oscura 0.0.1__py3-none-any.whl → 0.1.1__py3-none-any.whl

oscura/batch/aggregate.py

@@ -0,0 +1,300 @@
+"""Result aggregation for batch analysis.
+
+
+This module provides statistical aggregation and reporting for batch
+analysis results, including outlier detection and export capabilities.
+"""
+
+from pathlib import Path
+from typing import Any
+
+import numpy as np
+import pandas as pd
+
+
+def aggregate_results(
+    results: pd.DataFrame,
+    *,
+    metrics: list[str] | None = None,
+    outlier_threshold: float = 3.0,
+    include_plots: bool = False,
+    output_format: str = "dict",
+    output_file: str | Path | None = None,
+) -> dict[str, Any] | pd.DataFrame:
+    """Aggregate results from batch analysis into summary statistics.
+
+    : Computes comprehensive statistics (mean, std, min, max,
+    outliers) for each metric in the batch results. Supports export to various
+    formats and optional visualization generation.
+
+    Args:
+        results: DataFrame from batch_analyze() containing analysis results
+        metrics: List of column names to aggregate (default: all numeric columns)
+        outlier_threshold: Z-score threshold for outlier detection (default: 3.0)
+        include_plots: Generate comparison plots across files (default: False)
+        output_format: Output format - 'dict', 'dataframe', 'csv', 'excel', 'html'
+        output_file: Optional output file path for export formats
+
+    Returns:
+        Dictionary or DataFrame with summary statistics:
+        - count: Number of valid values
+        - mean: Mean value
+        - std: Standard deviation
+        - min: Minimum value
+        - max: Maximum value
+        - median: Median value
+        - q25: 25th percentile
+        - q75: 75th percentile
+        - outliers: List of outlier values
+        - outlier_files: List of files containing outliers
+
+    Raises:
+        ValueError: If no numeric metrics are found in results.
+
+    Examples:
+        >>> results = osc.batch_analyze(files, osc.characterize_buffer)
+        >>> summary = osc.aggregate_results(
+        ...     results,
+        ...     metrics=['rise_time', 'fall_time'],
+        ...     outlier_threshold=2.5
+        ... )
+        >>> print(summary['rise_time']['mean'])
+        >>> print(summary['rise_time']['outlier_files'])
+
+    Notes:
+        - Outliers detected using IQR method: values outside [Q1 - k*IQR, Q3 + k*IQR]
+          where k = (threshold / 3.0) * 1.5 (more robust than z-score for heavy-tailed data)
+        - Non-numeric columns are automatically skipped
+        - Missing values (NaN) are excluded from statistics
+        - CSV/Excel/HTML export requires output_file parameter
+
+    References:
+        BATCH-002: Result Aggregation
+    """
+    if results.empty:
+        return {} if output_format == "dict" else pd.DataFrame()
+
+    # Determine metrics to analyze
+    if metrics is None:
+        # Auto-select all numeric columns except 'file' and 'error'
+        metrics = results.select_dtypes(include=[np.number]).columns.tolist()
+        metrics = [m for m in metrics if m not in ["file", "error"]]
+
+    if not metrics:
+        raise ValueError("No numeric metrics found in results")
+
+    # Compute aggregated statistics
+    aggregated: dict[str, dict[str, Any]] = {}
+
+    for metric in metrics:
+        if metric not in results.columns:
+            continue
+
+        # Extract valid (non-null) values
+        values = results[metric].dropna()
+
+        if values.empty:
+            aggregated[metric] = {
+                "count": 0,
+                "mean": np.nan,
+                "std": np.nan,
+                "min": np.nan,
+                "max": np.nan,
+                "median": np.nan,
+                "q25": np.nan,
+                "q75": np.nan,
+                "outliers": [],
+                "outlier_files": [],
+            }
+            continue
+
+        # Basic statistics
+        stats = {
+            "count": len(values),
+            "mean": float(values.mean()),
+            "std": float(values.std()),
+            "min": float(values.min()),
+            "max": float(values.max()),
+            "median": float(values.median()),
+            "q25": float(values.quantile(0.25)),
+            "q75": float(values.quantile(0.75)),
+        }
+
+        # Outlier detection using IQR method (more robust than z-score)
+        # IQR method: outliers are values outside [Q1 - k*IQR, Q3 + k*IQR]
+        # where k = outlier_threshold * 1.5 (standard is k=1.5, we scale by threshold)
+        if len(values) > 3:  # Need at least 4 values for meaningful IQR
+            q1 = stats["q25"]
+            q3 = stats["q75"]
+            iqr = q3 - q1
+
+            # Scale IQR multiplier by threshold (default 3.0 -> 2.0 * 1.5 = 3.0)
+            k = (outlier_threshold / 3.0) * 1.5
+
+            lower_bound = q1 - k * iqr
+            upper_bound = q3 + k * iqr
+
+            outlier_mask = (values < lower_bound) | (values > upper_bound)
+            outlier_indices = values[outlier_mask].index.tolist()
+            stats["outliers"] = values[outlier_mask].tolist()
+
+            # Get corresponding filenames if available
+            if "file" in results.columns:
+                stats["outlier_files"] = results.loc[outlier_indices, "file"].tolist()
+            else:
+                stats["outlier_files"] = outlier_indices
+        else:
+            stats["outliers"] = []  # type: ignore[assignment]
+            stats["outlier_files"] = []  # type: ignore[assignment]
+
+        aggregated[metric] = stats
+
+    # Generate plots if requested
+    if include_plots:
+        # Import here to avoid circular dependency
+        try:
+            import matplotlib.pyplot as plt
+
+            for metric in metrics:
+                if metric not in aggregated:
+                    continue
+
+                _fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(12, 4))
+
+                # Histogram
+                results[metric].dropna().hist(ax=ax1, bins=30)
+                ax1.axvline(
+                    aggregated[metric]["mean"],
+                    color="r",
+                    linestyle="--",
+                    label="Mean",
+                )
+                ax1.axvline(
+                    aggregated[metric]["median"],
+                    color="g",
+                    linestyle="--",
+                    label="Median",
+                )
+                ax1.set_xlabel(metric)
+                ax1.set_ylabel("Count")
+                ax1.legend()
+                ax1.set_title(f"{metric} Distribution")
+
+                # Box plot
+                ax2.boxplot(results[metric].dropna())
+                ax2.set_ylabel(metric)
+                ax2.set_title(f"{metric} Box Plot")
+
+                plt.tight_layout()
+
+                # Save or show based on output_file
+                if output_file:
+                    plot_file = Path(output_file).with_suffix("") / f"{metric}_plot.png"
+                    plot_file.parent.mkdir(parents=True, exist_ok=True)
+                    plt.savefig(plot_file)
+                else:
+                    plt.show()
+
+                plt.close()
+
+        except ImportError:
+            pass  # Silently skip plotting if matplotlib not available
+
+    # Format output
+    if output_format == "dict":
+        return aggregated
+
+    elif output_format == "dataframe":
+        # Convert to DataFrame with metrics as rows
+        df = pd.DataFrame(aggregated).T
+        # Drop list columns for DataFrame format
+        df = df.drop(columns=["outliers", "outlier_files"], errors="ignore")
+        return df
+
+    elif output_format in ["csv", "excel", "html"]:
+        if not output_file:
+            raise ValueError(f"{output_format} format requires output_file parameter")
+
+        df = pd.DataFrame(aggregated).T
+        df = df.drop(columns=["outliers", "outlier_files"], errors="ignore")
+
+        if output_format == "csv":
+            df.to_csv(output_file)
+        elif output_format == "excel":
+            df.to_excel(output_file)
+        elif output_format == "html":
+            # Generate HTML report
+            html = _generate_html_report(results, aggregated, metrics)
+            Path(output_file).write_text(html)
+
+        return df
+
+    else:
+        raise ValueError(f"Unknown output_format: {output_format}")
+
+
+def _generate_html_report(
+    results: pd.DataFrame,
+    aggregated: dict[str, dict[str, Any]],
+    metrics: list[str],
+) -> str:
+    """Generate HTML report for batch analysis results."""
+    html = """
+    <!DOCTYPE html>
+    <html>
+    <head>
+        <title>Batch Analysis Report</title>
+        <style>
+            body { font-family: Arial, sans-serif; margin: 20px; }
+            h1 { color: #333; }
+            h2 { color: #666; margin-top: 30px; }
+            table { border-collapse: collapse; width: 100%; margin: 20px 0; }
+            th, td { border: 1px solid #ddd; padding: 8px; text-align: left; }
+            th { background-color: #4CAF50; color: white; }
+            tr:nth-child(even) { background-color: #f2f2f2; }
+            .outlier { background-color: #ffcccc; }
+        </style>
+    </head>
+    <body>
+        <h1>Batch Analysis Report</h1>
+    """
+    # Summary statistics table
+    html += "<h2>Summary Statistics</h2>\n<table>\n"
+    html += "<tr><th>Metric</th><th>Count</th><th>Mean</th><th>Std</th>"
+    html += "<th>Min</th><th>Median</th><th>Max</th><th>Outliers</th></tr>\n"
+
+    for metric in metrics:
+        if metric not in aggregated:
+            continue
+        stats = aggregated[metric]
+        html += "<tr>"
+        html += f"<td>{metric}</td>"
+        html += f"<td>{stats['count']}</td>"
+        html += f"<td>{stats['mean']:.4g}</td>"
+        html += f"<td>{stats['std']:.4g}</td>"
+        html += f"<td>{stats['min']:.4g}</td>"
+        html += f"<td>{stats['median']:.4g}</td>"
+        html += f"<td>{stats['max']:.4g}</td>"
+        html += f"<td>{len(stats['outliers'])}</td>"
+        html += "</tr>\n"
+
+    html += "</table>\n"
+
+    # Outlier details
+    has_outliers = any(len(aggregated[m]["outliers"]) > 0 for m in metrics if m in aggregated)
+
+    if has_outliers:
+        html += "<h2>Outliers Detected</h2>\n"
+        for metric in metrics:
+            if metric not in aggregated:
+                continue
+            stats = aggregated[metric]
+            if stats["outliers"]:
+                html += f"<h3>{metric}</h3>\n<table>\n"
+                html += "<tr><th>File</th><th>Value</th></tr>\n"
+                for file, value in zip(stats["outlier_files"], stats["outliers"], strict=False):
+                    html += f"<tr class='outlier'><td>{file}</td><td>{value:.4g}</td></tr>\n"
+                html += "</table>\n"
+
+    html += "</body>\n</html>"
+    return html

oscura/batch/analyze.py

@@ -0,0 +1,139 @@
+"""Multi-file batch analysis with parallel execution support.
+
+
+This module provides parallel batch processing of signal files using
+concurrent.futures for efficient multi-core utilization.
+"""
+
+from collections.abc import Callable
+from concurrent.futures import ProcessPoolExecutor, ThreadPoolExecutor, as_completed
+from pathlib import Path
+from typing import Any
+
+import pandas as pd
+
+
+def batch_analyze(
+    files: list[str | Path],
+    analysis_fn: Callable[[str | Path], dict[str, Any]],
+    *,
+    parallel: bool = False,
+    workers: int | None = None,
+    progress_callback: Callable[[int, int, str], None] | None = None,
+    use_threads: bool = False,
+    **config: Any,
+) -> pd.DataFrame:
+    """Analyze multiple files with the same analysis configuration.
+
+    : Multi-file analysis with parallel execution support
+    via concurrent.futures. Returns aggregated results as a DataFrame for
+    easy statistical analysis and export.
+
+    Args:
+        files: List of file paths to analyze
+        analysis_fn: Analysis function to apply to each file.
+            Must accept a file path and return a dict of results.
+        parallel: Enable parallel processing (default: False)
+        workers: Number of parallel workers (default: CPU count)
+        progress_callback: Optional callback for progress updates.
+            Called with (current, total, filename) after each file.
+        use_threads: Use ThreadPoolExecutor instead of ProcessPoolExecutor
+            (useful for I/O-bound tasks, default: False)
+        **config: Additional keyword arguments passed to analysis_fn
+
+    Returns:
+        DataFrame with one row per file, columns from analysis results.
+        Always includes a 'file' column with the input filename.
+
+    Examples:
+        >>> import oscura as osc
+        >>> import glob
+        >>> files = glob.glob('captures/*.wfm')
+        >>> results = osc.batch_analyze(
+        ...     files,
+        ...     analysis_fn=osc.characterize_buffer,
+        ...     parallel=True,
+        ...     workers=4
+        ... )
+        >>> print(results[['file', 'rise_time', 'fall_time', 'status']])
+        >>> results.to_csv('batch_results.csv')
+
+    Notes:
+        - Use parallel=True for CPU-bound analysis functions
+        - Use use_threads=True for I/O-bound operations (file loading)
+        - Progress callback is called from worker threads/processes
+        - All exceptions during analysis are caught and stored in 'error' column
+
+    References:
+        BATCH-001: Multi-File Analysis
+    """
+    if not files:
+        return pd.DataFrame()
+
+    # Wrapper to include config in analysis calls
+    def _wrapped_analysis(filepath: str | Path) -> dict[str, Any]:
+        try:
+            result = analysis_fn(filepath, **config)
+            # Ensure result is a dict
+            if not isinstance(result, dict):
+                result = {"result": result}  # type: ignore[unreachable]
+            result["file"] = str(filepath)
+            result["error"] = None
+            return result
+        except Exception as e:
+            # Return error info on failure
+            return {
+                "file": str(filepath),
+                "error": str(e),
+            }
+
+    results: list[dict[str, Any]] = []
+    total = len(files)
+
+    if parallel:
+        # Use concurrent.futures for parallel execution
+        executor_class = ThreadPoolExecutor if use_threads else ProcessPoolExecutor
+        with executor_class(max_workers=workers) as executor:
+            # Submit all tasks
+            future_to_file = {executor.submit(_wrapped_analysis, f): f for f in files}
+
+            # Process results as they complete
+            for i, future in enumerate(as_completed(future_to_file), 1):
+                filepath = future_to_file[future]
+                try:
+                    result = future.result()
+                    results.append(result)
+
+                    if progress_callback:
+                        progress_callback(i, total, str(filepath))
+                except Exception as e:
+                    # Catch execution errors
+                    results.append(
+                        {
+                            "file": str(filepath),
+                            "error": f"Execution error: {e}",
+                        }
+                    )
+
+    else:
+        # Sequential processing
+        for i, filepath in enumerate(files, 1):
+            result = _wrapped_analysis(filepath)
+            results.append(result)
+
+            if progress_callback:
+                progress_callback(i, total, str(filepath))
+
+    # Convert to DataFrame
+    df = pd.DataFrame(results)
+
+    # Reorder columns: file first, error last
+    cols = df.columns.tolist()
+    if "file" in cols:
+        cols.remove("file")
+        cols = ["file", *cols]
+    if "error" in cols:
+        cols.remove("error")
+        cols = [*cols, "error"]
+
+    return df[cols]