@aaswe/codebase-ai

# Changelog All notable changes to this project will be documented in this file. The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html). ## [1.1.4] - 2025-08-25 ### 🚀 CRITICAL BUG FIX - Comprehensive TTL Aggregation + Automatic Memory Scaling ### Major TTL System Fixes - **TTL Output Directory Aggregation**: Fixed critical `--output` parameter bug preventing TTL file aggregation - **Root Cause**: Phase 3 of AutoAnalysisWorkflow was being skipped entirely with "Skipping enhanced TTL generation" comment - **Fix**: Replaced skipped Phase 3 with active [`aggregateTTLFilesToOutputDirectory`](src/services/project-analysis/AutoAnalysisWorkflow.ts:1408-1530) method - **Impact**: TTL files now properly aggregate to specified output directory instead of staying scattered in source directories - **Comprehensive Module Analysis**: Fixed TTL files containing incomplete module data - **Root Cause**: TTL aggregation was copying existing single-file TTL files instead of creating comprehensive module analysis - **Before Fix**: Each TTL file contained analysis of only 1 source file (e.g., 1 class instead of 3 classes in a module) - **After Fix**: Each TTL file contains analysis from ALL source files in the module directory - **Technical Solution**: Rewrote [`createModuleAnalysis`](src/services/project-analysis/AutoAnalysisWorkflow.ts:763-901) method to combine `file.analysisResult` from ALL files ### Revolutionary Automatic Memory Scaling - **Zero-Configuration Neo4j Memory Optimization**: System now automatically detects codebase size and configures Neo4j memory without any user intervention - **Automatic Detection**: [`MemoryScalingService`](src/services/infrastructure/MemoryScalingService.ts) analyzes project file count across 15+ programming languages - **Smart Configuration**: Dynamically sets `NEO4J_HEAP_SIZE` and `NEO4J_PAGECACHE_SIZE` based on project category: - **Micro Projects** (< 500 files): 1g heap / 512m cache - **Small Projects** (< 2K files): 2g heap / 1g cache - **Medium Projects** (< 10K files): 4g heap / 2g cache - **Large Projects** (< 50K files): 8g heap / 4g cache - **Enterprise** (< 100K files): 12g heap / 6g cache - **Massive Codebases** (100K+ files): 16g heap / 8g cache - **System Validation**: Automatically checks available system memory and provides upgrade recommendations - **Integration**: Seamlessly integrated into [`full-start`](src/cli/index.ts:266-287) command as Phase 0 before container startup ### Enhanced User Experience & Progress Tracking - **Real-Time Progress Indicators**: Added visual progress tracking during long-running analysis operations - **Progress Updates**: Shows elapsed time and file count every 10 seconds (e.g., "⏳ Analysis in progress... (40s elapsed, processing 184 files)") - **User Feedback**: Clear visual confirmation that system is working during multi-minute analysis operations - **Smart Thresholds**: Progress indicators only appear for codebases with 50+ files to avoid spam - **Comprehensive File Filtering**: Enhanced ignore patterns filter out 80+ types of irrelevant files - **Multi-Language Support**: Covers 15+ programming languages and frameworks - **Build Artifacts**: Filters node_modules, dist, build, target, bin, obj, vendor, venv, __pycache__, coverage, logs - **IDE Files**: Excludes .vscode, .idea, .vs, swap files, OS files (.DS_Store, Thumbs.db) - **Version Control**: Ignores .git, .svn, .hg directories and associated files - **Language-Specific**: Intelligent filtering for Java (.class, .jar), Python (.pyc, .pyo), C++ (.o, .so), etc. - **Improved File Discovery**: Enhanced from 181 to 184 files detected with better pattern matching ### Augmented Code Context Engine Validation - **Successfully Tested**: Revolutionary three-layer intelligence system fully operational - **Layer 1**: 19 comprehensive TTL knowledge files aggregated correctly (keycloak-config-cli validation) - **Layer 2**: Neo4j graph database processed 184 Java files with graceful memory management - **Layer 3**: Complete source code indexing serving 200+ files via MCP protocol - **Production Resilience**: System demonstrates graceful fallback from Neo4j to TTL-based analysis when memory limits are reached - **Real-World Testing**: Validated with 5-minute analysis run processing 184 files - **Error Recovery**: Graceful handling of Neo4j transaction limits with continued TTL generation - **System Monitoring**: Phase 0 memory configuration with 62GB system validation and 1g/512m Neo4j allocation --- ## [1.1.3] - 2025-08-25 ### 🔄 Critical Full-Start Workflow Fix + Auto-Analysis ### Major Workflow Improvements - **Fixed Full-Start Command**: Completely redesigned `full-start` workflow to solve critical initialization order issues - **Before**: MCP server started immediately → `TTL Files: 0 loaded` → Analysis ran separately - **After**: 4-phase workflow ensures proper initialization order and TTL file loading --- ## [1.1.2] - 2025-08-22 ### Critical TTL Directory Isolation Fix + Visual Documentation ### Critical Fixes - **TTL Directory Restriction Bug**: Fixed critical issue causing MCP server to load data from multiple projects simultaneously - **Root Cause**: TTLContextLoader used global glob patterns (`**/*.module-knowledge.ttl`) that searched entire filesystem - **Impact**: Before fix: 82 TTL files loaded (41 AIDe + 41 target project), After fix: 41 TTL files (target project only) - **Technical Solution**: Updated `loadAllTTLFiles()` and `initializeFileWatcher()` methods in `TTLContextLoader.ts` - **CLI Integration**: Added missing `directories: ttlDirectories` configuration at line 650 in `cli/index.ts` - **Docker Configuration**: Enhanced `Dockerfile` with `--ttl-directories ./knowledge` parameter ### Validation Results - **TTL Context Loader**: `Successfully loaded 41/41 TTL files` (target project only) - **WebSocket MCP Server**: `Successfully loaded 41 TTL files` (clean data isolation) - **Stdio MCP Server**: `Loaded 41 TTL files in stdio server` (no cross-project contamination) - **MCP Data Verification**: Confirmed serving correct Java code data from keycloak project only ### Documentation Enhancements - **Visual Overview**: Added Neo4j knowledge graph visualization to README - Interactive module nodes with dependency relationships - Entity metrics and architectural patterns visualization - **Technical Architecture**: Complete documentation of 5-layer AASWE system - **Multi-Project Support**: Clear project isolation and data boundaries ### System Reliability - **Project Data Isolation**: Complete separation between different analyzed codebases - **MCP Server Integrity**: All three TTL loading mechanisms now respect directory restrictions - **Docker Volume Mapping**: Proper `${ANALYSIS_PROJECT_PATH:-./}/knowledge:/app/knowledge` configuration - **Container Health**: Validated full system operation with Neo4j + Redis + MCP server stack --- ## [1.1.1] - 2025-08-22 ### 🔒 Security & Reliability Release - CodeRabbit Fixes ### Security Fixes - **SQL/Cypher Injection Prevention**: Added input validation and sanitization in `MCPStdioServer.ts` - Strict type checking for filePath parameters - Regex-based parameter sanitization to prevent injection attacks - Proper validation with MCPServerError responses - **Error Response ID Correlation**: Fixed error response handling in MCP protocol - Updated `sendError()` method to accept and use proper request IDs - All error responses now correlate with originating requests - Maintains JSON-RPC protocol specification compliance ### Performance & Reliability Improvements - **Memory Leak Prevention**: Fixed memory accumulation in CLI process management - Replaced infinite promise with heartbeat intervals in `cli/index.ts` - Proper SIGINT/SIGTERM signal handling and cleanup mechanisms - Prevents long-running process memory leaks - **Metrics Division by Zero**: Fixed calculation errors in server metrics - Proper edge case handling when totalRequests = 0 - Accurate average response time calculation from first request - Reliable performance monitoring metrics ### Build & Infrastructure Fixes - **Docker Native Module Compilation**: Fixed container build issues - Removed `--ignore-scripts` flag to allow proper native module builds - Fixed source file copy sequence in Dockerfile - All dependencies now compile correctly in containerized environments - **Docker Context Optimization**: Improved build performance - Excludes local `dist/` directory to prevent stale artifacts - Added `!docs/**/*.md` inclusion for complete documentation - Optimized context size for faster builds ### Code Quality Enhancements - **Path Normalization**: Enhanced cross-platform compatibility - Robust Windows/Linux path handling with `toPosix()` helper function - Efficient deduplication using `Set` instead of array filtering - Cleaner, more maintainable path variation generation - **Package Lock Synchronization**: Updated package-lock.json to version 1.1.1 ### Known Non-Critical Issues - **RDF Storage Layer Warning**: Non-critical `baseDirectory` initialization error during startup - System continues to function normally despite warning - TTL files load successfully (38/38 files loaded) - MCP server starts and operates correctly - Will be addressed in future release ### Testing & Validation - **Docker Build Success**: All security fixes validated with successful container builds - **TypeScript Compilation**: No errors during build process - **NPM Dependencies**: All packages install correctly with synchronized lock file - **System Functionality**: Complete MCP server operation confirmed (4 tools, 38 TTL files) --- ## [1.1.0] - 2025-08-22 ### 🎉 Major Release - Complete MCP Integration & Bug Fixes ### Added - **Dual MCP Transport Support**: Full Model Context Protocol integration with both WebSocket and stdio transports - **RooCode/Cline IDE Integration**: Direct CLI IDE support via stdio MCP transport - **Enhanced TTL Context Loading**: Recursive TTL file discovery and loading (38+ files for keycloak-config-cli) - **Docker Health Check Dependencies**: Proper container startup orchestration - **Comprehensive Resource Discovery**: 146+ resources accessible via MCP protocol - **4 MCP Tools Available**: `get_context`, `query_knowledge`, `analyze_code`, `get_file_content` ### Fixed - **Critical TTL Aggregation Bug**: Fixed Neo4j Cypher query syntax in `AutoAnalysisWorkflow.ts` (line 247) - TTL files now show real entity counts instead of zeros - Replaced invalid `ANY()` clauses with direct string interpolation - **Container Startup Race Conditions**: Neo4j now starts before AASWE server - **RDF Storage Configuration Issues**: Disabled problematic monitoring that caused recurring "baseDirectory" errors - **Docker Volume Mounting**: Proper `ANALYSIS_PROJECT_PATH` environment variable support - **Neo4j Connection Reliability**: Eliminated connection refused errors during startup ### Changed - **Version Bump**: 1.0.19 → 1.1.0 - **Repository URLs**: Updated to `https://github.com/AssahBismarkabah/AIDe.git` - **README Documentation**: Comprehensive setup and MCP integration instructions - **Package Self-Dependency**: Updated to version 1.1.0 ### Technical Improvements - **TTL File Processing**: Complete recursive directory scanning for knowledge files - **MCP Server Architecture**: Unified server manager supporting dual transports - **Container Health Monitoring**: Proper health check implementations - **CLI Enhanced Detection**: Prefers local docker-compose over NPM package versions - **Error Handling**: Improved error messages and graceful failure handling ### Testing & Validation - **End-to-End Testing**: Complete system tested with keycloak-config-cli Java project - **MCP Connectivity**: Confirmed working with 4 tools and 146 resources - **Docker Orchestration**: Full containerized deployment operational - **IDE Integration**: RooCode configuration tested and validated - **Multi-Language Support**: Validated with Java, TypeScript, and configuration files ### Dependencies - All dependencies maintained at current stable versions - No breaking changes introduced - Backward compatible with existing configurations --- ## [1.0.19] - Previous Release - Initial stable release with basic functionality