overview-upload

#!/usr/bin/env python3
#
# Upload a file, or directory, to an Overview server. 
# From https://github.com/overview/overview-upload-directory

import argparse
import logging
import os
import pathlib
from overview_upload import Upload, create_document_set

# ---- Main ----

def main():
    parser = argparse.ArgumentParser(description='Upload a file or directory to an Overview server.')
    parser.add_argument('file', help='file or directory to upload')
    parser.add_argument('-t', '--token', help='API token corresponding to document set', required=True)
    parser.add_argument('-s', '--server', help='url of Overview server, defaults to http://localhost:9000', default='http://localhost:9000')
    parser.add_argument('--skip-duplicate', dest='skip_duplicate', help='Skip files already on the server', action="store_true")
    parser.add_argument('-n', '--noskip', dest='skip_duplicate', help='Don\'t skip files already on server', action="store_false")
    parser.add_argument('--split-by-page', dest='split_by_page', help='Turn input files into one document per page', action="store_true", default=False)
    parser.add_argument('--lang', dest='lang', default='en', help='2-char ISO document language code (used for OCR and text analysis)')

    group = parser.add_mutually_exclusive_group()
    group.add_argument('--ocr', dest='ocr', help='Run OCR on PDF pages that are only images', action="store_true")
    group.add_argument('--no-ocr', dest='ocr', help='Skip OCR always (for speed)', action="store_false")

    parser.add_argument('--create-document-set-with-title', dest='create_with_title', help='Create a new document set and then add files')
    parser.set_defaults(ocr=True, skip_duplicate=True)
    args = parser.parse_args()

    filename = args.file

    logger = logging.getLogger("overview-upload")
    logger.setLevel(logging.DEBUG)
    logger.addHandler(logging.StreamHandler())

    if not os.path.exists(filename):
        print("Cannot find file or directory " + filename)
    else:
        if args.create_with_title:
            response = create_document_set(args.server, args.token, args.create_with_title, logger=logger)
            logger.info('Created document set "{}" with ID {}', args.create_with_title, response['documentSet']['id'])
            api_token = response['apiToken']['token']
        else:
            api_token = args.token

        upload = Upload(args.server, api_token, logger=logger)
        upload.clear_previous_upload()

        upload_kwargs = {
            'skip_unhandled_extension': True,
            'skip_duplicate': args.skip_duplicate,
        }

        if os.path.isdir(filename):
            # Send a directory.
            upload.send_directory(filename, **upload_kwargs)
        else:
            # Send a single file.
            # use a basename on the server -- no directories
            path = pathlib.Path(filename)
            upload.send_path_if_conditions_met(path, filename=path.name, **upload_kwargs)

        upload.finish(
            ocr=args.ocr,
            split_by_page=args.split_by_page,
            lang=args.lang
        )

if __name__ == '__main__':
    main()