Today, we’ll cache GPT product recommendations so that users don’t trigger a new API call on every page load. This reduces costs, improves speed, and keeps recommendations fresh for a limited time.
🧠 Step 1: Wrap GPT logic with cache
In ProductRecommendationService, update the recommend() method:
use Illuminate\Support\Facades\Cache;
public function recommend(User $user)
{
return Cache::remember("recommendations_user_{$user->id}", now()->addHours(6), function () use ($user) {
// existing GPT logic...
$viewed = ...;
$clicked = ...;
$messages = [
[
'role' => 'user',
'content' => 'Here is what the user viewed: ' . json_encode($viewed)
],
[
'role' => 'user',
'content' => 'Here is what the user clicked: ' . json_encode($clicked)
],
[
'role' => 'user',
'content' => 'Suggest 5 relevant product ideas based on both viewing and clicking behavior.'
],
];
$function = [
'name' => 'suggest_products',
'description' => 'Suggest relevant products for a user based on viewed and clicked products',
'parameters' => [
'type' => 'object',
'properties' => [
'suggestions' => [
'type' => 'array',
'items' => ['type' => 'string'],
],
],
'required' => ['suggestions'],
],
];
$response = OpenAI::chat()->create([
'model' => 'gpt-4-0613',
'messages' => $messages,
'functions' => [$function],
'function_call' => ['name' => 'suggest_products'],
]);
$arguments = json_decode(
$response['choices'][0]['message']['function_call']['arguments'] ?? '{}',
true
);
$suggestions = $arguments['suggestions'] ?? [];
return Product::query()
->where(function ($query) use ($suggestions) {
foreach ($suggestions as $term) {
$query->orWhere('name', 'LIKE', "%$term%");
}
})
->limit(5)
->get();
});
}
🔁 Step 2: Invalidate cache when needed
In your controller or view route, you can optionally allow forced refresh:
Route::get('/recommendations', function (Request $request, ProductRecommendationService $service) {
$user = \App\Models\User::first(); // demo
if ($request->get('refresh')) {
Cache::forget("recommendations_user_{$user->id}");
}
$products = $service->recommend($user);
return view('recommendations.index', compact('products'));
});
Now add ?refresh=1 to force-refresh GPT recommendations.
✅ Summary
✅ Today you:
- Cached GPT product recommendations using Laravel’s Cache
- Reduced OpenAI API usage for repeated page loads
- Enabled cache expiry and manual refresh
✅ Up next (Day 9): We’ll create a dashboard chart showing top clicked GPT-recommended products across all users for analytics and trends.
