Added Non-Ascii filename support

matweb/frontend.py

@@ -68,7 +68,7 @@ def upload_file():
             parser, mime = utils.get_file_parser(filepath)
         except ValueError:
             flash('The filetype is not supported')
-            current_app.logger.error('Unsupported filetype',)
+            current_app.logger.error('Unsupported filetype')
             return redirect(url_for('routes.upload_file'))
 
         try:

matweb/utils.py

@@ -2,6 +2,8 @@ import hmac
 import os
 import hashlib
 import mimetypes as mtype
+import pathlib
+import uuid
 from typing import Tuple
 
 from flask_restful import abort, current_app
@@ -68,9 +70,16 @@ def get_supported_extensions():
 
 
 def save_file(file, upload_folder):
-    filename = secure_filename(file.filename)
+    path = pathlib.Path(file.filename)
+    extension = path.suffix
+    stem = path.stem
+
+    filename = secure_filename(stem)
     if not filename:
-        raise ValueError('Invalid Filename')
+        filename = str(uuid.uuid4())
+
+    if extension:
+        filename = str(pathlib.Path(filename).with_suffix(extension))
     filepath = os.path.join(upload_folder, filename)
     file.save(os.path.join(filepath))
     return filename, filepath

test/test.py

@@ -187,17 +187,28 @@ class Mat2WebTestCase(TestCase):
         self.assertIn(b'.mp2', rv.data)
         self.assertEqual(rv.status_code, 200)
 
-    def test_get_upload_naughty_input(self):
+    def test_get_upload_no_ascii_no_ext_input(self):
         rv = self.client.post(
             '/',
             data=dict(
-                file=(io.BytesIO(b"a"), '﷽'),
+                file=(io.BytesIO(b"a"), '﷽.txt'),
             ),
             follow_redirects=True
         )
         self.assertEqual(rv.status_code, 200)
-        self.assertIn(b'Invalid Filename', rv.data)
+        self.assertIn(b'.cleaned.txt', rv.data)
 
+    def test_get_upload_no_ascii_stem_input(self):
+        pdfBytes = b"%PDF-1.\n 1 0 obj<</Pages 2 0 R>>endobj\n2 0 obj<</Kids[3 0 R]/Count 1>>endobj\n3 0 obj<</Parent 2 0 R>>endobj\ntrailer <</Root 1 0 R>>"
+        rv = self.client.post(
+            '/',
+            data=dict(
+                file=(io.BytesIO(pdfBytes), '한국어.pdf'),
+            ),
+            follow_redirects=True
+        )
+        self.assertEqual(rv.status_code, 200)
+        self.assertIn(b'.cleaned.pdf', rv.data)
 
 if __name__ == '__main__':
     unittest.main()

test/test_api.py

@@ -451,8 +451,8 @@ class Mat2APITestCase(unittest.TestCase):
                            headers={'content-type': 'application/json'}
                            )
         error_message = request.get_json()['message']
-        self.assertEqual(400, request.status_code)
-        self.assertEqual("Invalid Filename", error_message)
+        self.assertEqual(415, request.status_code)
+        self.assertEqual("The filetype is not supported", error_message)
 
         request = self.app.post('/api/upload',
                                 data='{"file_name": "﷽", '
@@ -534,8 +534,8 @@ class Mat2APITestCase(unittest.TestCase):
             ),
             follow_redirects=False
         )
-        self.assertEqual(r.get_json()['message'], 'Invalid Filename')
-        self.assertEqual(r.status_code, 400)
+        self.assertEqual(r.get_json()['message'], 'The filetype is not supported')
+        self.assertEqual(r.status_code, 415)
 
         r = self.app.post(
             '/api/remove_metadata',