Fields Ai Laravel Package

Getting Started

Minimal Setup

1. Installation

   composer require broqit/fields-ai
   

   Publish the config file (if available):

   php artisan vendor:publish --provider="Broqit\FieldsAI\FieldsAIServiceProvider"
   

2. Basic Usage Inject the FieldsAI facade into your controller or service:

   use Broqit\FieldsAI\Facades\FieldsAI;
   
   // Generate a field description
   $description = FieldsAI::describeField('user', 'email');
   

3. First Use Case Use in a form builder to auto-generate field labels, placeholders, or validation rules:

   $label = FieldsAI::generateLabel('user', 'email'); // "User Email Address"
   $placeholder = FieldsAI::generatePlaceholder('user', 'email'); // "example@domain.com"
   

Implementation Patterns

Common Workflows

1. Dynamic Form Field Generation

   // In a FormRequest or FormService
   public function rules()
   {
       $rules = [];
       $fields = ['name', 'email', 'bio'];
   
       foreach ($fields as $field) {
           $rules[$field] = FieldsAI::inferValidation($field);
       }
   
       return $rules;
   }
   

2. Localization Support

   // Generate labels in multiple languages
   $labels = [
       'en' => FieldsAI::generateLabel('product', 'price', 'en'),
       'es' => FieldsAI::generateLabel('product', 'price', 'es'),
   ];
   

3. Integration with Laravel Blade

   @foreach($formFields as $field)
       <div>
           <label>{{ FieldsAI::generateLabel($model, $field) }}</label>
           <input type="text" name="{{ $field }}">
           <span class="help-text">{{ FieldsAI::generateHelpText($model, $field) }}</span>
       </div>
   @endforeach
   

4. API Response Enhancement

   // Auto-generate API response metadata
   $response = [
       'data' => $user,
       'fields' => [
           'email' => FieldsAI::describeField('user', 'email'),
           'created_at' => FieldsAI::describeField('user', 'created_at'),
       ],
   ];
   

5. Model Casting

   // Auto-generate attribute descriptions for API docs
   class User extends Model
   {
       public function getAttributeDescriptions()
       {
           return collect($this->attributes)
               ->mapWithKeys(fn ($value, $key) => [
                   $key => FieldsAI::describeField('user', $key)
               ]);
       }
   }
   

Gotchas and Tips

Pitfalls

1. Rate Limiting

   • The package may hit API rate limits if used excessively in bulk operations (e.g., generating labels for 100+ fields at once).
   • Solution: Cache responses aggressively:

     $label = Cache::remember("fields_ai_{$model}_{$field}", now()->addHours(1), function() use ($model, $field) {
         return FieldsAI::generateLabel($model, $field);
     });
     

2. Model/Field Mismatches

   • The AI may generate incorrect descriptions if the model or field names are ambiguous (e.g., user vs. users).
   • Solution: Use fully qualified names:

     FieldsAI::describeField('App\Models\User', 'email'); // Explicit namespace
     

3. Contextual Overrides

   • AI-generated descriptions may not account for custom business logic (e.g., a password field that’s actually a "PIN").
   • Solution: Extend the package via service providers:

     // app/Providers/FieldsAIServiceProvider.php
     public function boot()
     {
         FieldsAI::extend('user', 'pin', function ($model, $field) {
             return "4-digit security code (e.g., 1234)";
         });
     }
     

4. Configuration Quirks

   • The fields-ai config may not be published by default. Check for:

     config('fields-ai.api_key'); // Ensure this is set
     config('fields-ai.cache_enabled'); // Toggle caching
     

Debugging Tips

1. Enable Verbose Logging Add to config/fields-ai.php:

   'debug' => env('FIELDS_AI_DEBUG', false),
   

   Then check storage/logs/laravel.log for AI response details.

2. Mocking for Testing Use the FieldsAI facade’s mocking methods:

   FieldsAI::shouldReceive('generateLabel')
           ->once()
           ->with('user', 'email')
           ->andReturn('Test Email');
   
   $this->assertEquals('Test Email', FieldsAI::generateLabel('user', 'email'));
   

3. Fallback Logic Provide defaults for when the AI fails:

   $label = FieldsAI::generateLabel('user', 'email', [
       'fallback' => ucfirst(str_replace('_', ' ', $field)),
   ]);
   

Extension Points

1. Custom Prompts Override the default AI prompts via config:

   'prompts' => [
       'label' => 'Generate a user-friendly label for the {field} field in a {model} resource.',
       'validation' => 'Suggest Laravel validation rules for the {field} field in a {model}.',
   ],
   

2. Field Type Hints Pass additional context to improve accuracy:

   FieldsAI::describeField('order', 'status', [
       'type' => 'enum',
       'options' => ['pending', 'shipped', 'delivered'],
   ]);
   

3. Batch Processing Use the batch() method for efficiency:

   $descriptions = FieldsAI::batch(['user' => ['email', 'password']]);
   // Returns: ['user.email' => "...", 'user.password' => "..."]