feat: GoZen v3.0.0 - Context Compression & Middleware Pipeline

[john-zhh]

Overview

This PR implements GoZen v3.0.0 with two major BETA features:

1. Context Compression - Transparent context compression for large conversations
2. Middleware Pipeline - Pluggable middleware architecture for extensibility

All features are disabled by default and marked as BETA.

Features

Context Compression (BETA)

Transparent context compression that intercepts large conversation histories, summarizes them with a cheap model, and forwards compressed requests upstream.

• Token estimation for messages
• Configurable threshold and target tokens
• Preserves recent messages uncompressed
• Web API for configuration and statistics

Middleware Pipeline (BETA)

Transform GoZen into a programmable AI API gateway with a pluggable middleware chain.

Architecture:

• Middleware interface for custom middleware development
• Pipeline executor with priority-based ordering
• Registry for middleware lifecycle management
• PluginLoader for local (.so) and remote plugin support

Built-in Middleware:

Name	Priority	Description
context-injection	10	Auto-inject .cursorrules, CLAUDE.md
session-memory	15	Cross-session intelligence (v3.1)
request-logger	20	Log requests and responses
orchestration	50	Multi-model orchestration (v3.2)

Third-Party Plugin Support

• Local plugins: Load Go plugins (.so files) from disk
• Remote plugins: Download from URL with SHA256 verification
• Middleware development guide for third-party developers

Configuration

{
  "compression": {
    "enabled": false,
    "threshold_tokens": 50000,
    "target_tokens": 20000,
    "summary_model": "claude-3-haiku-20240307",
    "preserve_recent": 4
  },
  "middleware": {
    "enabled": false,
    "middlewares": [
      {
        "name": "context-injection",
        "enabled": true,
        "source": "builtin"
      }
    ]
  }
}

Web API

Compression

• GET/PUT /api/v1/compression - Configuration
• GET /api/v1/compression/stats - Statistics

Middleware

• GET/PUT /api/v1/middleware - Configuration
• GET /api/v1/middleware/{name} - Details
• POST /api/v1/middleware/{name}/enable - Enable
• POST /api/v1/middleware/{name}/disable - Disable
• POST /api/v1/middleware/reload - Reload all

Files Changed

New Files

• internal/proxy/compression.go - Context compressor
• internal/middleware/*.go - Middleware package
• internal/web/api_compression.go - Compression API
• internal/web/api_middleware.go - Middleware API
• docs/middleware-development.md - Development guide

Modified Files

• internal/config/config.go - New config types
• internal/config/store.go - New getters/setters
• internal/proxy/server.go - Integration
• internal/daemon/server.go - Initialization
• cmd/root.go - Version bump to 3.0.0

Testing

All existing tests pass. New tests added for:

• Context compression
• Middleware pipeline
• Built-in middlewares
• Registry and loader

go test ./...

Breaking Changes

None. All features are opt-in and disabled by default.