venice-ai-sdk-apl/dist/types/chat.d.ts

A comprehensive SDK for the Venice AI API with CLI support, programmatic CLI usage, CLI-style interface, and interactive demo

/**
 * Chat API types
 *
 * This file contains type definitions for the Chat API.
 */
import { VeniceParameters } from './common';
/**
 * Role of a message in a chat conversation
 */
export type ChatRole = 'system' | 'user' | 'assistant' | 'function' | 'tool';
/**
 * A message in a chat conversation
 */
export interface ChatMessage {
    /**
     * Role of the message author
     */
    role: ChatRole;
    /**
     * Content of the message
     */
    content: string | null;
    /**
     * Name of the author (optional)
     */
    name?: string;
    /**
     * Function call (optional)
     */
    function_call?: {
        name: string;
        arguments: string;
    };
    /**
     * Tool calls (optional)
     */
    tool_calls?: Array<{
        id: string;
        type: 'function';
        function: {
            name: string;
            arguments: string;
        };
    }>;
}
/**
 * Function definition for function calling
 */
export interface FunctionDefinition {
    /**
     * Name of the function
     */
    name: string;
    /**
     * Description of the function
     */
    description?: string;
    /**
     * Parameters schema in JSON Schema format
     */
    parameters: Record<string, any>;
}
/**
 * Tool definition for tool calling
 */
export interface ToolDefinition {
    /**
     * Type of the tool
     */
    type: 'function';
    /**
     * Function definition
     */
    function: FunctionDefinition;
}
/**
 * Parameters for creating a chat completion
 */
export interface CreateChatCompletionParams {
    /**
     * ID of the model to use
     */
    model: string;
    /**
     * Messages to generate chat completions for
     */
    messages: ChatMessage[];
    /**
     * Functions that the model may call
     * @deprecated Use tools instead
     */
    functions?: FunctionDefinition[];
    /**
     * Controls how the model calls functions
     * @deprecated Use tool_choice instead
     */
    function_call?: 'auto' | 'none' | {
        name: string;
    };
    /**
     * Tools that the model may call
     */
    tools?: ToolDefinition[];
    /**
     * Controls how the model calls tools
     */
    tool_choice?: 'auto' | 'none' | {
        type: 'function';
        function: {
            name: string;
        };
    };
    /**
     * Maximum number of tokens to generate
     */
    max_tokens?: number;
    /**
     * Maximum number of tokens to generate in the completion
     */
    max_completion_tokens?: number;
    /**
     * Sampling temperature (0-2)
     */
    temperature?: number;
    /**
     * Nucleus sampling parameter (0-1)
     */
    top_p?: number;
    /**
     * Number of chat completion choices to generate
     */
    n?: number;
    /**
     * Whether to stream the response
     */
    stream?: boolean;
    /**
     * Stop sequences that trigger early stopping
     */
    stop?: string | string[];
    /**
     * Presence penalty (-2 to 2)
     */
    presence_penalty?: number;
    /**
     * Frequency penalty (-2 to 2)
     */
    frequency_penalty?: number;
    /**
     * Logit bias
     */
    logit_bias?: Record<string, number>;
    /**
     * User identifier
     */
    user?: string;
    /**
     * Venice-specific parameters
     */
    venice_parameters?: VeniceParameters;
}
/**
 * Chat completion choice
 */
export interface ChatCompletionChoice {
    /**
     * Index of the choice
     */
    index: number;
    /**
     * Generated message
     */
    message: ChatMessage;
    /**
     * Reason the generation stopped
     */
    finish_reason: 'stop' | 'length' | 'function_call' | 'tool_calls' | 'content_filter';
}
/**
 * Usage statistics for a chat completion
 */
export interface ChatCompletionUsage {
    /**
     * Number of tokens in the prompt
     */
    prompt_tokens: number;
    /**
     * Number of tokens in the completion
     */
    completion_tokens: number;
    /**
     * Total number of tokens used
     */
    total_tokens: number;
}
/**
 * Response from creating a chat completion
 */
export interface ChatCompletionResponse {
    /**
     * Object type
     */
    object: 'chat.completion';
    /**
     * Unique identifier for the completion
     */
    id: string;
    /**
     * Unix timestamp of when the completion was created
     */
    created: number;
    /**
     * ID of the model used
     */
    model: string;
    /**
     * List of completion choices
     */
    choices: ChatCompletionChoice[];
    /**
     * Usage statistics
     */
    usage: ChatCompletionUsage;
    /**
     * System fingerprint
     */
    system_fingerprint?: string;
    /**
     * Metadata about the response
     */
    _metadata?: {
        /**
         * Rate limit information
         */
        rateLimit?: {
            /**
             * Number of requests made in the current period
             */
            limit: number;
            /**
             * Number of requests remaining in the current period
             */
            remaining: number;
            /**
             * Timestamp when the rate limit will reset
             */
            reset: number;
        };
        /**
         * Balance information
         */
        balance?: {
            /**
             * VCU balance
             */
            vcu: number;
            /**
             * USD balance
             */
            usd: number;
        };
    };
}
/**
 * Delta in a streaming chat completion
 */
export interface ChatCompletionChunkDelta {
    /**
     * Role of the message author
     */
    role?: ChatRole;
    /**
     * Content of the message
     */
    content?: string;
    /**
     * Function call
     */
    function_call?: {
        name?: string;
        arguments?: string;
    };
    /**
     * Tool calls
     */
    tool_calls?: Array<{
        index?: number;
        id?: string;
        type?: 'function';
        function?: {
            name?: string;
            arguments?: string;
        };
    }>;
}
/**
 * Choice in a streaming chat completion
 */
export interface ChatCompletionChunkChoice {
    /**
     * Index of the choice
     */
    index: number;
    /**
     * Delta of the message
     */
    delta: ChatCompletionChunkDelta;
    /**
     * Reason the generation stopped
     */
    finish_reason: 'stop' | 'length' | 'function_call' | 'tool_calls' | 'content_filter' | null;
}
/**
 * Chunk in a streaming chat completion
 */
export interface ChatCompletionChunk {
    /**
     * Object type
     */
    object: 'chat.completion.chunk';
    /**
     * Unique identifier for the completion
     */
    id: string;
    /**
     * Unix timestamp of when the completion was created
     */
    created: number;
    /**
     * ID of the model used
     */
    model: string;
    /**
     * List of completion choices
     */
    choices: ChatCompletionChunkChoice[];
    /**
     * System fingerprint
     */
    system_fingerprint?: string;
}