lang options for captions downloader

.gitignore

@@ -163,7 +163,7 @@ cython_debug/
 .DS_Store
 
 # our stuff
-captions/
+captions*/
 db/
 *.json
 *.conf

src/download_captions.py

@@ -1,5 +1,6 @@
 #!/usr/bin/env python3
 import os
+import sys
 import json
 import html
 from downloader import Downloader
@@ -11,6 +12,9 @@ def main():
   if not os.path.exists('captions'):
     os.mkdir('captions')
 
+  # lang
+  lang = None if len(sys.argv) == 1 else sys.argv[1]
+
   videos = json.load(open('videos.json'))
   for video in videos:
 
@@ -26,7 +30,7 @@ def main():
       original = open(f'captions/{id}.original.vtt', 'r').read()
     else:
       print(f'[youtube] downloading captions for {id}: {title}...')
-      original = downloader.download_captions(id)
+      original = downloader.download_captions(id, lang)
       with open(f'captions/{id}.original.vtt', 'w') as f:
         f.write(original)
 

src/downloader.py

@@ -59,7 +59,7 @@ def _cleanup_captions(self, original_captions):
     # remove header lines
     contents = re.sub(r'WEBVTT\n', '', contents)
     contents = re.sub(r'Kind: captions\n', '', contents)
-    contents = re.sub(r'Language: en\n', '', contents)
+    contents = re.sub(r'Language: .*?\n', '', contents)
 
     # remove timestamp lines
     contents = re.sub(r'\d\d:\d\d:\d\d\.\d\d\d --> .*\n', '', contents)