Switch extract to return bytes instead of decoding to avoid errors when downloading subtitles marked with the wrong encoding

Author: CosmicHorrorDev

docs/Utility-Functions.md

@@ -13,12 +13,12 @@ from subwinder.utils import extract, special_hash
 * [`extract()`](#extractbytes-encoding)
 * [`special_hash()`](#special_hashfilepath)
 
-### `extract(bytes, encoding)`
+### `extract(bytes)`
 
-Small helper function that base64 decodes and gzip decompresses `bytes` into from an `encoding` encoded string to a python `str`. This likely won't be useful to many people, but this is the format used to transfer subtitles and previews from the opensubtitles API.
+Small helper function that base64 decodes and gzip decompresses `bytes`. This likely won't be useful to many people, but this is the format used to transfer subtitles and previews from the opensubtitles API.
 
 ```python
-assert "Hi!" == extract(b"H4sIAIjurl4C//PIVAQA2sWeeQMAAAA=", "UTF-8")
+assert b"Hi!" == extract(b"H4sIAIjurl4C//PIVAQA2sWeeQMAAAA=")
 ```
 
 ### `special_hash(filepath)`

subwinder/core.py

@@ -275,30 +275,25 @@ def download_subtitles(
         return download_paths
 
     def _download_subtitles(self, sub_containers, filepaths):
-        encodings = []
         sub_file_ids = []
-        # Unpack stored info
-        for sub_container in sub_containers:
-            encodings.append(sub_container.encoding)
-            sub_file_ids.append(sub_container.file_id)
+        # Get all the file ids
+        sub_file_ids = [sub_container.file_id for sub_container in sub_containers]
 
         data = self._request(Endpoints.DOWNLOAD_SUBTITLES, sub_file_ids)["data"]
 
-        for encoding, result, fpath in zip(encodings, data, filepaths):
-            # Currently pray that python supports all the encodings and is called the
-            # same as what opensubtitles returns
-            subtitles = utils.extract(result["data"], encoding)
+        for result, fpath in zip(data, filepaths):
+            subtitles = utils.extract(result["data"])
 
             # Create the directories if needed, then save the file
             dirpath = fpath.parent
             dirpath.mkdir(exist_ok=True)
 
             # Write atomically if possible, otherwise fall back to regular writing
             if ATOMIC_DOWNLOADS_SUPPORT:
-                with atomic_write(fpath) as f:
+                with atomic_write(fpath, mode="wb") as f:
                     f.write(subtitles)
             else:
-                with fpath.open("w") as f:
+                with fpath.open("wb") as f:
                     f.write(subtitles)
 
     def get_comments(self, sub_containers):
@@ -563,6 +558,7 @@ def _preview_subtitles(self, ids):
             encoding = preview["encoding"]
             contents = preview["contents"]
 
-            previews.append(utils.extract(contents, encoding))
+            # Extract and decode the previews
+            previews.append(utils.extract(contents).decode(encoding))
 
         return previews

subwinder/utils.py

@@ -5,14 +5,12 @@
 from subwinder.exceptions import SubHashError
 
 
-def extract(bytes, encoding):
+def extract(bytes):
     """
     Extract `bytes` from being gzip'd and base64 encoded.
     """
     compressed = base64.b64decode(bytes)
-    # TODO: only previewing cares about getting this as a string. downloading can take
-    #       it directly as bytes
-    return gzip.decompress(compressed).decode(encoding)
+    return gzip.decompress(compressed)
 
 
 # As per API spec with some tweaks to make it a bit nicer

tests/test_utils.py

@@ -8,11 +8,11 @@ def test_extract():
         "wa+2QmHRYOxiZfzuNRrVZv8dQcVk3xP08dSMFps5/4WhRKSPvwBzf2OXZqAAAA"
     )
     IDEAL = (
-        "Hello there, I'm that good ole compressed and encoded subtitle information"
-        " that you so dearly want to save"
+        b"Hello there, I'm that good ole compressed and encoded subtitle information"
+        b" that you so dearly want to save"
     )
 
-    assert extract(COMPRESSED, "UTF-8") == IDEAL
+    assert extract(COMPRESSED) == IDEAL
 
 
 def test_special_hash():