Unquote the filename from URL in fetch_http

Signed-off-by: tdruez <[email protected]>

scanpipe/models.py

@@ -1514,10 +1514,14 @@ def fetch(self):
             raise Exception("No `download_url` value to be fetched.")
 
         downloaded = fetch_url(url=self.download_url)
-        self.filename = downloaded.filename
-        self.save()
-
         destination = self.project.move_input_from(downloaded.path)
+
+        # Force a commit to the database to ensure the file on disk is not rendered
+        # as "manually uploaded" in the UI.
+        with transaction.atomic():
+            self.filename = downloaded.filename
+            self.save()
+
         return destination
 
 

scanpipe/pipes/fetch.py

@@ -29,6 +29,7 @@
 import tempfile
 from collections import namedtuple
 from pathlib import Path
+from urllib.parse import unquote
 from urllib.parse import urlparse
 
 import requests
@@ -85,7 +86,7 @@ def fetch_http(uri, to=None):
     if not filename:
         # Using `response.url` in place of provided `Scan.uri` since the former
         # will be more accurate in case of HTTP redirect.
-        filename = Path(urlparse(response.url).path).name
+        filename = unquote(Path(urlparse(response.url).path).name)
 
     download_directory = to or tempfile.mkdtemp()
     output_file = Path(download_directory, filename)

scanpipe/tests/pipes/test_fetch.py

@@ -41,12 +41,12 @@ def test_scanpipe_pipes_fetch_http(self, mock_get):
         downloaded_file = fetch.fetch_http(url)
         self.assertTrue(Path(downloaded_file.directory, "filename.zip").exists())
 
-        redirect_url = "https://example.com/redirect.zip"
+        url_with_spaces = "https://example.com/space%20in%20name.zip"
         mock_get.return_value = mock.Mock(
-            content=b"\x00", headers={}, status_code=200, url=redirect_url
+            content=b"\x00", headers={}, status_code=200, url=url_with_spaces
         )
         downloaded_file = fetch.fetch_http(url)
-        self.assertTrue(Path(downloaded_file.directory, "redirect.zip").exists())
+        self.assertTrue(Path(downloaded_file.directory, "space in name.zip").exists())
 
         headers = {
             "content-disposition": 'attachment; filename="another_name.zip"',