# Claude Trigger Router 完整配置示例
# 这是偏进阶的参考模板，不是新用户的最小起步配置。
# 如果你是首次使用，优先运行 `ctr setup`，或先参考 README 中的最小配置。
# 复制到 ~/.claude-trigger-router/config.yaml 后，先替换 API Key 和模型名。
#
# 常见 provider preset（与 `ctr setup` 保持一致）：
# - openrouter: https://openrouter.ai/api/v1/chat/completions
# - deepseek: https://api.deepseek.com/chat/completions
# - openai-compatible: https://api.openai.com/v1/chat/completions
# - anthropic: https://api.anthropic.com/v1/messages
# - siliconflow: https://api.siliconflow.cn/v1/chat/completions
# - custom/local: 自定义兼容地址，例如 Ollama / vLLM / one-api

HOST: "127.0.0.1"
PORT: 5678

LOG: true
LOG_LEVEL: "debug"

# 1) Models：主入口。每个模型直接配置成一个可引用的接入项
Models:
  - id: sonnet
    api: "https://openrouter.ai/api/v1/chat/completions"
    key: "sk-xxx"
    interface: "openai"
    model: "anthropic/claude-sonnet-4"
    thinking: "auto"
    metadata:
      context_window_tokens: 200000
      safe_input_tokens: 180000

  - id: opus
    api: "https://openrouter.ai/api/v1/chat/completions"
    key: "sk-xxx"
    interface: "openai"
    model: "anthropic/claude-opus-4"
    metadata:
      context_window_tokens: 200000
      safe_input_tokens: 180000

  - id: deepseek_reasoner
    api: "https://api.deepseek.com/chat/completions"
    key: "sk-xxx"
    interface: "openai"
    model: "deepseek-reasoner"
    thinking: "high"
    metadata:
      context_window_tokens: 64000
      safe_input_tokens: 56000

  - id: ollama_qwen
    api: "http://localhost:11434/v1/chat/completions"
    key: "ollama"
    interface: "openai"
    model: "qwen2.5-coder:latest"
    thinking: "off"

# 2) Router：基础路由，统一直接引用 model id
Router:
  default: "sonnet"
  background: "ollama_qwen"
  think: "deepseek_reasoner"
  longContext: "opus"
  longContextThreshold: 60000
  webSearch: "sonnet"

# 3) SmartRouter：统一路由入口。先走高确定性规则，再用候选模型智能兜底
SmartRouter:
  enabled: true
  analysis_scope: "last_message"
  rules:
    - name: "architecture"
      priority: 90
      enabled: true
      description: "架构设计与系统方案任务"
      patterns:
        - type: exact
          keywords:
            - "架构设计"
            - "系统设计"
            - "技术方案"
            - "architecture"
            - "system design"
        - type: regex
          pattern: "(架构|系统设计|技术方案)"
      model: "opus"

    - name: "code_review"
      priority: 80
      enabled: true
      description: "代码审查与风险评估任务"
      patterns:
        - type: exact
          keywords:
            - "代码审查"
            - "code review"
            - "review code"
            - "检查代码"
        - type: regex
          pattern: "(代码|code).{0,5}(审查|review|检查|审核)"
      model: "sonnet"

    - name: "complex_reasoning"
      priority: 70
      enabled: true
      description: "复杂推理与严谨分析任务"
      patterns:
        - type: exact
          keywords:
            - "深入分析"
            - "复杂推理"
            - "逻辑推理"
            - "deep analysis"
            - "reasoning"
        - type: regex
          pattern: "(深入|复杂|详细).{0,5}(分析|推理|解析)"
      model: "deepseek_reasoner"

    - name: "simple_task"
      priority: 10
      enabled: true
      description: "快速简单任务"
      patterns:
        - type: exact
          keywords:
            - "快速"
            - "简单"
            - "quick"
            - "simple"
      model: "ollama_qwen"

  router_model: "sonnet"
  candidates:
    - model: "sonnet"
      description: "通用编程、代码生成、日常调试、多轮对话"

    - model: "deepseek_reasoner"
      description: "复杂推理、逻辑分析、需要严谨思考的任务"

    - model: "opus"
      description: "架构设计、系统规划、长文档与复杂评审任务"

  cache_ttl: 600000
  max_tokens: 256
  fallback: "default"

  sticky:
    enabled: true
    session_ttl_ms: 3600000
    fingerprint_similarity_threshold: 0.82
    break_on_explicit_route: true
    # Alignment 为显式开启项：Claude Code 已经携带会话上下文。
    # 只有跨模型交接摘要值得额外一次调用成本时再开启。
    alignment:
      enabled: true
      summarizer_model: "sonnet"
      max_summary_tokens: 256

  semantic:
    enabled: true
    mode: "embedding"
    threshold: 0.2
    prototypes:
      architecture: "重构 系统 结构 模块 拆分 架构 设计"
      code_review: "代码 审查 风险 评审 review"
      complex_reasoning: "复杂 推理 分析 严谨 逻辑 reasoning"

# 4) Governance：保留响应治理、级联与观测能力；semantic / sticky 已归入 SmartRouter
Governance:
  enabled: true

  cascade:
    enabled: true
    max_attempts: 2
    triggers:
      compile_failure: true
      test_failure: true
      placeholder_patterns:
        - "TODO"
        - "...rest of code"
        - "placeholder"
    levels:
      - from: "sonnet"
        to: "opus"
        reasoning: "high"

  shadow:
    enabled: true
    mode: "async_audit"
    verifier_model: "sonnet"
    sample_rate: 0.2
    checks:
      placeholder_patterns: true
      length_anomaly: true
      missing_code_block: true

  observability:
    anomaly_thresholds:
      min_sample_size: 3
      cascade_warn_rate: 0.4
      cascade_critical_rate: 0.6
      shadow_warn_rate: 0.5
      shadow_critical_rate: 0.7
      latency_warn_ms: 1500
      latency_critical_ms: 3000
      spike_warn_rate: 0.5
      spike_delta_rate: 0.3

# 5) Legacy 兼容：旧 `Providers` 仅作为兼容层保留，不建议新配置继续扩展
# Providers:
#   - name: openrouter
#     api_base_url: "https://openrouter.ai/api/v1/chat/completions"
#     api_key: "sk-xxx"
#     models:
#       - "anthropic/claude-sonnet-4"
#     transformer:
#       use: ["openrouter"]