Automatic image blob creation doesn't handle P images with JPEG

[BluePigman]

Description of the bug:

Currently P mode images cannot be uploaded to Gemini. https://pillow.readthedocs.io/en/stable/handbook/concepts.html#modes

Actual vs expected behavior:

import requests
from PIL import Image
import google.generativeai as genai

genai.configure(api_key)

media_url = "https://cdn.betterttv.net/emote/5f5a87416084af6c1719f9b4/3x"   

image = Image.open(requests.get(media_url, stream=True).raw)
response = genai.GenerativeModel("gemini-1.5-flash").generate_content(["Describe this image.", image])
description = response.text.replace('\n', ' ')
print(description)

Actual: Running this code gives the errors: KeyError: 'P' OSError: cannot write mode P as JPEG

Expected: If we add the following line before generating the response

image = image.convert("RGB")

The code runs with no errors.

Any other information you'd like to share?

This is similar to this issue solved previously: https://github.com/google-gemini/generative-ai-python/issues/160

[manojssmk]

I tried to run your code on Colab and faced the same issue Please find the gist here for reference. Thanks!

[BluePigman]

I tried to run your code on Colab and faced the same issue Please find the gist here for reference. Thanks!

You leaked your API key, just letting you know.

[MarkDaoust]

Ah, I see, thanks for reporting.

It looks like my first fix was junk. It looks like I can probably use img.get_format_mimetype() to resolve this for all supported mime-types.

[MarkDaoust]

@BluePigman @manojssmk please have a look at the fix in #523 and LMK if you have any feedback.

[BluePigman]

Looks good to me!

[MarkDaoust]

Fixed.