Configuration Module¶

The config module provides comprehensive configuration management for Perspt, supporting multiple LLM providers, flexible authentication, and intelligent defaults.

Core Structures¶

AppConfig¶

#[derive(Debug, Clone, Deserialize, PartialEq)]
pub struct AppConfig {
    pub providers: HashMap<String, String>,
    pub api_key: Option<String>,
    pub default_model: Option<String>,
    pub default_provider: Option<String>,
    pub provider_type: Option<String>,
}

The main configuration structure containing all configurable aspects of Perspt.

Fields:

providers - Map of provider names to their API base URLs
api_key - Universal API key for authentication
default_model - Default model identifier for LLM requests
default_provider - Name of default provider configuration
provider_type - Provider type classification for API compatibility

Supported Provider Types:

Provider Type	Description
`openai`	OpenAI GPT models (GPT-4o, GPT-4o-mini, o1-preview, o1-mini, o3-mini, o4-mini)
`anthropic`	Anthropic Claude models (Claude 3.5 Sonnet, Claude 3 Opus/Sonnet/Haiku)
`gemini`	Google Gemini models (Gemini 1.5 Pro/Flash, Gemini 2.0 Flash)
`groq`	Groq ultra-fast inference (Llama 3.x models)
`cohere`	Cohere Command models (Command R, Command R+)
`xai`	XAI Grok models (grok-3-beta, grok-3-fast-beta)
`deepseek`	DeepSeek models (deepseek-chat, deepseek-reasoner)
`ollama`	Local model hosting via Ollama (requires local setup)

Example Configuration:

{
  "api_key": "sk-your-api-key",
  "provider_type": "openai",
  "default_model": "gpt-4o-mini",
  "default_provider": "openai",
  "providers": {
    "openai": "https://api.openai.com/v1",
    "anthropic": "https://api.anthropic.com",
    "gemini": "https://generativelanguage.googleapis.com/v1beta",
    "groq": "https://api.groq.com/openai/v1",
    "cohere": "https://api.cohere.com/v1",
    "xai": "https://api.x.ai/v1",
    "deepseek": "https://api.deepseek.com/v1",
    "ollama": "http://localhost:11434/v1"
  }
}

Core Functions¶

process_loaded_config¶

pub fn process_loaded_config(mut config: AppConfig) -> AppConfig

Processes and validates loaded configuration, applying intelligent defaults and provider type inference.

Parameters:

config - The configuration to process

Returns:

AppConfig - Processed configuration with inferred values

Provider Type Inference Logic:

If provider_type is None, attempts inference from default_provider:

Default Provider	Inferred Type	Notes
`openai`	`openai`	Direct mapping
`anthropic`	`anthropic`	Direct mapping
`google`, `gemini`	`gemini`	Multiple aliases supported
`groq`	`groq`	Direct mapping
`cohere`	`cohere`	Direct mapping
`xai`	`xai`	Direct mapping
`deepseek`	`deepseek`	Direct mapping
`ollama`	`ollama`	Direct mapping
Unknown	`openai`	Fallback default

Example:

let mut config = AppConfig {
    providers: HashMap::new(),
    api_key: None,
    default_model: None,
    default_provider: Some("anthropic".to_string()),
    provider_type: None, // Will be inferred as "anthropic"
};

let processed = process_loaded_config(config);
assert_eq!(processed.provider_type, Some("anthropic".to_string()));

load_config¶

pub async fn load_config(config_path: Option<&String>) -> Result<AppConfig>

Loads application configuration from a file or provides comprehensive defaults.

Parameters:

config_path - Optional path to JSON configuration file

Returns:

Result<AppConfig> - Loaded configuration or error

Behavior:

With Configuration File (Some(path))

Reads JSON file from filesystem
Parses JSON into AppConfig structure
Processes configuration with process_loaded_config()
Returns processed configuration

Without Configuration File (None)

Creates default configuration with all supported provider endpoints pre-configured and OpenAI as default provider.

Default Provider Endpoints:

{
    "openai": "https://api.openai.com/v1",
    "anthropic": "https://api.anthropic.com",
    "gemini": "https://generativelanguage.googleapis.com/v1beta",
    "groq": "https://api.groq.com/openai/v1",
    "cohere": "https://api.cohere.com/v1",
    "xai": "https://api.x.ai/v1",
    "deepseek": "https://api.deepseek.com/v1",
    "ollama": "http://localhost:11434/v1"
}

Possible Errors:

File system errors (file not found, permission denied)
JSON parsing errors (invalid syntax, missing fields)
I/O errors during file reading

Examples:

// Load from specific file
let config = load_config(Some(&"config.json".to_string())).await?;

// Use defaults
let default_config = load_config(None).await?;

// Error handling
match load_config(Some(&"missing.json".to_string())).await {
    Ok(config) => println!("Loaded: {:?}", config),
    Err(e) => eprintln!("Failed to load config: {}", e),
}

Configuration Examples¶

Basic OpenAI Configuration¶

{
  "api_key": "sk-your-openai-key",
  "provider_type": "openai",
  "default_model": "gpt-4o-mini"
}

Multi-Provider Configuration¶

{
  "api_key": "your-default-key",
  "provider_type": "anthropic",
  "default_model": "claude-3-5-sonnet-20241022",
  "default_provider": "anthropic",
  "providers": {
    "openai": "https://api.openai.com/v1",
    "anthropic": "https://api.anthropic.com",
    "gemini": "https://generativelanguage.googleapis.com/v1beta",
    "groq": "https://api.groq.com/openai/v1",
    "cohere": "https://api.cohere.com/v1",
    "xai": "https://api.x.ai/v1",
    "deepseek": "https://api.deepseek.com/v1",
    "ollama": "http://localhost:11434/v1"
  }
}

Minimal Configuration with Provider Inference¶

{
  "default_provider": "groq",
  "default_model": "llama-3.1-70b-versatile"
}

Provider type will be automatically inferred as “groq”

Development with Multiple Models¶

{
  "provider_type": "openai",
  "default_model": "gpt-4o-mini",
  "default_provider": "openai",
  "providers": {
    "openai": "https://api.openai.com/v1",
    "groq": "https://api.groq.com/openai/v1",
    "ollama": "http://localhost:11434/v1"
  }
}