Skip to content

CHANGELOG.md

Latest commit

 

History

History
72 lines (63 loc) · 4.16 KB

CHANGELOG.md

File metadata and controls

72 lines (63 loc) · 4.16 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Added

  • REST Admin API (port 8080) for pattern CRUD and semantic search operations
  • Pattern enrichment pipeline: automatic embedding generation and concept extraction via OpenAI LLM
  • PostgreSQL data persistence with PGVector support for vector similarity search
  • Neo4j backing store for pattern concept relationships and graph traversal
  • MCP server (port 8081) for Claude Code integration with read-only pattern search capability
  • OpenTelemetry observability: distributed tracing, metrics collection, and structured logging
  • Gin HTTP server framework with middleware for tracing and request metrics
  • Configuration management (internal/config): layered loading from defaults, files, and environment variables
  • Server integration with configurable timeouts, TLS support, and graceful shutdown
  • Telemetry package (internal/telemetry) with otelx integration for unified OpenTelemetry setup
  • Middleware package (internal/middleware) with tracing and request metrics for Gin
  • Metrics package (internal/metrics) with domain-specific counters and histograms for patterns and database operations
  • Distributed tracing support via otelgin middleware with trace ID correlation
  • Request metrics: operation counts, duration histograms, and in-flight request counters
  • Version package (internal/version) for build metadata and release information
  • Repository layer (internal/repository) with PostgreSQL implementation for:
    • Patterns: CRUD, similarity search, and enrichment status tracking
    • Skills: definition management and versioning
    • Skill files: content persistence and metadata
    • Chunks: text segmentation for pattern processing
    • Enrichment jobs: status tracking and result storage
    • Agents: metadata and configuration
    • Graph: Neo4j pattern relationships and concept linkage
  • Database schema migrations for all entity types
  • Repository error types: domain-specific errors for conflict, not found, validation, and persistence failures
  • List options for pagination support in repository queries
  • Service layer (internal/service) for business logic:
    • Pattern service: enrichment orchestration, search, and lifecycle management
    • Skill and skill file services for capability management
    • Agent service for user/agent tracking
    • Enrichment service for LLM pipeline coordination
    • Search service for semantic similarity queries
  • OpenAI integration service (internal/service/openai):
    • Embedding generation using text-embedding-3-large model
    • Concept extraction and entity identification via structured chat completions
    • Token usage tracking and error handling
  • Health check endpoint for service readiness and dependency status
  • Docker multi-stage build for optimized image size
  • E2E test suite via Docker Compose:
    • Tests for all API endpoints (agents, skills, skill files, patterns, enrichment operations)
    • MCP server integration tests
    • Database and dependency initialization
  • GitHub Actions CI/CD workflows for automated testing and image publication
  • Comprehensive unit and integration tests with pgxmock for database isolation
  • Makefile targets for building, testing, and documentation generation
  • Swagger UI at /swagger/index.html with OpenAPI 3.0 specification
  • Build script with cleanup traps for Docker Compose teardown

Fixed

  • E2E test Docker Compose configuration naming (mnemonic_tests container reference)
  • CI config and E2E Docker Compose setup for proper service initialization
  • E2E test execution and assertions for all API endpoints

Changed

  • Project extracted and refocused: REST Admin API only, removed routing engine and regex matching components
  • API version path structure: /v1/api/ prefix for all endpoints
  • Server startup now initializes telemetry and observability middleware by default
  • Configuration validation includes log level and timeout validation with fail-fast error reporting
  • Build and CI/CD workflows optimized for mnemonic-api specific requirements