match_filter fails when using in embedded Python code

[GijsTimmers]

I want to only download YouTube videos with a valid upload date. If the upload date is not available, I don't want to download the video. Therefore, match_filter could be interesting.

In [8]: import youtube_dl

In [9]: ydl = youtube_dl.YoutubeDL({"match_filter": "upload_date"})

In [10]: ydl.download(["https://www.youtube.com/watch?v=cU2dZz18P0c"])
[youtube] cU2dZz18P0c: Downloading webpage
[youtube] cU2dZz18P0c: Extracting video information
[youtube] cU2dZz18P0c: Downloading DASH manifest
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-10-fc60485bb9c2> in <module>()
----> 1 ydl.download(["https://www.youtube.com/watch?v=cU2dZz18P0c"])

/usr/local/lib/python2.7/dist-packages/youtube_dl/YoutubeDL.pyc in download(self, url_list)
   1385             try:
   1386                 # It also downloads the videos
-> 1387                 res = self.extract_info(url)
   1388             except UnavailableVideoError:
   1389                 self.report_error('unable to download video')

/usr/local/lib/python2.7/dist-packages/youtube_dl/YoutubeDL.pyc in extract_info(self, url, download, ie_key, extra_info, process)
    642                 self.add_default_extra_info(ie_result, ie, url)
    643                 if process:
--> 644                     return self.process_ie_result(ie_result, download, extra_info)
    645                 else:
    646                     return ie_result

/usr/local/lib/python2.7/dist-packages/youtube_dl/YoutubeDL.pyc in process_ie_result(self, ie_result, download, extra_info)
    688         if result_type == 'video':
    689             self.add_extra_info(ie_result, extra_info)
--> 690             return self.process_video_result(ie_result, download=download)
    691         elif result_type == 'url':
    692             # We have to add extra_info to the results because it may be

/usr/local/lib/python2.7/dist-packages/youtube_dl/YoutubeDL.pyc in process_video_result(self, info_dict, download)
   1143                 new_info = dict(info_dict)
   1144                 new_info.update(format)
-> 1145                 self.process_info(new_info)
   1146         # We update the info dict with the best quality format (backwards compatibility)
   1147         info_dict.update(formats_to_download[-1])

/usr/local/lib/python2.7/dist-packages/youtube_dl/YoutubeDL.pyc in process_info(self, info_dict)
   1168             info_dict['format'] = info_dict['ext']
   1169 
-> 1170         reason = self._match_entry(info_dict, incomplete=False)
   1171         if reason is not None:
   1172             self.to_screen('[download] ' + reason)

/usr/local/lib/python2.7/dist-packages/youtube_dl/YoutubeDL.pyc in _match_entry(self, info_dict, incomplete)
    597             match_filter = self.params.get('match_filter')
    598             if match_filter is not None:
--> 599                 ret = match_filter(info_dict)
    600                 if ret is not None:
    601                     return ret

TypeError: 'str' object is not callable

[phihag]

As documented, in the API match_filter is a callable (i.e. function), not a string. You have two options:

1. Configure {"match_filter": youtube_dl.utils.match_filter_func("upload_date")}
2. Cut out the middle-man and implement your own filtering function

[GijsTimmers]

OK, thank you for clarifying this.