app.py

import os
import requests
import shutil
import gradio as gr
from concurrent.futures import ThreadPoolExecutor
from zipfile import ZipFile
import logging

logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')

IIIF_URL = "https://lbiiif.riksarkivet.se" #"https://iiifintern.ra.se"

def get_image_ids(batch_id: str) -> list[str]:
    """A list of image IDs in the given batch"""
    logging.info(f"Fetching image IDs for batch {batch_id}")
    response = requests.get(f"{IIIF_URL}/arkis!{batch_id}/manifest")
    response.raise_for_status()
    response = response.json()
    image_ids = [item["id"].split("!")[1][:14] for item in response["items"]]
    logging.info(f"Found {len(image_ids)} images in batch {batch_id}")
    return image_ids

def download_image(url: str, dest: str) -> None:
    """
    Download an image
    Arguments:
        url: Image url
        dest: Destination file name
    """
    logging.info(f"Downloading image from {url} to {dest}")
    response = requests.get(url, stream=True)
    with open(dest, "wb") as out_file:
        shutil.copyfileobj(response.raw, out_file)
    del response

def download_image_by_image_id(image_id: str):
    """
    Download the image with the given image ID
    Creates a directory named after the batch ID and saves the image in
    that directory.
    """
    batch_id = image_id[:8]
    os.makedirs(batch_id, exist_ok=True)
    url = f"{IIIF_URL}/arkis!{image_id}/full/max/0/default.jpg"
    dest = os.path.join(batch_id, image_id + ".jpg")
    download_image(url, dest)

def download_batch_images(batch_id: str, workers: int = 2, progress=None):
    logging.info(f"Starting download for batch {batch_id}")
    image_ids = get_image_ids(batch_id)
    total_images = len(image_ids)

    if progress:
        progress(0, desc=f"Starting download for {batch_id}...")

    def track_download(image_id):
        download_image_by_image_id(image_id)
        logging.info(f"Downloaded image {image_id}")
        if progress:
            # Update progress after each image
            current_progress = (image_ids.index(image_id) + 1) / total_images
            progress(current_progress, desc=f"Downloading {image_id}...")

    with ThreadPoolExecutor(max_workers=workers) as executor:
        for image_id in image_ids:
            executor.submit(track_download, image_id)

    logging.info(f"Zipping downloaded images for batch {batch_id}")
    zip_filename = f"{batch_id}.zip"
    with ZipFile(zip_filename, 'w') as zipf:
        for image_id in image_ids:
            img_path = os.path.join(batch_id, f"{image_id}.jpg")
            zipf.write(img_path, arcname=os.path.basename(img_path))

    if progress:
        progress(1, desc=f"Completed {batch_id}")

    logging.info(f"Completed download and zip for batch {batch_id}")
    return zip_filename


def gradio_interface(batch_ids_input, progress=gr.Progress()):
    batch_ids = [batch_id.strip() for batch_id in batch_ids_input.split("\n") if batch_id.strip()]
    
    zip_files = []
    try:
        for batch_id in progress.tqdm(batch_ids, desc="Processing batches"):
            logging.info(f"Processing batch {batch_id}")
            zip_file = download_batch_images(batch_id, progress=progress)
            zip_files.append(zip_file)
        return zip_files  # Return the list of zip files for download
    except Exception as e:
        logging.error(f"Error processing batches: {e}")
        return str(e)

with gr.Blocks() as app:
    gr.Markdown("# Batch Image Downloader")

    with gr.Row():
        with gr.Column():
            batch_ids_input = gr.Textbox(label="Batch IDs (one per line)", placeholder="Enter batch IDs, one per line.")
            download_button = gr.Button("Download Images")
        with gr.Column():
            output_files = gr.File(label="Download Zip Files", file_count="multiple")


    download_button.click(
        gradio_interface,
        inputs=[batch_ids_input],
        outputs=[output_files]
    )

app.queue()
app.launch()