Mercurial > repos > iuc > data_manager_gtdbtk_database_installer
view data_manager/gtdbtk_database_installer.py @ 1:2814c058a087 draft
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_gtdbtk_database_installer commit 196c4976945b02d23227836ae9cf70a9f44eed59
author | iuc |
---|---|
date | Tue, 03 Jan 2023 09:05:09 +0000 |
parents | 629464b96c2e |
children | 6ab422fba1a3 |
line wrap: on
line source
#!/usr/bin/env python import argparse import json import os import shutil import sys import tarfile from urllib.parse import urlparse from urllib.request import Request from urllib.request import urlopen def url_download(url, target_directory): url_parts = urlparse(url) tarball = os.path.abspath(os.path.join(target_directory, os.path.basename(url_parts.path))) src = None dst = None try: req = Request(url) src = urlopen(req) with open(tarball, 'wb') as dst: while True: chunk = src.read(2**10) if chunk: dst.write(chunk) else: break except Exception as e: sys.exit(str(e)) finally: if src is not None: src.close() if tarfile.is_tarfile(tarball): fh = tarfile.open(tarball, 'r:*') else: return tarball fh.extractall(target_directory) fh.close() os.remove(tarball) # The tarball extraction will create a directory named # something like release202 in the target_directory, so # we need to move the items in that directory to the # target directory. subdir = next(os.walk(target_directory))[1][0] subdir_path = os.path.join(target_directory, subdir) items = os.listdir(subdir_path) for item in items: item_path = os.path.join(subdir_path, item) shutil.move(item_path, target_directory) os.rmdir(subdir_path) return target_directory def download(database_id, database_name, url, out_file): with open(out_file) as fh: params = json.load(fh) target_directory = params['output_data'][0]['extra_files_path'] os.makedirs(target_directory) file_path = url_download(url, target_directory) data_manager_json = {"data_tables": {}} data_manager_entry = {} data_manager_entry['value'] = database_id data_manager_entry['name'] = database_name data_manager_entry['path'] = file_path data_manager_json["data_tables"]["gtdbtk_database"] = data_manager_entry with open(out_file, 'w') as fh: json.dump(data_manager_json, fh, sort_keys=True) parser = argparse.ArgumentParser() parser.add_argument('--database_name', dest='database_name', help='GTDB-Tk database display name') parser.add_argument('--database_id', dest='database_id', help='Unique GTDB-Tk database id') parser.add_argument('--url', dest='url', help='URL to download GTDB-Tk databse version') parser.add_argument('--out_file', dest='out_file', help='JSON output file') args = parser.parse_args() download(args.database_id, args.database_name, args.url, args.out_file)