2%}")
+        click.echo(f"Total Trades: {result.total_trades}")
+
+    except Exception as e:
+        click.echo(f"Error running backtest: {e}", err=True)
+        sys.exit(1)
+    finally:
+        TimeContext.reset()
+
+
+@backtest.command("list")
+@click.pass_context
+def backtest_list(ctx):
+    """列出历史回测结果"""
+    click.echo("Historical backtests:")
+    click.echo("(Not implemented yet)")
+
+
+# =============================================================================
+# Config 命令
+# =============================================================================
+
+@quantcli.group()
+def config():
+    """配置管理"""
+    pass
+
+
+@config.command("show")
+@click.pass_context
+def config_show(ctx):
+    """显示当前配置"""
+    config = DataConfig()
+    click.echo("QuantCLI Configuration:")
+    click.echo(f"  data.source: {config.source}")
+    click.echo(f"  data.cache_dir: {config.cache_dir}")
+    click.echo(f"  data.parallel: {config.parallel}")
+    click.echo(f"  data.fillna: {config.fillna}")
+    click.echo(f"  data.outlier_method: {config.outlier_method}")
+    click.echo(f"  data.outlier_threshold: {config.outlier_threshold}")
+
+
+@config.command("set")
+@click.argument("key", type=str)
+@click.argument("value", type=str)
+@click.pass_context
+def config_set(ctx, key, value):
+    """设置配置项"""
+    click.echo(f"Setting {key} = {value}")
+    # TODO: 实现配置持久化
+    click.echo("(Configuration persistence not implemented yet)")
+
+
+# =============================================================================
+# 辅助函数
+# =============================================================================
+
+def load_strategy(strategy_spec: str):
+    """加载策略类
+
+    Args:
+        strategy_spec: 策略文件路径或类名
+
+    Returns:
+        Strategy 类
+    """
+    from pathlib import Path
+
+    # 尝试作为文件加载
+    strategy_path = Path(strategy_spec)
+    if strategy_path.exists() and strategy_path.suffix == ".py":
+        import importlib.util
+        spec = importlib.util.spec_from_file_location("strategy", strategy_path)
+        module = importlib.util.module_from_spec(spec)
+        spec.loader.exec_module(module)
+
+        # 查找 Strategy 类
+        for attr_name in dir(module):
+            attr = getattr(module, attr_name)
+            if isinstance(attr, type) and issubclass(attr, type) and attr.__name__ != "Strategy":
+                # 找到继承自 object 的策略类 (非 Strategy 基类)
+                if hasattr(attr, "__bases__") and Strategy in attr.__mro__[1:]:
+                    return attr
+
+        raise ValueError(f"No Strategy class found in {strategy_spec}")
+
+    # 尝试作为内置策略名加载
+    built_in_strategies = {
+        "ma_cross": MaCrossStrategy,
+    }
+
+    if strategy_spec in built_in_strategies:
+        return built_in_strategies[strategy_spec]
+
+    raise ValueError(f"Unknown strategy: {strategy_spec}")
+
+
+class MaCrossStrategy(Strategy):
+    """简单均线交叉策略 (内置示例)"""
+
+    name = "MA Cross"
+    params = {"fast": 5, "slow": 20}
+
+    def __init__(self):
+        super().__init__()
+        import backtrader as bt
+        self.ma_fast = bt.indicators.SMA(self.data.close, period=self.params.fast)
+        self.ma_slow = bt.indicators.SMA(self.data.close, period=self.params.slow)
+
+    def next(self):
+        if self.ma_fast[0] > self.ma_slow[0] and self.ma_fast[-1] <= self.ma_slow[-1]:
+            self.buy()
+        elif self.ma_fast[0] < self.ma_slow[0] and self.ma_fast[-1] >= self.ma_slow[-1]:
+            self.sell()
+
+
+# =============================================================================
+# 入口点
+# =============================================================================
+
+def main():
+    """CLI 入口点"""
+    quantcli(obj={})
+
+
+if __name__ == "__main__":
+    main()