--- a/data_manager/data_manager.py	Sun Nov 22 12:47:50 2020 +0000
+++ b/data_manager/data_manager.py	Mon Apr 24 12:26:46 2023 +0000
@@ -1,71 +1,47 @@
+#!/usr/bin/env python
+#
+# Data manager for reference data for the 'BUSCO' Galaxy tools
 import argparse
 import datetime
 import json
 import os
 import shutil
-import tarfile
-import zipfile
-try:
-    # For Python 3.0 and later
-    from urllib.request import Request, urlopen
-except ImportError:
-    # Fall back to Python 2 imports
-    from urllib2 import Request, urlopen
-
-
-def url_download(url, workdir):
-    file_path = os.path.join(workdir, 'download.dat')
-    if not os.path.exists(workdir):
-        os.makedirs(workdir)
-    src = None
-    dst = None
-    try:
-        req = Request(url)
-        src = urlopen(req)
-        with open(file_path, 'wb') as dst:
-            while True:
-                chunk = src.read(2**10)
-                if chunk:
-                    dst.write(chunk)
-                else:
-                    break
-    finally:
-        if src:
-            src.close()
-    if tarfile.is_tarfile(file_path):
-        fh = tarfile.open(file_path, 'r:*')
-    elif zipfile.is_zipfile(file_path):
-        fh = zipfile.ZipFile(file_path, 'r')
-    else:
-        return
-    fh.extractall(workdir)
-    os.remove(file_path)
+import subprocess
+from pathlib import Path
 
 
 def main(args):
-    workdir = os.path.join(os.getcwd(), 'busco')
-    url_download(args.url, workdir)
+    workdir = os.path.join(os.getcwd(), "busco_downloads")
+    cmd = "busco --download %s" % args.database
+    subprocess.check_call(cmd, shell=True)
+    with open(args.json) as fh:
+        params = json.load(fh)
+    target_directory = params["output_data"][0]["extra_files_path"]
     data_manager_entry = {}
-    data_manager_entry['value'] = args.name.lower()
-    data_manager_entry['name'] = args.name
-    data_manager_entry['path'] = '.'
-    data_manager_json = dict(data_tables=dict(busco=data_manager_entry))
-    with open(args.output) as fh:
-        params = json.load(fh)
-    target_directory = params['output_data'][0]['extra_files_path']
+    data_manager_entry["value"] = args.name.lower()
+    data_manager_entry["name"] = args.name
+    data_manager_entry["version"] = args.version
+    data_manager_entry["path"] = str(Path(target_directory))
+    data_manager_json = dict(data_tables=dict(busco_database=data_manager_entry))
+
     os.mkdir(target_directory)
-    output_path = os.path.abspath(os.path.join(os.getcwd(), 'busco'))
+    output_path = os.path.abspath(os.path.join(os.getcwd(), "busco_downloads"))
     for filename in os.listdir(workdir):
         shutil.move(os.path.join(output_path, filename), target_directory)
-    with open(args.output, 'w') as fh:
+    with open(args.json, "w") as fh:
         json.dump(data_manager_json, fh, sort_keys=True)
 
 
-if __name__ == '__main__':
-    parser = argparse.ArgumentParser(description='Create data manager json.')
-    parser.add_argument('--out', dest='output', action='store', help='JSON filename')
-    parser.add_argument('--name', dest='name', action='store', default=str(datetime.date.today()), help='Data table entry unique ID')
-    parser.add_argument('--url', dest='url', action='store', help='Download URL')
+if __name__ == "__main__":
+
+    # Read command line
+    parser = argparse.ArgumentParser(description="Download BUSCO database")
+    parser.add_argument("--database", help="Database name")
+    parser.add_argument(
+        "--name", default=str(datetime.date.today()), help="Data table entry unique ID"
+    )
+    parser.add_argument("--version", help="BUSCO version")
+    parser.add_argument("--json", help="Path to JSON file")
     args = parser.parse_args()
 
     main(args)