import type { HttpClient } from "@typeDefs";
import { HttpModelProvider, InferHttpClientOptions } from "../http";
import type { MistralAuthConfig } from "./authConfig";
/**
 * Creates a Mistral {@link ModelProvider} with the {@link MistralAiApi}
 *
 * ```ts
 * import { createMistralModelProvider } from "generative-ts";
 *
 * const mistralLarge = createMistralModelProvider({
 *   modelId: "mistral-large-latest", // Mistral defined model ID
 *   // you can explicitly pass auth here, otherwise by default it is read from process.env
 * });
 *
 * const response = await mistralLarge.sendRequest({
 *   $prompt: "Brief History of NY Mets:"
 *   // all other Mistral ChatCompletion API options available here
 * });
 *
 * console.log(response.choices[0]?.message.content);
 * ```
 *
 * ### Provider Setup and Notes
 *
 * Create an API account at {@link https://mistral.ai | Mistral}
 *
 * Obtain a {@link https://console.mistral.ai/api-keys/ | Mistral API key} and either pass them explicitly in `auth` or set them in the environment as `MISTRAL_API_KEY`
 *
 * ### Model Parameters
 *
 * - {@link https://docs.mistral.ai/api/#operation/createChatCompletion | Mistral ChatCompletion API}
 *
 * ### Model IDs
 *
 * - {@link https://docs.mistral.ai/getting-started/models/ | Mistral Models}
 *
 * @see {@link https://console.mistral.ai/api-keys/ | Mistral API keys}
 * @see {@link https://docs.mistral.ai/api/#operation/createChatCompletion | Mistral ChatCompletion API}
 * @see {@link https://docs.mistral.ai/getting-started/models/ | Mistral Models}
 *
 * @category Providers
 * @category Provider: Mistral
 *
 * @param {Object} params
 * @param {string} params.modelId - The model ID as defined by Mistral.
 * @param {HttpClient} [params.client] - HTTP client to use for requests. If not supplied, the built-in fetch-based implementation will be used.
 * @param {MistralAuthConfig} [params.auth] - Authentication configuration for Mistral. If not supplied, it will be loaded from the environment.
 * @returns {HttpModelProvider<MistralGenerateOptions, MistralGenerateResponse, BaseModelProviderConfig>} The Mistral Model Provider with the {@link MistralAiApi}
 * @throws {Error} If no auth is passed and MISTRAL_API_KEY is not found in process.env
 *
 * @example Usage
 * ```ts
 * import { createMistralModelProvider } from "generative-ts";
 *
 * const mistralLarge = createMistralModelProvider({
 *   modelId: "mistral-large-latest", // Mistral defined model ID
 *   // you can explicitly pass auth here, otherwise by default it is read from process.env
 * });
 *
 * const response = await mistralLarge.sendRequest({
 *   $prompt: "Brief History of NY Mets:"
 *   // all other Mistral ChatCompletion API options available here
 * });
 *
 * console.log(response.choices[0]?.message.content);
 * ```
 */
export declare function createMistralModelProvider<THttpClientOptions = InferHttpClientOptions<HttpModelProvider>>({ modelId, client, auth, }: {
    modelId: string;
    client?: HttpClient<THttpClientOptions>;
    auth?: MistralAuthConfig;
}): HttpModelProvider<import("../../apis/mistral").MistralAiOptions, import("../../apis/mistral").MistralAiResponse, THttpClientOptions, {
    modelId: string;
}>;