config

type CacheConfig struct {
	// Enable turns caching on/off (default: false)
	Enable bool `yaml:"enable"`

	// Type specifies the caching strategy:
	// - "memory": In-memory cache (cleared on restart)
	// - "redis": Redis-based persistent cache
	// - "file": File-based persistent cache
	Type string `yaml:"type"`

	// TTL specifies how long to keep cached responses (default: 24h)
	TTL time.Duration `yaml:"ttl"`

	// MaxSize limits the cache size:
	// - For memory cache: maximum number of entries
	// - For file cache: maximum total size in bytes
	MaxSize int64 `yaml:"max_size"`

	// Dir specifies the directory for file-based cache
	Dir string `yaml:"dir,omitempty"`

	// Redis configuration (only used if Type is "redis")
	Redis *RedisCacheConfig `yaml:"redis,omitempty"`
}

type Config struct {
	Server  ServerConfig  `yaml:"server"`
	LLM     LLMConfig     `yaml:"llm"`
	Logging LoggingConfig `yaml:"logging"`
	Routes  []RouteConfig `yaml:"routes"`
}

type LLMConfig struct {
	// Provider specifies the LLM provider (e.g., "openai", "anthropic", "ollama")
	Provider string `yaml:"provider"`

	// Model is the name of the model to use (e.g., "gpt-4", "claude-3-haiku")
	Model string `yaml:"model"`

	// APIKey is the authentication key for the provider's API
	// Use environment variables (e.g., ${OPENAI_API_KEY}) for secure configuration
	APIKey string `yaml:"api_key"`

	// Endpoint is the API endpoint URL
	// For Ollama, this is typically "http://localhost:11434"
	// For OpenAI/Anthropic, this is their respective API endpoints
	Endpoint string `yaml:"endpoint"`

	// SystemPrompt is the default system message for all conversations
	SystemPrompt string `yaml:"system_prompt"`

	// MaxContextTokens specifies the maximum combined tokens (input + output)
	// This is model-dependent:
	// - GPT-4 Turbo: 128,000 tokens
	// - Claude-3-Haiku: 200,000 tokens
	// - Llama2: Varies by version
	MaxContextTokens int `yaml:"max_context_tokens"`

	// Cache configuration (optional)
	Cache *CacheConfig `yaml:"cache,omitempty"`

	// Retry configuration (optional)
	Retry *RetryConfig `yaml:"retry,omitempty"`

	// Options contains provider-specific generation parameters
	Options map[string]interface{} `yaml:"options"`
}

type LoggingConfig struct {
	// Level sets logging verbosity: debug, info, warn, error
	Level string `yaml:"level"`

	// Format specifies log output format: json or text
	Format string `yaml:"format"`
}

type RedisCacheConfig struct {
	// Address is the Redis server address (e.g., "localhost:6379")
	Address string `yaml:"address"`

	// Password for Redis authentication (optional)
	Password string `yaml:"password"`

	// DB is the Redis database number to use
	DB int `yaml:"db"`
}

type RetryConfig struct {
	// MaxRetries is the maximum number of retry attempts (default: 3)
	MaxRetries int `yaml:"max_retries"`

	// InitialDelay is the delay before the first retry (default: 1s)
	InitialDelay time.Duration `yaml:"initial_delay"`

	// MaxDelay caps the maximum delay between retries (default: 30s)
	MaxDelay time.Duration `yaml:"max_delay"`

	// Multiplier increases the delay after each retry (default: 2)
	// The delay pattern will be: initial_delay * (multiplier ^ retry_count)
	Multiplier float64 `yaml:"multiplier"`

	// RetryableErrors specifies which error types should trigger retries
	// Common values: "rate_limit", "timeout", "server_error"
	RetryableErrors []string `yaml:"retryable_errors"`
}

type RouteConfig struct {
	// Path is the URL path to match
	Path string `yaml:"path"`

	// Handler specifies which handler to use for this route
	Handler string `yaml:"handler"`
}

type ServerConfig struct {
	// Port specifies the HTTP server port (default: 8080)
	Port int `yaml:"port"`

	// ReadTimeout is the maximum duration for reading the entire request,
	// including the body (default: 30s)
	ReadTimeout time.Duration `yaml:"read_timeout"`

	// WriteTimeout is the maximum duration before timing out writes of the response
	// (default: 30s)
	WriteTimeout time.Duration `yaml:"write_timeout"`

	// MaxHeaderBytes controls the maximum number of bytes the server will
	// read parsing the request header's keys and values (default: 1MB)
	MaxHeaderBytes int `yaml:"max_header_bytes"`

	// ShutdownTimeout specifies how long to wait for the server to shutdown
	// gracefully before forcing termination (default: 30s)
	ShutdownTimeout time.Duration `yaml:"shutdown_timeout"`
}