dartframe 0.8.4

dartframe: ^0.8.4 copied to clipboard

[MAJOR FEATURE] Window Functions - Exponentially Weighted Moving (EWM) Operations

• NEW: DataFrame.ewm() - Create exponentially weighted window with span, alpha, halflife, or com parameters
• NEW: ewm().mean() - Exponentially weighted moving average for smoothing time series data
• NEW: ewm().std() - Exponentially weighted moving standard deviation for volatility analysis
• NEW: ewm().var_() - Exponentially weighted moving variance for risk measurement
• NEW: ewm().corr() - Exponentially weighted moving correlation (pairwise and with other DataFrame)
• NEW: ewm().cov() - Exponentially weighted moving covariance (pairwise and with other DataFrame)
• ENHANCEMENT: Support for adjustWeights and ignoreNA parameters for flexible weighting schemes
• COMPATIBILITY: Pandas-like API for familiar exponential smoothing workflows

[MAJOR FEATURE] Window Functions - Expanding Window Operations

• NEW: DataFrame.expanding() - Create expanding window with minPeriods parameter
• NEW: expanding().mean() - Expanding mean (cumulative average) for running statistics
• NEW: expanding().sum() - Expanding sum (cumulative sum) for accumulation analysis
• NEW: expanding().std() - Expanding standard deviation for growing window volatility
• NEW: expanding().min() - Expanding minimum (running minimum) for tracking lowest values
• NEW: expanding().max() - Expanding maximum (running maximum) for tracking highest values
• ENHANCEMENT: All expanding operations support minPeriods parameter for minimum observation requirements

[FEATURE] DataFrame Statistical Methods - Data Manipulation Operations

• NEW: DataFrame.clip() - Trim values at input thresholds with lower/upper bounds for outlier control
• NEW: DataFrame.abs() - Compute absolute values for all numeric columns
• NEW: DataFrame.pctChange() - Calculate percentage change between consecutive rows for growth analysis
• NEW: DataFrame.diff() - Calculate first discrete difference between consecutive rows
• NEW: DataFrame.idxmax() - Return index labels of maximum values for each column
• NEW: DataFrame.idxmin() - Return index labels of minimum values for each column
• NEW: DataFrame.qcut() - Quantile-based discretization for specified columns into equal-sized bins
• ENHANCEMENT: Enhanced DataFrame.round() with parameter validation and error handling

[FEATURE] Series Statistical Methods

• NEW: Series.clip() - Trim values at input thresholds with lower/upper bounds
• FIX: Resolved duplicate abs() method causing ambiguity errors in Series extensions
• FIX: Fixed Series extension methods not working on Series<dynamic> by proper type handling

[MAJOR FEATURE] GroupBy Enhancements - Advanced Aggregation and Operations

• NEW: DataFrame.groupBy2() - Create GroupBy object for advanced operations with chainable API
• NEW: GroupBy.transform() - Transform values within groups while maintaining original DataFrame shape
• NEW: GroupBy.filter() - Filter entire groups based on conditions for group-level selection
• NEW: GroupBy.pipe() - Apply chainable functions for method chaining and custom operations
• NEW: GroupBy.nth() - Get nth row from each group (supports negative indexing)
• NEW: GroupBy.head() / GroupBy.tail() - Get first/last n rows from each group
• NEW: Cumulative Operations within Groups:
  • GroupBy.cumsum() - Cumulative sum within each group for running totals
  • GroupBy.cumprod() - Cumulative product within each group
  • GroupBy.cummax() - Cumulative maximum within each group
  • GroupBy.cummin() - Cumulative minimum within each group
• NEW: Flexible Aggregation with Enhanced GroupBy.agg():
  • Single function mode: agg('sum') for simple aggregations
  • Multiple functions mode: agg(['sum', 'mean', 'count']) for multiple statistics
  • Column-specific mode: agg({'col1': 'sum', 'col2': ['mean', 'max']}) for targeted aggregations
  • Named aggregations mode: agg({'total': NamedAgg('amount', 'sum')}) for custom column names
• NEW: NamedAgg class for custom aggregation column names and multiple function support
• NEW: Convenience aggregation methods: sum(), mean(), count(), min(), max(), std(), var_(), first(), last()
• NEW: Utility methods: ngroups property, size() method, groups property for group inspection
• FIX: Fixed list equality issue in groupBy() when using multiple columns - now uses string representation for proper map key equality
• ARCHITECTURE: Lazy evaluation in GroupBy operations for memory efficiency

[FEATURE] Advanced Slicing Methods (Previously Implemented)

• NEW: DataFrame.slice() - Slice with step parameter (forward and reverse slicing)
• NEW: DataFrame.sliceByLabel() - Label-based range slicing (inclusive endpoints)
• NEW: DataFrame.sliceByPosition() - Combined position slicing with step parameter
• NEW: DataFrame.sliceByLabelWithStep() - Label + step combination for flexible slicing
• NEW: DataFrame.everyNthRow() / DataFrame.everyNthColumn() - Convenience sampling methods
• NEW: DataFrame.reverseRows() / DataFrame.reverseColumns() - Simple reversal operations

[ENHANCEMENT] Method Chaining Support

• All new operations return DataFrames/Series for seamless method chaining
• Example: df.clip(lower: 0, upper: 100).abs().round(2) for fluent API usage
• Consistent API design across all statistical and manipulation methods

[ENHANCEMENT] Null Value Handling

• Consistent null value handling across all new operations
• Null values are preserved appropriately in all transformations
• Graceful handling of edge cases (empty DataFrames, single rows, mixed types)

[ENHANCEMENT] Performance Optimizations

• Efficient O(n) and O(n*m) implementations for all new methods
• Lazy evaluation in GroupBy operations for reduced memory footprint
• Memory-efficient implementations suitable for large datasets (1000+ rows tested)
• Performance targets: < 1 second for typical operations

[ENHANCEMENT] Error Handling

• Comprehensive parameter validation with descriptive error messages
• Proper type checking and conversion for mixed data types
• Clear error messages for invalid operations and edge cases

[MAJOR FEATURE] Data Type System - Nullable Types and Type Management

• NEW: Comprehensive DType system with nullable integer, boolean, and string types
  • Int8DType, Int16DType, Int32DType, Int64DType - Nullable integer types with range validation (-128 to 127, -32768 to 32767, etc.)
  • BooleanDType - Nullable boolean with flexible string parsing ('true', 'yes', '1', 'false', 'no', '0')
  • StringDType - Nullable string with optional max length constraints
  • Float32DType, Float64DType - Nullable float types with NaN handling
  • DateTimeDType - Nullable datetime with string and timestamp parsing
  • ObjectDType - Generic object type for mixed data
• NEW: DTypeRegistry - Custom data type registration system
  • register(name, constructor) - Register custom types with string names
  • get(name) - Retrieve registered types by name
  • Built-in type lookup with fallback to custom types
  • Type validation and management
• NEW: DTypes convenience class - Easy type creation
  • DTypes.int8(), DTypes.int16(), DTypes.int32(), DTypes.int64() - Integer type constructors
  • DTypes.float32(), DTypes.float64() - Float type constructors
  • DTypes.boolean(), DTypes.string(), DTypes.datetime() - Other type constructors
  • All types support nullable/non-nullable variants via nullable parameter
• NEW: DataFrame.dtypesDetailed - Automatic type inference
  • Detects optimal types based on data content and value ranges
  • Chooses smallest integer type that fits the data range (Int8 for [-128, 127], etc.)
  • Handles nullable vs non-nullable type detection
  • Smart string parsing: infers numeric types from parsable strings (e.g., '123' → Int8DType)
• NEW: DataFrame.astype() - Enhanced type conversion with categorical support
  • Convert columns to specific DType objects: df.astype({'col': DTypes.int32()})
  • Support for Map<String, DType> and Map<String, String> formats
  • Error handling modes: 'raise', 'ignore', 'coerce' for flexible conversion
  • CATEGORICAL SUPPORT: df.astype({'col': 'category'}) delegates to existing categorical system
  • Automatic fallback for categorical and other existing types
  • Full compatibility with existing astype() behavior
• NEW: DataFrame.inferDTypes() - Automatic type optimization
  • Infer and convert to optimal types automatically
  • Downcast options: 'integer', 'float', 'all' for memory optimization
  • Smart string-to-number inference for data cleaning
  • Reduces memory usage by selecting smallest appropriate types
• NEW: DataFrame.memoryUsageByDType() - Memory usage analysis
  • Calculate memory usage per column based on dtype
  • Fixed-size type calculations (Int8=1 byte, Int16=2 bytes, etc.)
  • Variable-size type estimation for strings and objects
  • Returns Map<String, int> of column names to bytes
• NEW: Series.dtypeInfo - Series type information
  • Get DType information for Series data
  • Automatic type inference with range-based integer selection
  • Avoids conflict with existing dtype property
• NEW: Series.astype() - Series type conversion with categorical support
  • Convert Series to specific DType: s.astype(DTypes.int8())
  • CATEGORICAL SUPPORT: s.astype('category', categories: [...], ordered: true)
  • In-place conversion for categorical type (modifies original Series)
  • Returns new Series for other type conversions
  • Support for 'int', 'float', 'string', 'object' string names
  • Error handling: 'raise', 'ignore', 'coerce' modes
  • Full compatibility with existing categorical system
• NEW: Series.memoryUsageByDType() - Series memory analysis
  • Calculate memory usage based on inferred dtype
  • Range-based integer type detection for accurate estimates
• NEW: Series public methods for dtype management
  • toCategorical(categories, ordered) - Convert to categorical type
  • setDType(dtype) - Set dtype string identifier
  • clearCategorical() - Clear categorical data
• ENHANCEMENT: Smart Type Inference
  • Parses string content to infer numeric types: ['1', '2', '3'] → Int8DType
  • Range-based integer type selection: values [1,2,3] → Int8, [1000,2000] → Int16
  • Automatic detection of parsable integers and floats in string data
• ENHANCEMENT: Robust Error Handling
  • All numeric and datetime types throw FormatException for unparsable strings
  • Proper exception propagation with 'raise' error mode
  • Null coercion with 'coerce' error mode
  • Value preservation with 'ignore' error mode
• COMPATIBILITY: Works alongside existing type system
  • Categorical conversion delegates to existing _Categorical implementation
  • No breaking changes to existing DataFrame/Series behavior
  • Extension-based implementation for clean separation of concerns
• PERFORMANCE: Memory-efficient type storage and conversion
  • Optimal type selection reduces memory footprint
  • Efficient conversion algorithms
  • Lazy evaluation where appropriate
• VALIDATION: Comprehensive type validation and range checking
  • Integer range validation (Int8: -128 to 127, Int16: -32768 to 32767, etc.)
  • Type compatibility checks
  • Null handling for nullable types
• TESTING: 52 comprehensive tests (22 dtype + 30 categorical)
  • Full integration testing with DataFrame and Series
  • Categorical compatibility testing
  • Error handling and edge case coverage
  • Memory usage calculation verification

[MAJOR FEATURE] Database Support - SQL Database Integration

• NEW: DatabaseConnection abstract interface for database operations
  • query() - Execute SQL queries and return DataFrame
  • execute() - Execute SQL commands (INSERT, UPDATE, DELETE) and return affected rows
  • executeBatch() - Execute multiple SQL commands efficiently
  • beginTransaction() - Start database transactions for ACID compliance
  • close() - Close database connection
  • isConnected() - Check connection status
  • databaseType - Get database type identifier
• NEW: DatabaseTransaction interface for transaction management
  • query() - Execute queries within transaction
  • execute() - Execute commands within transaction
  • commit() - Commit transaction changes
  • rollback() - Rollback transaction on errors
• NEW: ConnectionPool class for efficient connection management
  • getConnection() - Get connection from pool
  • releaseConnection() - Return connection to pool
  • close() - Close all pooled connections
  • activeConnectionCount - Monitor active connections
  • availableConnectionCount - Monitor available connections
  • Configurable max connections (default: 5)
  • Automatic connection lifecycle management
• NEW: Database-specific implementations
  • SQLiteConnection - SQLite database support
  • PostgreSQLConnection - PostgreSQL database support
  • MySQLConnection - MySQL database support
  • Each with transaction support and batch operations
• NEW: DatabaseReader utility class
  • readSqlQuery() - Read SQL query results into DataFrame (pandas-like read_sql_query)
  • readSqlTable() - Read entire SQL table into DataFrame (pandas-like read_sql_table)
  • createConnection() - Factory method for creating database connections
  • Support for WHERE clauses, LIMIT, OFFSET, column selection
• NEW: DataFrame.toSql() extension method (pandas-like to_sql)
  • Write DataFrame to SQL database tables
  • ifExists modes: 'fail', 'replace', 'append' for table handling
  • Automatic table creation with type inference
  • Custom data type mapping via dtype parameter
  • Chunked inserts for large datasets (configurable chunkSize)
  • Index column support with custom labels
  • Automatic SQL type inference from Dart types
• FEATURE: Parameterized Queries
  • All query methods support parameters argument
  • SQL injection prevention through parameter binding
  • Type-safe parameter handling
• FEATURE: Batch Operations
  • executeBatch() for bulk inserts/updates
  • Optimized for high-performance bulk operations
  • Reduces database round-trips
• FEATURE: Transaction Support
  • Full ACID compliance with begin/commit/rollback
  • Automatic rollback on errors
  • Nested transaction prevention
  • Transaction state management
• ARCHITECTURE: Production-ready structure
  • Abstract interfaces for easy extension
  • Mock implementations for testing
  • Ready for real database driver integration (sqflite, postgres, mysql1)
  • Separation of concerns with transaction classes
• ERROR HANDLING: Comprehensive exception types
  • DatabaseConnectionError - Connection failures
  • DatabaseQueryError - Query execution failures
  • DatabaseTransactionError - Transaction failures
  • UnsupportedDatabaseError - Unsupported database types
• EXAMPLES: Complete example files
  • example/database_example.dart - Mock examples with 10 scenarios
  • example/database_real_example.dart - Real database examples (SQLite Northwind, PostgreSQL PostGIS)
  • DATABASE_SETUP.md - Comprehensive setup guide
• COMPATIBILITY: Pandas-like API for familiar usage patterns

[DOCUMENTATION] Comprehensive Documentation Added

• WINDOW_FUNCTIONS_SUMMARY.md - Complete window functions documentation with examples
• EWM_CORR_COV_SUMMARY.md - EWM correlation and covariance documentation
• GROUPBY_ENHANCEMENTS_SUMMARY.md - GroupBy operations documentation with use cases
• ADVANCED_SLICING_SUMMARY.md - Advanced slicing documentation
• CLIP_ABS_SUMMARY.md - DataFrame operations documentation
• SERIES_CLIP_SUMMARY.md - Series clip documentation
• DATAFRAME_COMPLETE_METHODS_SUMMARY.md - Complete statistical methods documentation
• example/window_functions_example.dart - Window functions usage examples
• example/dataframe_operations_example.dart - DataFrame operations examples
• example/series_clip_example.dart - Series clip usage examples
• example/dtype_example.dart - Comprehensive DType system examples with 10 usage scenarios

[TESTING] Comprehensive Test Coverage - 221+ New Tests Added

• 47 tests for window functions (test/window_functions_test.dart)
• 47 tests for GroupBy operations (test/groupby_enhanced_test.dart)
• 44 tests for advanced slicing (test/advanced_slicing_test.dart)
• 36 tests for DataFrame operations (test/dataframe_operations_test.dart)
• 28 tests for missing DataFrame methods (test/dataframe_missing_methods_test.dart)
• 19 tests for Series clip (test/series_clip_test.dart)
• Edge case coverage: empty DataFrames/Series, single row/column, mixed types, null values
• Large dataset performance validation (1000+ rows)

[FEATURE] Export Formats - Multiple Output Formats

• NEW: toLatex() - Export DataFrame to LaTeX table format
  • Support for captions, labels, and position specifiers
  • Automatic escaping of special LaTeX characters
  • Longtable environment for multi-page tables
  • Custom column format strings
  • Bold headers and configurable styling
• NEW: toMarkdown() - Export DataFrame to Markdown table format
  • GitHub-flavored markdown (pipe format)
  • Grid and simple table formats
  • Column alignment options (left, center, right)
  • Float formatting for numeric precision
  • Maximum column width for truncation
• NEW: toStringFormatted() - Enhanced formatted string representation
  • Intelligent truncation for large DataFrames
  • Configurable max rows and columns
  • Float formatting support
  • Shape information footer
  • Pandas-like display with ellipsis
• NEW: toRecords() - Convert DataFrame to list of record maps
  • Optional index inclusion
  • Custom index column naming
  • Perfect for JSON serialization
  • Row-by-row iteration support
• COMPATIBILITY: Pandas-like API for familiar export workflows

[FEATURE] Web & API - HTML and XML Support

• NEW: toHtml() - Export DataFrame to HTML table format
  • CSS classes and table ID support
  • Notebook styling for Jupyter-like display
  • Configurable borders and alignment
  • Automatic HTML entity escaping
  • Truncation for large DataFrames
  • Dimension display footer
• NEW: toXml() - Export DataFrame to XML format
  • Custom root and row element names
  • Attribute and element column modes
  • XML entity escaping
  • Pretty print with indentation
  • Index inclusion control
• NEW: DataFrame.readHtml() - Read HTML tables from string
  • Automatic table detection and parsing
  • Header row specification
  • Numeric value parsing
  • HTML entity decoding
  • Multiple table support
• NEW: DataFrame.readXml() - Read XML data into DataFrame
  • Custom row element selection
  • Attribute extraction with prefix
  • Numeric value parsing
  • XML entity decoding
  • Flexible column detection
• COMPATIBILITY: Round-trip support for HTML and XML formats

[API COMPLETENESS] Feature Parity Achievement

• DataFrame and Series now have complete statistical method coverage
• All 11 statistical methods implemented for both DataFrame and Series
• Pandas-like API with consistent method signatures across all operations
• Full support for method chaining and functional programming patterns
• Complete export format support for documentation and reporting

[MAJOR FEATURE] Comprehensive String Operations Extension

• NEW: Pattern extraction methods - str.extract(), str.extractall(), str.findall() for regex-based text processing
• NEW: String padding and justification - str.pad(), str.center(), str.ljust(), str.rjust(), str.zfill() for text alignment
• NEW: String slicing and manipulation - str.slice(), str.get() for advanced substring operations
• NEW: String concatenation and repetition - str.cat(), str.repeat() for text composition
• NEW: String type checking methods - str.isalnum(), str.isalpha(), str.isdigit(), str.isspace(), str.islower(), str.isupper(), str.istitle(), str.isnumeric(), str.isdecimal() for character validation
• COMPATIBILITY: Pandas-like string accessor API for familiar text operations

[MAJOR FEATURE] Enhanced Categorical Data Operations

• NEW: cat.reorderCategories() - Reorder category levels with ordering control
• NEW: cat.addCategories() - Add new categories to existing categorical data
• NEW: cat.removeCategories() - Remove unused categories with validation
• NEW: cat.renameCategories() - Rename categories using mapping dictionaries
• NEW: cat.setCategories() - Set categories with recode and rename modes
• NEW: cat.asOrdered() / cat.asUnordered() - Convert between ordered and unordered categorical types
• NEW: cat.min() / cat.max() - Min/max operations for ordered categories
• NEW: cat.memoryUsage() - Memory usage analysis and optimization metrics for categorical storage
• ENHANCEMENT: All categorical operations integrated with CategoricalAccessor interface

[FEATURE] DataFrame Duplicate Handling and Selection Methods

• NEW: duplicated() - Identify duplicate rows with configurable subset and keep options
• NEW: dropDuplicates() - Remove duplicate rows from DataFrame
• NEW: nlargest() - Select N rows with largest values in specified column
• NEW: nsmallest() - Select N rows with smallest values in specified column

[FEATURE] Functional Programming Extensions

• NEW: apply() - Apply function along axis (rows or columns) with flexible operation support
• NEW: applymap() - Element-wise function application across entire DataFrame
• NEW: agg() - Aggregate with multiple functions simultaneously for complex aggregations
• NEW: transform() - Transform values while preserving DataFrame structure
• NEW: pipe() - Apply chainable functions for method composition

[MAJOR FEATURE] GroupBy Enhancements with Advanced Operations

• NEW: GroupBy class providing chainable API for grouped operations
• NEW: groupBy2() - Returns GroupBy object for method chaining and advanced groupby workflows
• NEW: Transform operations - transform(), transformMean(), transformSum() for group-wise transformations
• NEW: Filter operations - filter() method for group-wise filtering based on conditions
• NEW: Cumulative operations - cumsum(), cumprod(), cummax(), cummin() for cumulative calculations within groups
• NEW: Row selection - nth(), head(), tail() for selecting specific rows within groups
• NEW: NamedAgg class for named aggregations with multiple function support
• NEW: pipe() for method chaining and custom group operations
• ARCHITECTURE: Seamless integration with existing groupBy functionality

[MAJOR FEATURE] Advanced Time Series Operations (12 new methods)

• NEW: Shift operations - shift(), lag(), lead() for time series data alignment
• NEW: Time index operations - tshift() for shifting by time period, asfreq() for frequency conversion
• NEW: Time-based filtering - atTime(), betweenTime() for time window selection, first(), last() for period endpoints
• NEW: Timezone operations - tzLocalize() for adding timezone info, tzConvert() for timezone conversion, tzNaive() for removing timezone
• COMPATIBILITY: Pandas-like API for seamless time series workflows

[FEATURE] Enhanced Resampling Operations

• NEW: resampleOHLC() - Open, High, Low, Close resampling for OHLC data aggregation
• NEW: resampleNunique() - Count unique values per resampling period
• NEW: resampleWithOffset() - Resampling with custom time offset support
• ENHANCEMENT: Advanced time series data transformations with period-based aggregation

[FEATURE] Advanced Data Slicing Methods (6 new methods)

• NEW: slice() - Flexible slicing with step parameter support
• NEW: sliceByLabel() - Label-based range slicing for index-based selection
• NEW: sliceByPosition() - Combined position and range slicing operations
• NEW: sliceByLabelWithStep() - Label-based slicing with step increments
• NEW: everyNthRow() / everyNthColumn() - Convenience methods for sampling every nth element
• NEW: reverseRows() / reverseColumns() - Row and column reversal operations

[FEATURE] Expression Evaluation and Querying

• NEW: eval() - Evaluate string expressions for computed columns and values
• NEW: query() - Query DataFrame using intuitive string expressions with variable binding
• ENHANCEMENT: Chainable expression evaluation for complex data transformations

[FEATURE] MultiIndex and Advanced Indexing Support

• NEW: MultiIndex - Hierarchical indexing for multi-level row/column structures
• NEW: DatetimeIndex - Timezone-aware datetime indexing with frequency support
• NEW: TimedeltaIndex - Time difference indexing for duration-based operations
• NEW: PeriodIndex - Time period indexing for period-based time series
• ARCHITECTURE: Native support for multi-dimensional hierarchical data structures

[MAJOR FEATURE] Enhanced File I/O Support with Web Compatibility

• NEW: Full CSV support using csv package - read/write with custom delimiters, headers, encoding
• NEW: Full Excel support using excel package- read/write .xlsx/.xls files with multi-sheet operations
• NEW: Multi-sheet Excel operations - readAllExcelSheets() and writeExcelSheets() for working with entire workbooks
• NEW: Platform-agnostic FileIO abstraction - works on desktop, mobile, and web without code changes
• NEW: Binary file support - readBytesFromFile() and writeBytesToFile() for Excel and other binary formats
• NEW: deleteFile() method added to FileIO interface for temporary file cleanup
• ENHANCEMENT: All file readers/writers now use FileIO for cross-platform compatibility
• ENHANCEMENT: DataFrame I/O methods now support both file paths and string content
  • DataFrame.fromCSV() supports both path and csv parameters with full DataFrame options (formatData, missingDataIndicator, replaceMissingValueWith, allowFlexibleColumns)
  • DataFrame.fromJson() supports both path and jsonString parameters with all orientations and DataFrame options
  • Automatic temporary file handling for string-based input with proper cleanup
• ENHANCEMENT: Unified toJSON() method combines in-memory conversion and file writing
  • Returns JSON structure when path is null (in-memory mode)
  • Writes to file when path is provided (file mode)
  • Supports all orientations: 'records', 'index', 'columns', 'values'
• WEB: Full web browser support - upload files for processing, download results
• Comprehensive documentation with examples for all file formats

[MAJOR FEATURE] HDF5 File Support - Pure Dart implementation

• NEW: Read HDF5 datasets with FileReader.readHDF5() - compatible with Python h5py, MATLAB v7.3, R
• NEW: All major datatypes supported (numeric, strings, compounds, arrays, enums, variable-length, timestamps)
• NEW: Multi-dimensional datasets (3D+) with automatic flattening and shape preservation
• NEW: Compression support (gzip, lzf, shuffle filter) and chunked storage with B-tree indexing
• NEW: Group navigation, attributes, metadata, and dataset slicing
• NEW: Cross-platform compatible (Windows, macOS, Linux, Web, iOS, Android) - no FFI dependencies

[BREAKING CHANGE] All functions with field name inputFilePath have been simplified to path.

• REMOVAL: The GeoDataFrame and GeoSeries classes, along with all related spatial analysis methods, have been completely removed from the dartframe package.
• REASON: This change was made to streamline the core library, reduce its size, and separate concerns. Geospatial functionality is now housed in a dedicated, specialized package.
• MIGRATION: All geospatial features have been migrated to the new geoengine package. To continue using GeoDataFrame and GeoSeries, please add geoengine to your pubspec.yaml dependencies.
• You can find the new package here: geoengine on pub.dev.
• This move allows for more focused development on both the core data manipulation features in dartframe and the geospatial capabilities in geoengine.

• NEW: TimeSeriesIndex class for time-based indexing and operations
  • Support for timestamps with frequency information (Daily, Hourly, Monthly, Yearly)
  • Factory constructor TimeSeriesIndex.dateRange() for creating time series ranges
  • Automatic frequency detection with detectFrequency() method
  • Utility methods: slice(), contains(), indexOf(), asFreq() for frequency conversion
  • Support for empty time series with proper error handling
• NEW: FrequencyUtils class with time series utilities
  • Frequency normalization and validation (normalizeFrequency(), isValidFrequency())
  • Human-readable frequency descriptions (frequencyDescription())
  • Duration calculations for supported frequencies (getFrequencyDuration())
  • Support for frequency aliases (daily, hourly, monthly, yearly, annual)
• NEW: DataFrameTimeSeries extension for DataFrame time series operations
  • resample() method for changing time series frequency with aggregation functions
  • Support for multiple aggregation functions: mean, sum, min, max, count, first, last
  • upsample() method for increasing frequency with fill methods (pad/ffill, backfill/bfill, nearest)
  • downsample() method for decreasing frequency with aggregation
  • Automatic date column detection in DataFrames
  • Comprehensive error handling for edge cases and invalid inputs
• NEW: Comprehensive test coverage with 43 passing tests
  • Tests for TimeSeriesIndex constructor, properties, and utility methods
  • Tests for dateRange() factory constructor with various frequencies (D, H, M, Y)
  • Tests for frequency detection and validation functionality
  • Tests for DataFrame resampling, upsampling, and downsampling operations
  • Edge case testing for empty DataFrames, null dates, and mixed data types
  • Error handling validation for unsupported frequencies and methods
• ARCHITECTURE: Seamless integration with existing DataFrame and Series classes
• COMPATIBILITY: Pandas-like API for familiar time series operations
• PERFORMANCE: Efficient time-based operations with proper indexing
• All time series functionality follows pandas conventions for easy migration

• NEW: Series.astype('category') method for converting Series to categorical dtype (pandas-compatible)
• NEW: CategoricalAccessor (.cat) providing pandas-like categorical operations interface
• NEW: Efficient memory storage using integer codes with category labels mapping
• NEW: Support for both ordered and unordered categorical data types
• NEW: Category management operations:
  • series.cat.addCategories() - Add new categories to existing categorical data
  • series.cat.removeCategories() - Remove unused categories with validation
  • series.cat.renameCategories() - Rename categories using mapping dictionary
  • series.cat.reorderCategories() - Reorder categories and set ordered flag
• NEW: Categorical properties and methods:
  • series.cat.categories - Access category labels
  • series.cat.codes - Access integer codes
  • series.cat.ordered - Check if categories are ordered
  • series.cat.nCategories - Get number of categories
  • series.cat.unique() - Get unique categories present in data
  • series.cat.contains() - Check if categorical contains specific value
• NEW: series.isCategorical property to check if Series is categorical
• NEW: series.isCategoricalLike() method to detect categorical-suitable data
• NEW: series.seriesDtype property for pandas-like dtype information
• NEW: Enhanced series.dtype getter to handle categorical data types
• NEW: Seamless DataFrame integration - categorical Series work in all DataFrame operations
• NEW: Automatic data synchronization between categorical codes and Series values
• NEW: Type conversion support: 'category', 'object', 'int', 'float', 'string' dtypes
• ARCHITECTURE: Integrated approach - no separate CategoricalSeries class needed
• ARCHITECTURE: Internal _Categorical class for efficient categorical storage
• PERFORMANCE: Memory optimization - categorical encoding only when beneficial
• COMPATIBILITY: Full pandas API compatibility for categorical operations
• All existing Series methods (length, nunique, valueCounts, etc.) work seamlessly with categorical data
• Maintains backward compatibility - no breaking changes to existing Series functionality

• NEW: Complete test coverage for Series interpolation methods (linear, polynomial, spline)
• NEW: Extensive testing of enhanced fill operations (ffill, bfill) with limit parameters
• NEW: Missing data analysis accuracy validation using isna() and notna() methods
• NEW: Integration for combining interpolation, fill operations, and missing data detection
• NEW: Edge case testing for invalid parameters, insufficient data, and mixed data types
• NEW: Custom missing value handling across different scenarios (-999, 'NA', etc.)
• NEW: DataFrame-level missing data operations testing for multi-column scenarios
• Validates error handling, data preservation, and method parameter combinations
• Ensures robust behavior for complex missing data workflows and edge cases

• Supports both named access (shape.rows, shape.columns) and indexed access (shape[0], shape[1])
• Future-proofed for 3D+ data structures (tensors)
• Added utility methods: addDimension(), removeDimension(), transpose(), size, isEmpty, etc.

• NEW: rollingWindow() method with pandas-like API for all columns simultaneously
• Basic operations: mean(), sum(), std(), variance(), min(), max()
• Advanced operations: median(), quantile(), skew(), kurt()
• Correlation operations: corr(), cov() with other DataFrames
• Custom functions: apply() method for user-defined operations
• Support for centered windows, minimum periods, and flexible parameters

• CONSOLIDATION: Moved all basic statistical functions from functions.dart to statistics.dart for better organization
  • Migrated count(), mean(), min(), max(), sum(), describe() with enhanced APIs
  • All functions now have consistent skipna parameter (defaults to true)
  • Improved error handling with ArgumentError instead of generic exceptions
  • Better missing value handling using internal _isMissing() method
  • Enhanced return types: dynamic for min/max/sum to handle missing values properly
• NEW: 19 additional statistical functions added to SeriesStatistics extension
• NEW: Basic Statistics Functions:
  • cumsum() - Cumulative sum over the Series with skipna support
  • nunique() - Count number of unique values with dropna parameter
  • value_counts() - Frequency count with normalize, sort, ascending, and dropna options
• NEW: Percentile Functions:
  • percentile() - Alternative to quantile using 0-100 scale for easier interpretation
  • iqr() - Interquartile Range (Q3 - Q1) for measuring statistical dispersion
• NEW: Advanced Statistics Functions:
  • sem() - Standard Error of the Mean with configurable degrees of freedom
  • mad() - Mean Absolute Deviation for robust central tendency measurement
  • range() - Range (max - min) for measuring data spread
• NEW: Correlation and Covariance Functions:
  • corr() - Pearson correlation coefficient with another Series
  • cov() - Covariance with another Series and configurable degrees of freedom
  • autocorr() - Autocorrelation with configurable lag periods for time series analysis
• NEW: Rank and Order Statistics Functions:
  • rank() - Rank values with 5 tie-breaking methods ('average', 'min', 'max', 'first', 'dense')
  • pct_change() - Percentage change between consecutive values for time series analysis
  • diff() - Difference between consecutive values with configurable periods
• NEW: Robust Statistics Functions:
  • trimmed_mean() - Mean after removing outliers from both tails (configurable proportion)
  • winsorized_mean() - Mean after capping outliers at boundary values
• NEW: Distribution Functions:
  • entropy() - Shannon entropy with configurable logarithm base for information theory
  • geometric_mean() - Geometric mean for positive values (nth root of product)
  • harmonic_mean() - Harmonic mean for positive values (reciprocal of arithmetic mean of reciprocals)
• ENHANCEMENT: All statistical functions feature:
  • Consistent API design with skipna parameters
  • Comprehensive documentation with mathematical explanations and examples
  • Proper edge case handling (empty data, insufficient samples, non-numeric data)
  • Return double.nan or _missingRepresentation for invalid cases instead of throwing exceptions
  • Full integration with existing Series functionality and DataFrame operations
• PERFORMANCE: Optimized mathematical operations using dart:math library functions
• COMPATIBILITY: Pandas-like API for familiar data science workflows
• ARCHITECTURE: Over 40 statistical functions now available in Series class
• Fixed broken describe() function that was calling undefined std() and quantile() methods
• Removed duplicate cumsum() function from functions.dart to avoid conflicts
• All existing tests pass with enhanced statistical functionality

• Added migration guide and compatibility documentation
• Maintained backward compatibility for existing code
• Enhanced performance for multiple rolling operations

• NEW: DataFrameReshaping extension with advanced reshaping operations
  • stack() and unstack() methods for hierarchical indexing and multi-level data structures
  • Enhanced meltEnhanced() method with additional pandas-like parameters (ignoreIndex, colLevel)
  • widen() method as intuitive inverse of melt for long-to-wide format conversion
  • Corrected transpose() method with proper matrix transposition logic
• NEW: Enhanced pivot table functionality
  • pivotTableEnhanced() with support for multiple index and column levels
  • crosstab() method for cross-tabulation analysis with margins and normalization
  • Advanced aggregation options and margin calculations
• NEW: DataFrameMerging extension with enhanced join operations
  • merge() method with pandas-like parameters and validation options
  • concat() method for concatenating DataFrames with flexible options
  • joinEnhanced() method with additional join strategies
  • Merge validation (one-to-one, one-to-many, many-to-one, many-to-many)
  • Framework for mergeAsof() (asynchronous merge operations)

• functions.dart: Contains standard, fully functional methods (melt, pivotTable, join)
• reshaping.dart: Contains enhanced versions with "Enhanced" suffixes (meltEnhanced, pivotTableEnhanced, joinEnhanced)
• Users can choose between standard functionality or enhanced features
• No breaking changes - both approaches coexist for maximum compatibility

• Fixed incorrect column structure that was using index values as column names
• Implemented proper matrix transposition where columns become rows and vice-versa
• Added proper copy logic with _copyValue() helper for different data types
• Enhanced documentation with clear examples of transpose behavior

• getCoordinates() - extracts coordinates as a DataFrame
• countCoordinates - counts coordinate pairs in each geometry
• countGeometries - counts geometries in multi-part geometries
• countInteriorRings - counts interior rings in polygonal geometries
• isClosed - checks if LineStrings are closed
• isEmpty - checks if geometries are empty
• isRing - checks if features are rings
• isValid - validates geometry structures
• hasZ - checks for 3D coordinates
• bounds - gets bounding boxes for geometries
• totalBounds - gets overall bounds of all geometries
• centroid - calculates centroids of geometries
• type - gets geometry types
• area - calculates areas of polygonal geometries
• lengths - calculates lengths of linear geometries
• isCCW - checks if rings are counterclockwise
• contains - checks spatial containment relationships