Enhancement: Enable resume on partially downloaded video stream

pytube/__main__.py

@@ -250,7 +250,7 @@ def vid_info(self):
     def bypass_age_gate(self):
         """Attempt to update the vid_info by bypassing the age gate."""
         innertube = InnerTube(
-            client='ANDROID_EMBED',
+            client='ANDROID',
             use_oauth=self.use_oauth,
             allow_cache=self.allow_oauth_cache
         )

pytube/request.py

@@ -4,6 +4,7 @@
 import logging
 import re
 import socket
+import sys
 from functools import lru_cache
 from urllib import parse
 from urllib.error import URLError
@@ -133,17 +134,16 @@ def seq_stream(
 def stream(
     url,
     timeout=socket._GLOBAL_DEFAULT_TIMEOUT,
-    max_retries=0
+    max_retries=0,
+    start_pos=0
 ):
     """Read the response in chunks.
     :param str url: The URL to perform the GET request for.
     :rtype: Iterable[bytes]
     """
-    file_size: int = default_range_size  # fake filesize to start
-    downloaded = 0
-    while downloaded < file_size:
-        stop_pos = min(downloaded + default_range_size, file_size) - 1
-        range_header = f"bytes={downloaded}-{stop_pos}"
+    file_size: int = sys.maxsize  # fake filesize to start
+    while start_pos < file_size:
+        stop_pos = min(start_pos + default_range_size, file_size) - 1
         tries = 0
 
         # Attempt to make the request multiple times as necessary.
@@ -155,7 +155,7 @@ def stream(
             # Try to execute the request, ignoring socket timeouts
             try:
                 response = _execute_request(
-                    url + f"&range={downloaded}-{stop_pos}",
+                    url + f"&range={start_pos}-{stop_pos}",
                     method="GET",
                     timeout=timeout
                 )
@@ -174,7 +174,7 @@ def stream(
                 break
             tries += 1
 
-        if file_size == default_range_size:
+        if file_size == sys.maxsize:
             try:
                 resp = _execute_request(
                     url + f"&range={0}-{99999999999}",
@@ -189,7 +189,7 @@ def stream(
             chunk = response.read()
             if not chunk:
                 break
-            downloaded += len(chunk)
+            start_pos += len(chunk)
             yield chunk
     return  # pylint: disable=R1711
 

pytube/streams.py

@@ -305,19 +305,29 @@ def download(
             logger.debug(f'file {file_path} already exists, skipping')
             self.on_complete(file_path)
             return file_path
-
-        bytes_remaining = self.filesize
+
         logger.debug(f'downloading ({self.filesize} total bytes) file to {file_path}')
-
-        with open(file_path, "wb") as fh:
+        downloaded = 0
+        bytes_remaining = self.filesize
+        if os.path.isfile(file_path):
+            partial_download_size = os.path.getsize(file_path)
+            if partial_download_size % request.default_range_size == 0:
+                downloaded = partial_download_size
+                bytes_remaining -= downloaded
+                logger.debug(f'appending to valid partial download file ({partial_download_size} bytes)')
+
+        file_mode = "ab" if downloaded > 0 else "wb"
+        with open(file_path, file_mode) as fh:
             try:
                 for chunk in request.stream(
                     self.url,
                     timeout=timeout,
-                    max_retries=max_retries
+                    max_retries=max_retries,
+                    start_pos=downloaded
                 ):
                     # reduce the (bytes) remainder by the length of the chunk.
                     bytes_remaining -= len(chunk)
+                    logger.debug(f"download chunks_left={bytes_remaining//len(chunk)}")
                     # send to the on_progress callback.
                     self.on_progress(chunk, fh, bytes_remaining)
             except HTTPError as e: