README

本包为您的Laravel应用程序提供与Together AI API的简单集成。

安装

通过Composer安装包

composer require ketchalegend/laravel-together-ai

配置

发布配置文件

php artisan vendor:publish --provider="ketchalegend\LaravelTogetherAI\TogetherAIServiceProvider" --tag="together-ai-config"

然后，将您的Together AI API密钥添加到.env文件中

TOGETHER_AI_API_KEY=your-api-key-here

使用方法

以下是使用Laravel Together AI包的各种示例

use ketchalegend\LaravelTogetherAI\Facades\TogetherAI;

// Basic chat completion
$response = TogetherAI::chat()->create([
    'messages' => [
        ['role' => 'user', 'content' => 'Hello, how are you?']
    ]
]);

// Chat completion with custom parameters
$response = TogetherAI::chat()->create([
    'messages' => [
        ['role' => 'user', 'content' => 'Explain quantum computing in simple terms.']
    ],
    'temperature' => 0.8,
    'max_tokens' => 150
]);

// Chat with system message and multiple user messages
$response = TogetherAI::chat()->create([
    'messages' => [
        ['role' => 'system', 'content' => 'You are a helpful assistant specializing in technology.'],
        ['role' => 'user', 'content' => 'What is cloud computing?'],
        ['role' => 'assistant', 'content' => 'Cloud computing is a technology that allows users to access and use computing resources over the internet instead of on their local computer.'],
        ['role' => 'user', 'content' => 'What are its main benefits?']
    ]
]);

// Streamed chat completion
$stream = TogetherAI::chat()->create([
    'messages' => [
        ['role' => 'user', 'content' => 'Tell me a short story about a robot.']
    ],
    'stream' => true
]);

foreach ($stream as $chunk) {
    if (isset($chunk['done']) && $chunk['done']) {
        echo "Story finished.\n";
        break;
    }
    echo $chunk['choices'][0]['delta']['content'] ?? '';
}

// Chat completion with function calling (if supported by Together AI)
$functions = [
    [
        'name' => 'get_current_weather',
        'description' => 'Get the current weather in a given location',
        'parameters' => [
            'type' => 'object',
            'properties' => [
                'location' => [
                    'type' => 'string',
                    'description' => 'The city and state, e.g. San Francisco, CA',
                ],
                'unit' => ['type' => 'string', 'enum' => ['celsius', 'fahrenheit']],
            ],
            'required' => ['location'],
        ],
    ]
];

$response = TogetherAI::chat()->create([
    'messages' => [
        ['role' => 'user', 'content' => 'What\'s the weather like in New York?']
    ],
    'functions' => $functions,
    'function_call' => 'auto'
]);

// Handle the response
if (isset($response['choices'][0]['function_call'])) {
    $functionCall = $response['choices'][0]['function_call'];
    // Process the function call...
} else {
    echo $response['choices'][0]['message']['content'];
}

// Error handling
try {
    $response = TogetherAI::chat()->create([
        'messages' => [
            ['role' => 'user', 'content' => 'Generate a very long response.']
        ],
        'max_tokens' => 10000 // Assuming this exceeds the API's limit
    ]);
} catch (\Exception $e) {
    echo "An error occurred: " . $e->getMessage();
}

// Using custom headers and base URI   

$client = TogetherAI::factory()
    ->withApiKey($newApiKey)
    ->withBaseUri($newBaseUri)
    ->withHttpHeader('Custom-Header', 'Custom-Value');
    ->make();

$response = $client->chat()->create([
    'messages' => [
        ['role' => 'user', 'content' => 'Hello with custom configuration!']
    ]
]);

这些示例演示了

1. 基本聊天完成
2. 使用自定义参数聊天
3. 系统消息和助手消息的多轮对话
4. 流式响应
5. 函数调用
6. 用户消息的简化输入
7. 错误处理

请记得在您的应用程序中适当地处理API响应和错误。实际响应的结构可能因Together AI API版本和特定模型而异。

许可证

MIT许可证（MIT）。有关更多信息，请参阅许可证文件。