anteater

Added

• Parallel Analysis API: Memory-efficient concurrent file analysis for large codebases
  • ParallelAnalyzer class with configurable concurrency and chunked processing
  • ParallelAnalysisConfig for tuning maxConcurrency (default: 4) and chunkSize (default: 50)
  • FileAnalysisResult for per-file analysis results with error handling
  • AnalysisProgressCallback for real-time progress tracking
  • analyzeWithProgress() method with progress callback support
  • analyzeStream() for sequential streaming results
  • analyzeParallelStream() for parallel streaming with chunked processing
  • Automatic exclusion of generated files (.g.dart, .freezed.dart)
• MetricsAggregator Enhancement: addPrecomputedResult() method for parallel aggregation without AST retention
• Library Documentation: Added parallel analysis usage examples to library doc comment

Changed

• Library Exports: Exported ParallelAnalyzer, ParallelAnalysisConfig, FileAnalysisResult, AnalysisProgressCallback from main library

What's Changed

• chore: update ignores by @brody-0125 in #3
• chore: update ignores (#3) by @brody-0125 in #4
• refactor : ssa versioning and flow facts by @brody-0125 in #5
• refactor : ssa versioning and flow facts (#5) by @brody-0125 in #6

Full Changelog: 0.3.0...0.3.2

Changed

• SDK Version: Updated minimum SDK constraint from ^3.10.0 to ^3.10.7 to fix pub.dev dependency resolution failure

Fixed

• SSA Builder: Added result field versioning for CallInstruction, LoadFieldInstruction, LoadIndexInstruction, NullCheckInstruction, CastInstruction, TypeCheckInstruction
• SSA Builder: Implemented trivial phi use-site replacement with substitution chain resolution
• SSA Builder: Added AwaitInstruction result versioning
• Datalog Facts: Preserved SSA versions in _getValueVarId using variable.toString()
• Datalog Facts: Added flow-sensitive *At facts (AssignAt, AllocAt, CallAt, LoadFieldAt, StoreFieldAt, PhiAt) with block IDs
• Abstract Interpreter: Added missing transfer functions for LoadField, Call, Await, LoadIndex, NullCheck, Cast, TypeCheck
• Abstract Interpreter: Fixed unreachable predecessor filtering in phi transfer function

Changed

• Halstead Metrics: Renamed N1/N2 fields to operatorTotal/operandTotal for Dart naming convention compliance
• SDK Version: Updated minimum SDK constraint from ^3.5.0 to ^3.10.0 for dart_bert_tokenizer compatibility
• Async I/O: Converted async file existence checks to sync versions for better performance

Fixed

• Library Documentation: Added library; directive to fix dangling doc comment warning
• Doc Comments: Escaped angle brackets in documentation to prevent unintended HTML interpretation
• String Interpolation: Fixed string concatenation to use interpolation
• Const Usage: Applied const constructors and declarations throughout test files
• Dynamic Calls: Fixed avoid_dynamic_calls warnings in server tests
• Unnecessary Import: Removed redundant import in metrics_test.dart
• Constructor Ordering: Moved constructors before field declarations per sort_constructors_first lint rule

What's Changed

• feature : support technical debt cost model and neural analysis by @brody-0125 in #1
• feature : support technical debt cost model and neural analysis (#1) by @brody-0125 in #2

Full Changelog: 0.2.0...0.3.0

Added

• Neural Analysis Setup: Semantic clone detection using ONNX Runtime
  • Pre-trained nomic-embed-text model (~522MB) for 768-dimensional embeddings
  • WordPiece tokenization via dart_bert_tokenizer
  • Model file search order: --model option → ./model/ → ~/.anteater/
  • Installation instructions for ONNX Runtime (brew install onnxruntime)
• CLI Command: anteater clones --path <path> for semantic clone detection
  • Detects semantically similar code (same functionality, different syntax)
  • Configurable similarity threshold (--threshold)
  • Custom model/vocab paths (--model, --vocab)
  • Fallback to ~/.anteater/ for global installation
  • Output formats: text, json
• Technical Debt Cost Model: Comprehensive debt tracking and cost estimation
  • DebtItem and DebtType models with 10 debt types (TODO, FIXME, ignore, as dynamic, etc.)
  • DebtSeverity with configurable cost multipliers (critical: 4x, high: 2x, medium: 1x, low: 0.5x)
  • DebtDetector for detecting debt from comments, casts, annotations, and metrics
  • DebtCostCalculator for item and total cost calculation
  • DebtReport with Markdown, JSON, and console output formats
  • DebtAggregator for project-level aggregation and hotspot analysis
  • DebtTrend for tracking debt changes over time
• CLI Command: anteater debt --path <path> for technical debt analysis
  • Output formats: text, json, markdown
  • Threshold-based CI gate with --fail-on-threshold
  • Customizable threshold with --threshold flag
• Debt Configuration: YAML configuration for debt costs and thresholds
  • Per-type cost configuration
  • Metrics-based debt thresholds (MI, complexity, LOC)
  • Configurable unit of measurement (hours, story points)

Full Changelog: https://github.com/brody-0125/anteater/commits/0.2.0

Added

• Style Consistency Rules: 10 production-ready lint rules
  • Safety rules: avoid-dynamic, avoid-global-state, avoid-late-keyword, no-empty-block, no-equal-then-else
  • Quality rules: prefer-first-last, prefer-async-await, avoid-unnecessary-cast, prefer-trailing-comma, binary-expression-order
• Rule Infrastructure
  • StyleRule abstract base class for custom rules
  • RuleRunner for coordinated rule execution
  • RuleRegistry for rule registration and configuration
  • RuleConfig for YAML-based configuration parsing
  • AnalysisResult and AnalysisResultBuilder for multi-file results
• Rule Configuration
  • Per-rule severity override
  • Per-rule file exclusion patterns
  • Global exclusion patterns
• CLI Command: anteater rules --path <path> for running style rules
• Performance Optimizations
  • RegExp pattern caching for glob matching
  • Lazy computed properties in AnalysisResult
  • Unmodifiable collections in RuleSettings
  • Const empty list returns

Changed

• Improved error handling: rule execution failures reported as warnings instead of silent failures

Added

• Style Consistency Rules: 10 production-ready lint rules
  • Safety rules: avoid-dynamic, avoid-global-state, avoid-late-keyword, no-empty-block, no-equal-then-else
  • Quality rules: prefer-first-last, prefer-async-await, avoid-unnecessary-cast, prefer-trailing-comma, binary-expression-order
• Rule Infrastructure
  • StyleRule abstract base class for custom rules
  • RuleRunner for coordinated rule execution
  • RuleRegistry for rule registration and configuration
  • RuleConfig for YAML-based configuration parsing
  • AnalysisResult and AnalysisResultBuilder for multi-file results
• Rule Configuration
  • Per-rule severity override
  • Per-rule file exclusion patterns
  • Global exclusion patterns
• CLI Command: anteater analyze -p <path> for running style rules analysis
• Performance Optimizations
  • RegExp pattern caching for glob matching
  • Lazy computed properties in AnalysisResult
  • Unmodifiable collections in RuleSettings
  • Const empty list returns

Changed

• Improved error handling: rule execution failures reported as warnings instead of silent failures