imageLoader.py

import os
import random
from fastdownload import download_url
from duckduckgo_search import ddg_images
from fastcore.all import *
from fastai.vision.all import *

def search_images(term, max_images=30):
    print(f"Searching for '{term}'")
    return L(ddg_images(term, max_results=max_images)).itemgot('image')

def download_images(term, folder, num_images):
    # Erstelle den Ordner, falls er nicht existiert
    if not os.path.exists(folder):
        os.makedirs(folder)

    # Suche nach den Bild-URLs
    urls = search_images(term, max_images=num_images)

    # Lade jedes Bild herunter und speichere es im Ordner
    for i, url in enumerate(urls):
        filename = f"{i+1:04}.jpg"
        dest = os.path.join(folder, filename)
        try:
            download_url(url, dest, show_progress=False)
            print(f"Downloaded image {i+1}/{num_images} to {dest}")
        except Exception:
                print(f"Error trying to load Image {urls[i]}, trying next...")

    print(f"Downloaded {num_images} images to folder {folder}.")

def cleanup_images():
    for path, subdirs, files in os.walk('data'):
        for directory in subdirs:
            path = Path(directory)
            failed = verify_images(get_image_files(path))
            failed.map(Path.unlink)
            print("Number of images in folder '",path,"' failed: ",len(failed))

def prepare_images():
    for path, subdirs, files in os.walk('data'):
        for directory in subdirs:
            print("Resizing images in dir: ",directory)
            resize_images(directory, max_size=500, dest=directory)

# Beispielaufruf
search_term = 'cumulonimbus cloud'
folder_name = 'cumulonimbus'
num_images = 250

download_images(search_term, folder_name, num_images)
prepare_images()
cleanup_images()