Changelog ========= All notable changes to edaflow are documented here. The format is based on `Keep a Changelog `_, and this project adheres to `Semantic Versioning `_. Version 0.12.33 (2025-01-11) - Major API Improvement 🚀 ------------------------------------------------------- **Added:** - **🚀 NEW CLEAN APIs**: Introduced ``apply_encoding()`` and ``apply_encoding_with_encoders()`` functions for consistent, predictable behavior - ``apply_encoding(df)`` - Always returns DataFrame (recommended for most users) - ``apply_encoding_with_encoders(df)`` - Always returns (DataFrame, encoders) tuple - Clear, explicit function names that indicate exactly what they return - Comprehensive documentation with usage examples **Fixed:** - **🐛 ROOT CAUSE RESOLVED**: Eliminated confusion from ``apply_smart_encoding()`` inconsistent return types - Previous issue: Function returned DataFrame OR tuple based on ``return_encoders`` parameter - Enhanced error messages with helpful guidance for wrong data types - Robust detection and handling of tuple inputs in visualization functions **Deprecated:** - **⚠️ DEPRECATION WARNING**: ``apply_smart_encoding()`` with ``return_encoders=True`` now shows deprecation warning - Existing code continues working with guidance toward better alternatives - Clear migration path to new consistent functions **Usage Examples:** .. code-block:: python # ✅ NEW RECOMMENDED - Always returns DataFrame df_encoded = edaflow.apply_encoding(df) # ✅ NEW EXPLICIT - Always returns tuple df_encoded, encoders = edaflow.apply_encoding_with_encoders(df) # ⚠️ DEPRECATED - Inconsistent return type (still works) df_encoded = edaflow.apply_smart_encoding(df, return_encoders=True) # tuple! Version 0.12.32 (2025-08-11) - Critical Input Validation Fix 🐛 --------------------------------------------------------------- **Fixed:** - **🐛 CRITICAL INPUT FIX**: Fixed AttributeError: 'tuple' object has no attribute 'empty' in visualization functions - Root cause: Users passing tuple result from ``apply_smart_encoding(..., return_encoders=True)`` directly to visualization functions - Enhanced input validation with helpful error messages for common usage mistakes - Better error handling in ``visualize_scatter_matrix`` and other visualization functions - Clear documentation showing correct vs incorrect usage patterns - Prevents crashes in step 14 of EDA workflows when encoding functions are misused **Technical Details:** - **Smart Error Detection**: Automatically detects when tuple is passed instead of DataFrame - **Helpful Error Messages**: Guides users to correct usage pattern with code examples - **Robust Input Validation**: Added comprehensive type checking for all visualization functions - **Workflow Stability**: Eliminates common crash point in automated EDA workflows **Usage Examples:** .. code-block:: python # ❌ WRONG - This causes AttributeError: df_encoded = edaflow.apply_smart_encoding(df, return_encoders=True) # Returns tuple! edaflow.visualize_scatter_matrix(df_encoded) # Crashes # ✅ CORRECT - Unpack the tuple: df_encoded, encoders = edaflow.apply_smart_encoding(df, return_encoders=True) edaflow.visualize_scatter_matrix(df_encoded) # Works perfectly! Version 0.12.31 (2025-01-05) - Critical KeyError Hotfix 🚨 ---------------------------------------------------------- **Fixed:** - **🚨 CRITICAL HOTFIX**: Fixed KeyError: 'type' in ``summarize_eda_insights()`` function during Google Colab usage - Enhanced exception handling when target analysis dictionary missing expected keys - Implemented safe dictionary access using ``.get()`` method to prevent KeyErrors - All existing functionality preserved - pure stability fix - Verified fix across all notebook platforms (Colab, JupyterLab, VS Code) **Technical Details:** - **Robust Error Handling**: Added comprehensive try-catch blocks for edge cases - **Safe Dictionary Access**: Uses ``.get()`` method instead of direct key access - **Platform Compatibility**: Tested and verified across all major notebook environments - **Zero Regression**: No functionality changes - purely stability improvements Version 0.12.30 (2025-01-05) - Universal Display Optimization Breakthrough 🎨 ------------------------------------------------------------------------------ **Added:** - **🎨 BREAKTHROUGH FEATURE**: Introduced ``optimize_display()`` function for universal notebook compatibility - Automatic detection of Google Colab, JupyterLab, VS Code Notebooks, Classic Jupyter - Dynamic CSS injection for perfect dark/light mode visibility across all platforms - Automatic matplotlib backend optimization for each notebook environment - Solves visibility issues in dark mode themes universally - Zero configuration required - automatically detects and optimizes for your platform - Works flawlessly across all major notebook platforms **Usage:** .. code-block:: python from edaflow import optimize_display optimize_display() # Automatically optimizes for your platform **Technical Details:** - **Smart Platform Detection**: Automatically identifies current notebook environment - **Universal CSS Injection**: Applies platform-specific styling for optimal visibility - **Backend Optimization**: Sets optimal matplotlib backend for each platform - **Dark Mode Excellence**: Perfect visibility in dark themes across all platforms - **Zero Dependencies**: Uses only standard library features for maximum compatibility Version 0.12.29 (2025-08-11) - Critical Bug Fix for Unhashable Types 🐛 ------------------------------------------------------------------------ **Fixed:** - **🐛 CRITICAL FIX**: Fixed TypeError in ``analyze_categorical_columns`` when processing columns with unhashable types (lists, dicts) - Added proper exception handling for ``nunique()`` and ``unique()`` operations on columns containing unhashable data types - Function now converts unhashable types to strings before processing unique value counts - Added comprehensive error handling to gracefully handle any processing errors - Added missing return statement to provide structured data output for programmatic use - Returns dictionary with keys: ``object_columns``, ``numeric_potential``, ``truly_categorical``, ``non_object_columns`` **Technical Details:** - **Enhanced Compatibility**: Function now handles complex nested data structures without crashing - **Robust Processing**: Multiple fallback mechanisms ensure analysis completes successfully - **Better API**: Consistent return values enable programmatic access to analysis results Version 0.12.28 (2025-08-11) - Comprehensive Display Formatting Excellence 🎨 ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ **Added:** - **✨ NEW FUNCTION**: ``summarize_eda_insights()`` - Generate comprehensive EDA insights and recommendations after completing analysis workflow - Analyzes dataset characteristics, data quality, class distribution, and feature types - Provides intelligent recommendations for modeling and preprocessing - Supports target column analysis for classification/regression tasks - Tracks which edaflow functions have been used in the workflow - Returns structured dictionary with organized insights and actionable recommendations **Fixed:** - **🎨 ADDITIONAL DISPLAY FIXES**: Resolved display formatting issues in multiple core functions - **FIXED**: ``convert_to_numeric`` - Removed unnecessary separator lines and changed to SIMPLE box style - **FIXED**: ``display_column_types`` - Removed separator lines and improved table border rendering - **FIXED**: ``impute_numerical_median`` - Cleaned up display formatting and fixed box joining issues - **IMPROVED**: All affected functions now use consistent SIMPLE box style for clean border joining - **REMOVED**: Unnecessary "====" separators that cluttered the professional output **Technical Details:** - **Unified Styling**: All core functions now share consistent, professional formatting standards - **Perfect Border Joining**: SIMPLE box style ensures clean table border connections - **Visual Excellence**: Removed all visual clutter for optimal user experience - **Production Ready**: Professional output suitable for client presentations and reports Version 0.12.26 (2025-08-09) - Categorical Display Polish 📊 ------------------------------------------------------------ **Fixed:** - **🎨 CATEGORICAL DISPLAY FIX**: Resolved display formatting issues in ``analyze_categorical_columns`` function - **FIXED**: Removed unnecessary blue line beneath main title - **IMPROVED**: CATEGORICAL COLUMNS sub heading now has clean background (removed black background) - **ENHANCED**: All table box styles changed to SIMPLE for proper line joining - **FIXED**: NON-OBJECT COLUMNS sub heading contrast improved (removed dark background) - **IMPROVED**: Column Type Analysis panel now uses SIMPLE box for clean borders - **REMOVED**: Unnecessary line under "Analysis complete" message Version 0.12.25 (2025-08-08) - Missing Data Display Enhancement 🎨 ------------------------------------------------------------------- **Fixed:** - **🎨 DISPLAY FORMATTING FIX**: Resolved unnecessary separator lines in ``check_null_columns`` function - **FIXED**: Removed redundant "====" separators above and below MISSING DATA ANALYSIS banner - **IMPROVED**: Table border rendering now uses SIMPLE box style for clean line joining - **ENHANCED**: Professional output formatting without visual clutter Version 0.12.24 (2025-08-08) - Texture Analysis Warning Fix 🔧 --------------------------------------------------------------- **Fixed:** - **🔧 TEXTURE ANALYSIS WARNING FIX**: Resolved scikit-image UserWarning in ``analyze_image_features`` function - **FIXED**: Local Binary Pattern (LBP) analysis now properly converts images to uint8 format - **RESOLVED**: "floating-point images may give unexpected results" warning from texture analysis - **ENHANCED**: Improved image preprocessing to handle both normalized [0,1] and [0,255] input images - **OPTIMIZED**: More robust texture feature extraction with proper data type handling **Technical Details:** - **Smart Data Type Detection**: Automatically detects normalized vs standard image formats - **Optimal Performance**: LBP analysis now uses recommended integer format for better results - **Professional Output**: Clean execution without warnings in production environments - **Backward Compatible**: All existing code continues to work unchanged Version 0.12.23 (2025-08-08) - Critical RTD Documentation Parameter Fix 🚨 --------------------------------------------------------------------------- **Fixed:** - **🚨 CRITICAL RTD DOCUMENTATION FIX**: Corrected parameter name mismatches in ``analyze_image_features`` function - **FIXED**: Changed ``analyze_colors`` → ``analyze_color`` in quickstart.rst documentation (3 instances) - **FIXED**: Changed ``bins`` → ``bins_per_channel`` in RTD documentation examples - **RESOLVED**: TypeError when users followed RTD documentation examples exactly - **ENHANCED**: Documentation now matches actual function signature perfectly - **TESTED**: Created comprehensive test suite to prevent future parameter mismatches **Impact:** - **User Experience**: Eliminated TypeError when following documentation examples - **Documentation Quality**: RTD examples now work correctly out-of-the-box - **Professional Standards**: Maintained edaflow's reputation for accurate documentation Version 0.12.22 (2025-08-08) - Google Colab Compatibility & Clean Workflow 🌟 ------------------------------------------------------------------------------ **Fixed:** - **🔧 GOOGLE COLAB COMPATIBILITY**: Fixed KeyError in ``apply_smart_encoding`` documentation examples - **FIXED**: Removed hardcoded 'target' column assumption in documentation examples - **FIXED**: Updated quickstart.rst and README.md with flexible column handling - **RESOLVED**: Documentation examples now work in Google Colab, Jupyter, and all environments - **ENHANCED**: More robust ML encoding workflow that adapts to user datasets **Enhanced:** - **📚 CLEAN WORKFLOW**: Removed redundant print statements from documentation examples - **IMPROVED**: Professional rich-styled output eliminates need for manual formatting - **MODERNIZED**: Documentation examples now showcase rich styling capabilities - **CREATED**: Google Colab compatibility test suite for validation **Impact:** - **Universal Compatibility**: Documentation works across all Python environments - **Modern Presentation**: Clean, professional output using rich styling - **Educational Value**: Enhanced learning experience for users across platforms Version 0.12.3 (2025-08-06) - Complete Positional Argument Compatibility Fix 🔧 -------------------------------------------------------------------------------- **CRITICAL BUG FIX:** **Fixed:** - **CRITICAL**: Resolved TypeError when calling ``visualize_image_classes(image_paths, ...)`` with positional arguments - **Positional Arguments**: Function now properly handles legacy positional argument usage from Jupyter notebooks - **Backward Compatibility**: Complete support for all three usage patterns: 1. ``visualize_image_classes(path, ...)`` - Positional (deprecated, shows warning) 2. ``visualize_image_classes(image_paths=path, ...)`` - Keyword deprecated (shows warning) 3. ``visualize_image_classes(data_source=path, ...)`` - Recommended (no warning) **Improved:** - **User Experience**: Clear deprecation warnings guide users toward recommended ``data_source=`` syntax - **Function Architecture**: Refactored to wrapper function pattern for robust argument handling - **Error Messages**: Enhanced error messages provide clear guidance for parameter usage - **Documentation**: Updated all examples to show modern ``data_source=`` syntax **Technical Details:** - **Implementation**: Split function into public wrapper and internal implementation - **Argument Handling**: Proper detection and mapping of positional arguments to correct parameters - **Warning System**: Contextual warnings for different deprecated usage patterns - **Testing**: Comprehensive test suite validates all backward compatibility scenarios **Notes:** - **Zero Breaking Changes**: All existing code continues to work unchanged - **Jupyter Notebook Fix**: Resolves the specific TypeError reported in Jupyter notebook usage - **Migration Path**: Users can migrate at their own pace with clear guidance Version 0.12.2 (2025-08-06) - Documentation Refresh Release 📚 --------------------------------------------------------------- **Documentation:** - **PyPI Description**: Refreshed PyPI package description with latest feature updates - **Changelog Display**: Fixed PyPI changelog display showing current version information - **Version Alignment**: Ensured all documentation reflects current v0.12.2 capabilities Version 0.12.1 (2025-08-05) - Enhanced Computer Vision EDA 🖼️ -------------------------------------------------------------- **Enhanced Functions:** - **visualize_image_classes()**: Now supports both directory paths and pandas DataFrames as data sources - **DataFrame Support**: Pass image metadata as DataFrame with 'image_path' and 'class' columns - **Directory Support**: Continues to support organized folder structure (folder = class) - **Flexible Input**: Unified interface for different dataset organizations - **Error Handling**: Clear error messages guide proper usage for both input types **Improvements:** - **Data Source Flexibility**: Switch between directory-based and metadata-based workflows - **DataFrame Integration**: Perfect for datasets with existing metadata and annotations - **Backward Compatibility**: Maintains all existing directory-based functionality - **User Experience**: Enhanced error messages and parameter validation Version 0.10.0 (2025-08-05) - Image Quality Assessment Release 🔍 ---------------------------------------------------------------- **Major New Feature: Comprehensive Image Quality Assessment** **NEW Functions:** - **assess_image_quality()**: Complete automated quality assessment for image datasets **Key Capabilities:** - **Corruption Detection**: Automatically identify unreadable or damaged image files - **Brightness Analysis**: Flag overly dark or bright images with statistical thresholds - **Contrast Assessment**: Detect low-contrast images that might impact training - **Blur Detection**: Use Laplacian variance to identify potentially blurry images - **Color Analysis**: Distinguish between grayscale and color images, detect mixed modes - **Dimension Consistency**: Find unusual aspect ratios and size outliers using statistical methods - **Artifact Detection**: Identify compression artifacts and unusual patterns - **Quality Scoring**: Statistical quality scoring system (0-100) for overall dataset health - **Automated Recommendations**: Actionable suggestions for dataset improvement - **Production Integration**: Quality gates with customizable thresholds for ML pipelines - **Scalable Analysis**: Sampling support for efficient processing of large datasets **Enhanced Capabilities:** - Expanded from 15 to 16 comprehensive EDA functions - Extended computer vision capabilities with production-ready quality assessment - Added scipy optimization for advanced blur detection algorithms - Comprehensive statistical analysis with detailed reporting - Educational focus on image quality standards and best practices **Perfect For:** - Medical and scientific imaging with strict quality requirements - Production ML pipelines with automated data validation - Research and development with dataset quality monitoring - Educational purposes for learning image quality assessment Version 0.9.0 (2025-08-05) - Computer Vision EDA Release 🖼️ ------------------------------------------------------------ **Added** ~~~~~~~~~ * **NEW**: ``visualize_image_classes()`` function for comprehensive image classification dataset analysis * **NEW**: Computer Vision EDA workflow support with class-wise sample visualization * **NEW**: Directory-based and DataFrame-based image dataset analysis capabilities * **NEW**: Automatic class distribution analysis with imbalance detection * **NEW**: Image quality assessment with corrupted image detection * **NEW**: Statistical insights for image datasets (balance ratios, sample counts, warnings) * **NEW**: Professional grid layouts for image sample visualization with smart sizing * **NEW**: Technical image information display (dimensions, file sizes, aspect ratios) * **NEW**: Comprehensive documentation for computer vision EDA workflows **Enhanced** ~~~~~~~~~~~~ * Complete EDA suite now includes 15 functions (expanded from 14) * Added Pillow dependency for robust image processing and visualization * Extended edaflow's educational philosophy to computer vision domains * Professional documentation with computer vision examples and workflows * Updated package metadata and dependencies for image processing capabilities **Technical Features** ~~~~~~~~~~~~~~~~~~~~~~ * **Flexible Input Support**: Both directory structures and DataFrame-based workflows * **Quality Assessment**: Automatic detection of corrupted images and data quality issues * **Statistical Analysis**: Comprehensive class balance analysis with actionable warnings * **Professional Visualization**: Smart grid layouts with customizable sampling strategies * **Educational Integration**: Maintains edaflow's core principle of teaching through analysis * **Production Ready**: Robust error handling and validation for real-world datasets Version 0.8.6 (2025-08-05) - PyPI Changelog Display Fix -------------------------------------------------------- **Fixed** ~~~~~~~~~ * **CRITICAL**: Fixed PyPI changelog not displaying latest releases (v0.8.4, v0.8.5) * **DOCUMENTATION**: Updated README.md changelog section that PyPI displays instead of CHANGELOG.md * **PYPI**: Synchronized README.md changelog with comprehensive CHANGELOG.md content * **ENHANCED**: Ensured PyPI users see complete version history and latest features Version 0.8.5 (2025-08-05) - Code Organization and Structure Improvement -------------------------------------------------------------------------- **Changed** ~~~~~~~~~~~ * **REFACTORED**: Renamed ``missing_data.py`` to ``core.py`` to better reflect comprehensive EDA functionality * **ENHANCED**: Updated module docstring to describe complete suite of analysis functions * **IMPROVED**: Better project structure with appropriately named core module containing all 14 EDA functions * **FIXED**: Updated all imports and tests to reference the new core module structure * **MAINTAINED**: Full backward compatibility - all functions work exactly the same Version 0.8.4 (2025-08-05) - Comprehensive Scatter Matrix Visualization Release -------------------------------------------------------------------------------- **Added** ~~~~~~~~~ * **NEW**: ``visualize_scatter_matrix()`` function with advanced pairwise relationship analysis * **NEW**: Flexible diagonal plots: histograms, KDE curves, and box plots * **NEW**: Customizable upper/lower triangles: scatter plots, correlation coefficients, or blank * **NEW**: Color coding by categorical variables for group-specific pattern analysis * **NEW**: Multiple regression line types: linear, polynomial (2nd/3rd degree), and LOWESS smoothing * **NEW**: Comprehensive statistical insights: correlation analysis, pattern identification * **NEW**: Professional scatter matrix layouts with adaptive figure sizing * **NEW**: Full integration with existing edaflow workflow and styling consistency **Enhanced** ~~~~~~~~~~~~ * Complete EDA visualization suite now includes 14 functions (from 13) * Added scikit-learn and statsmodels dependencies for advanced analytics * Updated package metadata and documentation for scatter matrix capabilities **Technical Features** ~~~~~~~~~~~~~~~~~~~~~~ * **Matrix Customization**: Independent control of diagonal, upper, and lower triangle content * **Statistical Analysis**: Automatic correlation strength categorization and reporting * **Regression Analysis**: Advanced trend line fitting with multiple algorithm options * **Color Intelligence**: Automatic categorical/numerical variable handling for color coding * **Performance Optimization**: Efficient handling of large datasets with smart sampling suggestions * **Error Handling**: Comprehensive validation with informative error messages * **Professional Output**: Publication-ready visualizations with consistent edaflow styling Version 0.8.3 (2025-08-04) - Critical Documentation Fix Release ---------------------------------------------------------------- **Fixed** ~~~~~~~~~ * **CRITICAL**: Updated README.md changelog section that PyPI was displaying instead of CHANGELOG.md * **PYPI**: Fixed PyPI changelog display by synchronizing README.md changelog with main CHANGELOG.md * **DOCUMENTATION**: Ensured consistent changelog information across all package files Version 0.8.2 (2025-08-04) - Metadata Enhancement Release ---------------------------------------------------------- **Fixed** ~~~~~~~~~ * **METADATA**: Enhanced PyPI metadata to ensure proper changelog display * **PYPI**: Forced PyPI cache refresh by updating package metadata * **LINKS**: Added additional project URLs for better discoverability Version 0.8.1 (2025-08-04) - Changelog Formatting Release ---------------------------------------------------------- **Fixed** ~~~~~~~~~ * Updated changelog dates to current date format * Removed duplicate changelog header that was causing PyPI display issues * Improved changelog formatting for better PyPI presentation Version 0.8.0 (2025-08-04) - Statistical Histogram Analysis Release -------------------------------------------------------------------- **Added** ~~~~~~~~~ * **NEW**: ``visualize_histograms()`` function with advanced statistical analysis and skewness detection * Comprehensive distribution analysis with normality testing (Shapiro-Wilk, Jarque-Bera, Anderson-Darling) * Advanced skewness interpretation: Normal (\|skew\| < 0.5), Moderate (0.5-1), High (≥1) * Kurtosis analysis: Normal, Heavy-tailed (leptokurtic), Light-tailed (platykurtic) * KDE curve overlays and normal distribution comparisons * Statistical text boxes with comprehensive distribution metrics * Transformation recommendations based on skewness analysis * Multi-column histogram visualization with automatic subplot layout **Enhanced** ~~~~~~~~~~~~ * Updated Complete EDA Workflow to include 12 functions (from 9) * Added histogram analysis as Step 10 in the comprehensive workflow * Enhanced README documentation with detailed histogram function examples * Comprehensive test suite with 7 test scenarios covering various distribution types **Fixed** ~~~~~~~~~ * Fixed Anderson-Darling test attribute error and improved statistical test error handling Version 0.7.0 (2025-08-03) - Comprehensive Heatmap Visualization Release ------------------------------------------------------------------------- **Added** ~~~~~~~~~ * **NEW**: ``visualize_heatmap()`` function with comprehensive heatmap visualizations * Four distinct heatmap types: correlation, missing data patterns, values, and cross-tabulation * Multiple correlation methods: Pearson, Spearman, and Kendall * Missing data pattern visualization with threshold highlighting * Data values heatmap for detailed small dataset inspection * Cross-tabulation heatmaps for categorical relationship analysis **Enhanced** ~~~~~~~~~~~~ * Complete EDA workflow now includes 11 steps with comprehensive heatmap analysis * Updated package features to highlight new heatmap visualization capabilities Version 0.6.0 (2025-08-02) - Interactive Boxplot Visualization Release ----------------------------------------------------------------------- **Added** ~~~~~~~~~ * **NEW**: ``visualize_interactive_boxplots()`` function with full Plotly Express integration * Interactive boxplot visualization with hover tooltips, zoom, and pan functionality * Statistical summaries with emoji-formatted output for better readability * Customizable styling options (colors, dimensions, margins) * Smart column selection for numerical data **Enhanced** ~~~~~~~~~~~~ * Complete EDA workflow now includes 10 steps with interactive final visualization * Added plotly>=5.0.0 dependency for interactive visualizations Version 0.5.1 (2024-01-14) - Documentation Enhancement ------------------------------------------------------- **Fixed** ~~~~~~~~~ * Updated PyPI documentation to properly showcase handle_outliers_median() function * Ensured PyPI page displays the complete 9-step EDA workflow including outlier handling * Synchronized local documentation improvements with PyPI display Version 0.5.0 (2025-08-04) - Outlier Handling Release ------------------------------------------------------ **Added** ~~~~~~~~~ * ``handle_outliers_median()`` function for automated outlier detection and replacement * Multiple outlier detection methods: IQR, Z-score, and Modified Z-score * Complete outlier analysis workflow integration with boxplot visualization * Median-based outlier replacement for robust statistical handling * Flexible column selection with automatic numerical column detection **Fixed** ~~~~~~~~~ * Dtype compatibility improvements to eliminate pandas FutureWarnings * Enhanced error handling and validation for numerical column processing Earlier Versions ---------------- For complete version history, see the `GitHub Releases `_ page. .. note:: This changelog covers the major releases. For detailed commit history and minor updates, visit the `GitHub repository `_.