Request for Auto-Verification of Images in Perplexity API

[dpatelTaskMe]

Request for Auto-Verification of Images in Perplexity API

Dear Team Perplexity,

I hope this message finds you well. As a verified user of the Perplexity API, I’ve noticed that the images provided in the API responses still require manual verification. This creates friction and negatively impacts the user experience on our platform.

Is there any way to enable automatic verification for images retrieved through the API? Resolving this would greatly streamline our workflow and enhance the overall user experience.

Looking forward to your assistance or guidance on this matter.

Best regards,
Dhruvish Patel
Python Developer
Email: [email protected]

[drodrig]

This would be a nice addition. However, I suspect the overhead on Perplexity's side would make it unlikely. You could test yourself, however.

import requests

def validate_images(data):
    """
    Checks if images exist in the given data. If an image does not exist, 
    its 'image_url' is set to None.
    
    :param data: Dictionary containing an 'images' key with a list of image objects.
    :return: The modified dictionary with invalid images set to None.
    """
    for item in data.get("images", []):  # Loop through the list of images
        image_url = item.get("image_url")
        if image_url:
            try:
                response = requests.head(image_url, timeout=2)  # Perform a HEAD request
                if response.status_code != 200:  # If not found, set image_url to None
                    item["image_url"] = None
            except requests.RequestException:  # Handles connection errors, timeouts, etc.
                item["image_url"] = None
    
    return data  # Return updated data

# Example usage
example_data = {
    "images": [
        {"height": 525, "image_url": "https://images.trvl-media.com/lodging/39000000/38270000/38266000/38265913/adbca281.jpg?impolicy=resizecrop&rw=575&rh=575&ra=fill"},
        {"height": 400, "image_url": "https://cf.bstatic.com/xdata/images/hotel/max1024x768/invalid_image.jpg"},
    ]
}

updated_data = validate_images(example_data)
print(updated_data)

The above is untested. Also, you could potentially add quite a few seconds to the reply.