config

Documentation

Overview

Package config provides file-based configuration loading with validation.

Stability: stable

Package config provides file-based configuration loading with validation and environment variable overrides.

Load reads a JSON file into a struct. LoadWithEnv additionally applies environment variable overrides using the "env" struct tag. Validate checks fields with the "validate" struct tag.

Index

• func Load[T any](path string) (T, error)
• func LoadWithEnv[T any](path, prefix string) (T, error)
• func Validate(v any) []string
• type DistributedConfig
• type EngineConfig
• type TrainingConfig

Functions

func Load

func Load[T any](path string) (T, error)

Load reads a JSON configuration file at path into a value of type T.

func LoadWithEnv

func LoadWithEnv[T any](path, prefix string) (T, error)

LoadWithEnv reads a JSON configuration file and applies environment variable overrides. Fields with an `env:"NAME"` struct tag are overridden by the environment variable PREFIX_NAME if set.

Supported field types: string, int, bool.

func Validate

func Validate(v any) []string

Validate checks struct fields tagged with `validate:"required"`. Returns a slice of error messages for each field that fails validation. Returns nil if all validations pass.

Types

type DistributedConfig

type DistributedConfig struct {
	CoordinatorAddress string `env:"COORDINATOR_ADDRESS" json:"coordinator_address" validate:"required"`
	TimeoutSeconds     int    `env:"TIMEOUT_SECONDS"     json:"timeout_seconds"`
	TLSEnabled         bool   `env:"TLS_ENABLED"         json:"tls_enabled"`
}

DistributedConfig holds configuration for distributed training.

type EngineConfig

type EngineConfig struct {
	Device        string `env:"DEVICE"          json:"device"          validate:"required"`
	MemoryLimitMB int    `env:"MEMORY_LIMIT_MB" json:"memory_limit_mb"`
	LogLevel      string `env:"LOG_LEVEL"       json:"log_level"`
}

EngineConfig holds configuration for the compute engine.

type TrainingConfig

type TrainingConfig struct {
	BatchSize          int    `env:"BATCH_SIZE"          json:"batch_size"          validate:"required"`
	LearningRate       string `env:"LEARNING_RATE"       json:"learning_rate"       validate:"required"`
	Optimizer          string `env:"OPTIMIZER"           json:"optimizer"           validate:"required"`
	Epochs             int    `env:"EPOCHS"              json:"epochs"`
	CheckpointInterval int    `env:"CHECKPOINT_INTERVAL" json:"checkpoint_interval"`
}

TrainingConfig holds configuration for training workflows.