README

Modelflow AI
Anthropic Adapter

Note: This is part of the modelflow-ai project create issues in the main repository.

Note: This project is heavily under development and any feedback is greatly appreciated.

Installation

To install the Anthropic Adapter package, you need to have PHP 8.2 or higher and Composer installed on your machine. Then, you can add the package to your project by running the following command:

composer require modelflow-ai/anthropic-adapter

Examples

Here are some examples of how you can use the Anthropic Adapter in your PHP applications. You can find more detailed examples in the examples directory.

Usage

First, initialize the client:

use ModelflowAi\Anthropic\Anthropic;

$client = Anthropic::client('your-api-key');

Then, you can use the AnthropicChatModelAdapter:

use ModelflowAi\Chat\Adapter\AIChatAdapterInterface;
use ModelflowAi\Chat\AIChatRequestHandler;
use ModelflowAi\Chat\Request\AIChatRequest;
use ModelflowAi\Chat\Request\Message\AIChatMessage;
use ModelflowAi\Chat\Request\Message\AIChatMessageRoleEnum;
use ModelflowAi\DecisionTree\DecisionTree;
use ModelflowAi\DecisionTree\Criteria\CapabilityCriteria;
use ModelflowAi\DecisionTree\DecisionRule;
use ModelflowAi\Anthropic\Model;
use ModelflowAi\AnthropicAdapter\Chat\AnthropicChatAdapter;
use ModelflowAi\PromptTemplate\ChatPromptTemplate;

$modelAdapter = new AnthropicChatAdapter($client, Model::CLAUDE_3_HAIKU);

/** @var DecisionTreeInterface<AIChatRequest, AIChatAdapterInterface> $decisionTree */
$decisionTree = new DecisionTree([
    new DecisionRule($modelAdapter, [CapabilityCriteria::SMART]),
]);
$handler = new AIChatRequestHandler($decisionTree);

$response = $handler->createRequest(
    ...ChatPromptTemplate::create(
        new AIChatMessage(AIChatMessageRoleEnum::SYSTEM, 'You are an {feeling} bot'),
        new AIChatMessage(AIChatMessageRoleEnum::USER, 'Hello {where}!'),
    )->format(['where' => 'world', 'feeling' => 'angry']),
)
    ->addCriteria(CapabilityCriteria::SMART)
    ->build()
    ->execute();

echo \sprintf('%s: %s', $response->getMessage()->role->value, $response->getMessage()->content);

Contributing

Contributions are welcome. Please open an issue or submit a pull request in the main repository at https://github.com/modelflow-ai/.github.

License

This project is licensed under the MIT License. For the full copyright and license information, please view the LICENSE file that was distributed with this source code.