train.py

import os
import slideio
from PIL import Image


####################Convert SVS files to JPEG files####################
# Set the path to the svs files
svs_path = "/Users/josedelgado/Downloads/"

# Set the path to the output directory
output_path = "/Users/josedelgado/Documents/GitHub/microbiology/Images"

# Set the size of the representative image
size = (224, 224)

# Loop through the svs files
for root, dirs, files in os.walk(svs_path):
    for filename in files:
        if filename.endswith(".svs"):
            print("processing " + os.path.join(root, filename)  )
            # Open the svs file
            slide = slideio.open_slide(os.path.join(root, filename))

            # Read a representative image from the slide
            scene = slide.get_scene(0)
            print("region size: " + str(scene.size))

            # Convert the image to RGB format
            image = scene.read_block(size=(500,0))

            # Set the output filename
            output_filename = os.path.splitext(filename)[0] + ".jpeg"

            # Set the output directory
            output_dir = os.path.join(output_path, os.path.relpath(root, svs_path))

            # Create the output directory if it doesn't exist
            if not os.path.exists(output_dir):
                os.makedirs(output_dir)

            # Convert the numpy array to a PIL image
            image = Image.fromarray(image)

            # Save the image in the ImageNet file folder format
            output_file = os.path.join(output_dir, output_filename)
            image.save(output_file)