s3 url as input folder

[pythonstuff8]

Search before asking

• I have searched the Autodistill issues and found no similar feature requests.

Question

How do I use my s3 bucket folder url as the input folder variable:

ontology=CaptionOntology({
    f"thing": f"thing"

})
from autodistill_grounded_sam import GroundedSAM

base_model = GroundedSAM(ontology=ontology)
dataset = base_model.label(
    input_folder="s3://bucket/folder/",
    extension=".jpg",
    output_folder="outputf")```

### Additional

this line 'input_folder="s3://bucket/folder",` could also be `input_folder="https://bucket.s3.amazonaws.com/folder/",`

[capjamesg]

Hello there! We don't currently have direct support for running Autodistill on an S3 bucket. With that said, you can download your images from S3 to the computer on which you are running Autodistill with a few commands.

To download data from AWS, first install the awscli command line tool. This tool allows you to interact with your AWS account on the command line. Once you have the command line tool installed, run the following command:

aws s3 sync s3://mybucket/folder_path .

Replace mybucket with the name of your bucket and folder_path with the name of the folder or file you want to export. This comamnd will download an asset from AWS to your current working directory (.).

[pythonstuff8]

thank you, I will wait until support comes ... I also have to files stored as PIL images in a varible can I use those `original_image=original_image.resize((224,224))

rotated_image = original_image.rotate(-30)



rotated_image2 = original_image.rotate(30)



flipped_image = original_image.transpose(Image.FLIP_LEFT_RIGHT)



blurred_image = original_image.filter(ImageFilter.GaussianBlur(radius=2))


enhancer = ImageEnhance.Brightness(original_image)
brightened_image = enhancer.enhance(1.5)


darkened_image = enhancer.enhance(0.5)

noisy_image = add_noise(original_image, noise_factor=50)  # Adjust noise_factor as needed
try:
    inverted_image = ImageOps.invert(original_image)
except:
   pass
grayscale_image = original_image.convert("L")
return original_image,rotated_image,rotated_image2,flipped_image,blurred_image,brightened_image,darkened_image,noisy_image,inverted_image,grayscale_image`

[pythonstuff8]

@capjamesg