Mistral API Wrapper for PHP

This PHP library provides a simple wrapper for the Mistral API, allowing you to easily integrate the Mistral API into your PHP projects.

Features

• Easy integration with Mistral API
• Supports all Mistral API endpoints
• Streaming support for real-time responses in chat completions
• Utilizes PSR-17 and PSR-18 compliant HTTP clients and factories for making API requests

Requirements

• PHP 8.1 or higher
• A PSR-17 HTTP Factory implementation (e.g., guzzle/psr7 or nyholm/psr7)
• A PSR-18 HTTP Client implementation (e.g., guzzlehttp/guzzle or symfony/http-client)

Installation

You can install the library via Composer:

composer require softcreatr/php-mistral-ai-sdk

Usage

First, include the library in your project:

<?php

require_once 'vendor/autoload.php';

Then, create an instance of the MistralAI class with your API key, organization (optional), an HTTP client, an HTTP request factory, and an HTTP stream factory:

use SoftCreatR\MistralAI\MistralAI;

$apiKey = 'your_api_key';

// Replace these lines with your chosen PSR-17 and PSR-18 compatible HTTP client and factories
$httpClient = new YourChosenHttpClient();
$requestFactory = new YourChosenRequestFactory();
$streamFactory = new YourChosenStreamFactory();
$uriFactory = new YourChosenUriFactory();

$mistral = new MistralAI($requestFactory, $streamFactory, $uriFactory, $httpClient, $apiKey);

Now you can call any supported MistralAI API endpoint using the magic method __call:

$response = $mistral->createChatCompletion([
    'model' => 'mistral-tiny',
    'messages' => [
        [
            'role' => 'user',
            'content' => 'Who is the most renowned French painter?'
        ],
    ],
]);

// Process the API response
if ($response->getStatusCode() === 200) {
    $responseObj = json_decode($response->getBody()->getContents(), true);
    
    print_r($responseObj);
} else {
    echo "Error: " . $response->getStatusCode();
}

Streaming Example

You can enable real-time streaming for chat completions:

$streamCallback = static function ($data) {
    if (isset($data['choices'][0]['delta']['content'])) {
        echo $data['choices'][0]['delta']['content'];
    }
};

$mistral->createChatCompletion([
    'model' => 'mistral-small-latest',
    'messages' => [
        [
            'role' => 'user',
            'content' => 'Tell me a story about a brave knight.',
        ],
    ],
    'stream' => true,
], $streamCallback);

For more details on how to use each endpoint, refer to the Mistral API documentation, and the examples provided in the repository.

Supported Methods

Chat Completions

• Create Chat Completion - Example
  • createChatCompletion(array $options = [], callable $streamCallback = null)

Embeddings

• Create Embedding - Example
  • createEmbedding(array $options = [])

Models

• List Models - Example
  • listModels()
• Retrieve Model - Example
  • retrieveModel(array $parameters = [])
• Delete Model - Example
  • deleteModel(array $parameters = [])
• Update Fine-Tuned Model - Example
  • updateFineTunedModel(array $parameters = [])
• Archive Model - Example
  • archiveModel(array $parameters = [])
• Unarchive Model - Example
  • unarchiveModel(array $parameters = [])

Files

• Upload File - Example
  • uploadFile(array $options = [])
• List Files - Example
  • listFiles()
• Retrieve File - Example
  • retrieveFile(array $parameters = [])
• Delete File - Example
  • deleteFile(array $parameters = [])

Fine-Tuning Jobs

• List Fine-Tuning Jobs - Example
  • listFineTuningJobs()
• Retrieve Fine-Tuning Job - Example
  • retrieveFineTuningJob(array $parameters = [])
• Cancel Fine-Tuning Job - Example
  • cancelFineTuningJob(array $parameters = [])
• Start Fine-Tuning Job - Example
  • startFineTuningJob(array $parameters = [])
• Create Fine-Tuning Job - Example
  • createFineTuningJob(array $options = [])

FIM Completion

• Create FIM Completion - Example
  • createFimCompletion(array $options = [])

Agents Completion

• Create Agents Completion - Example
  • createAgentsCompletion(array $options = [])

Changelog

For a detailed list of changes and updates, please refer to the CHANGELOG.md file. We adhere to Semantic Versioning and document notable changes for each release.

Known Problems and Limitations

Streaming Support

Streaming is now supported for real-time token generation in chat completions. Please make sure you are handling streams correctly using a callback, as demonstrated in the examples.

License

This library is licensed under the ISC License. See the LICENSE file for more information.

Maintainers 🛠️

Sascha Greuel

Contributors ✨