Weave Code
Ai Perplexity Platform Laravel Package
symfony/ai-perplexity-platform
Symfony AI bridge for the Perplexity Platform. Provides integration with Perplexity’s Sonar chat completions API for building AI chat experiences in Symfony apps, with links to Perplexity docs and contribution resources.
Perplexity platform bridge for Symfony AI
Frequently asked questions about Ai Perplexity Platform
- Can I use this package directly in Laravel without Symfony’s AI components?
- No, this package requires Symfony’s AI abstractions (e.g., `Message`, `ModelClient`). For Laravel, you’ll need an adapter like `spatie/laravel-symfony-support` or custom facades to bridge Symfony’s HttpClient and AI logic. Alternatively, consider a lightweight Guzzle wrapper for Perplexity’s API.
- How do I handle Perplexity API keys in Laravel’s `.env` with Symfony’s HttpClient?
- Store `PERPLEXITY_API_KEY` in `.env` as usual, but Symfony’s HttpClient may need a custom authentication plugin (e.g., `Bearer` token). Use `symfony/http-client-guzzle` to bridge Laravel’s Guzzle with Symfony’s middleware, or wrap the client in a Laravel service provider.
- Will this work with Laravel’s queue system for streaming responses?
- Symfony’s `DeltaInterface` streams may block Laravel’s synchronous requests. Use queue workers (e.g., Swoole, ReactPHP) or chunked processing to handle streaming responses. Laravel’s default queues won’t natively support Symfony’s streaming format without custom event listeners.
- What Laravel versions support this package, and are there dependency conflicts?
- This package requires PHP 8.2+ and Symfony 7.3+, compatible with Laravel 10+. Older versions (e.g., 9.x) may need dependency overrides (e.g., `ramsey/uuid`) or custom packages. Test thoroughly for conflicts with Laravel’s native HTTP client or DI container.
- How do I switch between Perplexity and other AI providers (e.g., OpenAI) in Laravel?
- Symfony’s `Provider` abstraction supports multi-provider routing, but Laravel needs custom service bindings to expose these as Laravel services. Use a facade or service container alias to dynamically switch providers while maintaining Laravel’s dependency injection.
- Are there performance risks with streaming responses in Laravel?
- Yes, large streaming responses (e.g., multi-turn chats) may exhaust Laravel’s memory or block requests. Offload processing to async workers (Swoole, ReactPHP) or implement chunked streaming with Laravel’s queue system to avoid timeouts or crashes.
- How do I handle Symfony’s API errors in Laravel’s exception system?
- Symfony’s `ApiError` exceptions must be mapped to Laravel’s `HttpResponse` or `ProblemDetail`. Create a custom exception handler in Laravel’s `app/Exceptions/Handler.php` to convert Symfony errors into Laravel-compatible formats for consistent debugging.
- Is this package stable for production, or should I wait for Laravel-native alternatives?
- This is an early-stage package (low adoption) tied to Symfony’s AI ecosystem. For production, consider a custom Guzzle wrapper or `php-ai/perplexity` until the bridge matures. Monitor Perplexity’s API stability and Symfony’s AI updates for breaking changes.
- Can I integrate this with Laravel Nova or Horizon for queue-based AI tasks?
- No native support exists. Use Laravel’s event listeners or custom service providers to bridge Symfony’s AI logic with Nova/Panel or Horizon queues. For example, dispatch Symfony AI tasks to Horizon jobs with `shouldQueue()` for async processing.
- What’s the best way to test this package in Laravel without Symfony’s test suite?
- Mock Symfony’s `ModelClient` and `HttpClient` interfaces in Laravel’s PHPUnit tests. Use Laravel’s `Mockery` or `PHPUnit` to stub Perplexity API responses, focusing on the bridge layer between Symfony and Laravel components. Test edge cases like rate limits or streaming failures.
Weaver
How can I help you explore Laravel packages today?