[Test]: Re-use code

[haesleinhuepf]

Copy over the create function from here: https://github.com/haesleinhuepf/darth-d/blob/main/darth_d/_create.py

[haesleinhuepf]

git-bob comment

[github-actions]

^{This message was generated by git-bob (version: 0.11.2dev, model: claude-3-5-sonnet-20241022, log), an experimental AI-based assistant. It can make mistakes and has limitations. Check its messages carefully.}

It seems there was a request to re-use the create function. I'll provide a Jupyter notebook with the function copied over:

%%notebook

from stackview import jupyter_displayable_output

@jupyter_displayable_output(library_name='darth-d', help_url='https://github.com/haesleinhuepf/darth-d')
def create(prompt:str=None, image_width:int=1024, image_height:int=1024, num_images:int=1, model:str="dall-e-3", style:str='vivid', quality:str='standard'):
    """Create an image from scratch using OpenAI's DALL-E 2 or 3.

    Parameters
    ----------
    prompt: str, text explaining what the image should show
    num_images: int, optional
        ignored for dall-e-3
    model: str, optional
        "dall-e-2", "dall-e-3"
    image_width: int, optional
        must be 256, 512 or 1024 for dall-e-2 or 1024, 1792 for dall-e-3
    image_height: int, optional
        must be 256, 512 or 1024 for dall-e-2 or 1024, 1792 for dall-e-3
    style: str, optional
        "vivid" or "natural", ignored for dall-e-2
    quality: str, optional
        "standard" or "hd", ignored for dall-e-2
    """
    from openai import OpenAI
    from ._utilities import images_from_url_responses

    client = OpenAI()

    size_str = f"{image_width}x{image_height}"

    kwargs = {}
    if model == "dall-e-3":
        kwargs['style'] = style
        kwargs['quality'] = quality

    response = client.images.generate(
      prompt=prompt,
      n=num_images,
      model=model,
      size=size_str,
        **kwargs
    )

    return images_from_url_responses(response)

Note: You'll need to implement the images_from_url_responses function or import it from the appropriate location to make this work.

[haesleinhuepf]

Ok cool. Take over the code of the `images_from_url_responses function from https://github.com/haesleinhuepf/darth-d/blob/main/darth_d/_utilities.py

git-bob comment

[github-actions]

^{This message was generated by git-bob (version: 0.11.2dev, model: claude-3-5-sonnet-20241022, log), an experimental AI-based assistant. It can make mistakes and has limitations. Check its messages carefully.}

I'll update the notebook to include both functions:

%%notebook

from stackview import jupyter_displayable_output
import numpy as np
from skimage.io import imread
from skimage import transform
from PIL import Image
import io

def numpy_to_bytestream(data):
    """Turn a NumPy array into a bytestream"""
    image = Image.fromarray(data.astype(np.uint8)).convert("RGBA")
    bytes_io = io.BytesIO()
    image.save(bytes_io, format='PNG')
    bytes_io.seek(0)
    return bytes_io.read()

def images_from_url_responses(response, input_shape = None):
    """Turns a list of OpenAI's URL responses into numpy images"""
    images = [imread(item.url) for item in response.data]

    if input_shape is not None:
        images = [transform.resize(image, input_shape, anti_aliasing=True, preserve_range=True).astype(image.dtype) for image in images]
        if len(input_shape) == 2 and len(images[0].shape) == 3:
            images = [image[:,:,0] for image in images]

    if len(images) == 1:
        return images[0]
    else:
        return np.asarray(images)

@jupyter_displayable_output(library_name='darth-d', help_url='https://github.com/haesleinhuepf/darth-d')
def create(prompt:str=None, image_width:int=1024, image_height:int=1024, num_images:int=1, model:str="dall-e-3", style:str='vivid', quality:str='standard'):
    """Create an image from scratch using OpenAI's DALL-E 2 or 3."""
    from openai import OpenAI

    client = OpenAI()
    size_str = f"{image_width}x{image_height}"

    kwargs = {}
    if model == "dall-e-3":
        kwargs['style'] = style
        kwargs['quality'] = quality

    response = client.images.generate(
      prompt=prompt,
      n=num_images,
      model=model,
      size=size_str,
        **kwargs
    )

    return images_from_url_responses(response)

All necessary functions are now included in one notebook. I also added the required imports at the top.

[haesleinhuepf]

Do not write a notebook but plain Python code instead. Furthermore, do not use numpy arrays as images, but PIL image instead.

git-bob comment on how to implement this.

[github-actions]

^{This message was generated by git-bob (version: 0.11.2dev, model: claude-3-5-sonnet-20241022, log), an experimental AI-based assistant. It can make mistakes and has limitations. Check its messages carefully.}

Here's how we can modify the code to use PIL images instead of numpy arrays:

from PIL import Image
import io
import requests

def images_from_url_responses(response, input_shape=None):
    """Turns a list of OpenAI's URL responses into PIL images"""
    images = []
    for item in response.data:
        response = requests.get(item.url)
        img = Image.open(io.BytesIO(response.content))
        if input_shape:
            img = img.resize(input_shape)
        images.append(img)
    
    return images[0] if len(images) == 1 else images

def create(prompt:str=None, image_width:int=1024, image_height:int=1024, num_images:int=1, 
          model:str="dall-e-3", style:str='vivid', quality:str='standard'):
    """Create an image using OpenAI's DALL-E"""
    from openai import OpenAI

    client = OpenAI()
    size_str = f"{image_width}x{image_height}"

    kwargs = {}
    if model == "dall-e-3":
        kwargs['style'] = style
        kwargs['quality'] = quality

    response = client.images.generate(
        prompt=prompt,
        n=num_images,
        model=model,
        size=size_str,
        **kwargs
    )

    return images_from_url_responses(response)

The main changes are in images_from_url_responses where we now use PIL's Image directly instead of converting between numpy arrays.