Download product with multiple workers

This patch introduces workers for downloading files using the executable.
For programs using the library the parallelization should be done in the calling code itself.

There could be a lot of md5sums to process when working on long download periods.
Therefore `multiprocessing` is used to be able to calculate those hashes on multiple cores.

sentinel5dl/__main__.py

@@ -12,6 +12,7 @@
 import dateutil.parser
 import certifi
 import logging
+import multiprocessing
 import textwrap
 import sentinel5dl
 from sentinel5dl import search, download
@@ -148,6 +149,13 @@ def main():
             providing a cabundle.'''
     )
 
+    parser.add_argument(
+        '--worker',
+        type=int,
+        default=1,
+        help='Number of parallel downloads',
+    )
+
     parser.add_argument(
         'download_dir',
         metavar='download-dir',
@@ -170,8 +178,11 @@ def main():
         processing_mode=args.mode
     )
 
-    # Download found products to the local folder
-    download(result.get('products'), args.download_dir)
+    # Download found products to the download directory with number of workers
+    with multiprocessing.Pool(args.worker) as p:
+        p.starmap(download, map(
+            lambda product: ((product,), args.download_dir),
+            result.get('products')))
 
 
 if __name__ == '__main__':