Support extracting metadata from bytes and stream objects (#1547)

This should enable more accurate uploads of in-memory files.

telethon/utils.py

@@ -601,16 +601,52 @@ def get_message_id(message):
 
 
 def _get_metadata(file):
-    # `hachoir` only deals with paths to in-disk files, while
+    if not hachoir:
-    # `_get_extension` supports a few other things. The parser
+        return
-    # may also fail in any case and we don't want to crash if
+
+    stream = None
+    close_stream = True
+    seekable = True
+
+    # The parser may fail and we don't want to crash if
     # the extraction process fails.
-    if hachoir and isinstance(file, str) and os.path.isfile(file):
+    try:
-        try:
+        # Note: aiofiles are intentionally left out for simplicity
-            with hachoir.parser.createParser(file) as parser:
+        if isinstance(file, str):
-                return hachoir.metadata.extractMetadata(parser)
+            stream = open(file, 'rb')
-        except Exception as e:
+        elif isinstance(file, bytes):
-            _log.warning('Failed to analyze %s: %s %s', file, e.__class__, e)
+            stream = io.BytesIO(file)
+        else:
+            stream = file
+            close_stream = False
+            if getattr(file, 'seekable', None):
+                seekable = file.seekable()
+            else:
+                seekable = False
+
+        if not seekable:
+            return None
+
+        pos = stream.tell()
+        filename = getattr(file, 'name', '')
+
+        parser = hachoir.parser.guess.guessParser(hachoir.stream.InputIOStream(
+            stream,
+            source='file:' + filename,
+            tags=[],
+            filename=filename
+        ))
+
+        return hachoir.metadata.extractMetadata(parser)
+
+    except Exception as e:
+        _log.warning('Failed to analyze %s: %s %s', file, e.__class__, e)
+
+    finally:
+        if stream and close_stream:
+            stream.close()
+        elif stream and seekable:
+            stream.seek(pos)
 
 
 def get_attributes(file, *, attributes=None, mime_type=None,
@@ -803,15 +839,31 @@ def is_gif(file):
 
 
 def is_audio(file):
-    """Returns `True` if the file extension looks like an audio file."""
+    """Returns `True` if the file has an audio mime type."""
-    file = 'a' + _get_extension(file)
+    ext = _get_extension(file)
-    return (mimetypes.guess_type(file)[0] or '').startswith('audio/')
+    if not ext:
+        metadata = _get_metadata(file)
+        if metadata and metadata.has('mime_type'):
+            return metadata.get('mime_type').startswith('audio/')
+        else:
+            return False
+    else:
+        file = 'a' + ext
+        return (mimetypes.guess_type(file)[0] or '').startswith('audio/')
 
 
 def is_video(file):
-    """Returns `True` if the file extension looks like a video file."""
+    """Returns `True` if the file has a video mime type."""
-    file = 'a' + _get_extension(file)
+    ext = _get_extension(file)
-    return (mimetypes.guess_type(file)[0] or '').startswith('video/')
+    if not ext:
+        metadata = _get_metadata(file)
+        if metadata and metadata.has('mime_type'):
+            return metadata.get('mime_type').startswith('video/')
+        else:
+            return False
+    else:
+        file = 'a' + ext
+        return (mimetypes.guess_type(file)[0] or '').startswith('video/')
 
 
 def is_list_like(obj):