plugin for npo.nl stoppped working.

Checklist

• This is a bug report.
• This is a plugin request.
• This is a feature request.
• I used the search function to find already opened/closed issues or pull requests.

Description

Trying to play any url at npo.nl/live results in an error.

Expected / Actual behavior

expecting a stream, error is:

[cli][info] Found matching plugin npo for URL npo.nl/live/npo-1
error: Unable to open URL: http://ida.omroep.nl/npoplayer/i.js?s=http%3A//npo.nl/live/npo-1 (410 Client Error: Gone for url: http://ida.omroep.nl/npoplayer/i.js?s=http%3A//npo.nl/live/npo-1)

Reproduction steps / Stream URLs to test

streamlink -l debug npo.nl/live/npo-1

Environment details (operating system, python version, etc.)

Multiple systems, all using streamlink 0.3.2, Python 2.7.12

[beardypig]

Cool. tvcatchup.py is a pretty basic plugin.

To make a Plugin you have to subclass Plugin and set the module level variable __plugin__ to that class (__plugin__ = TVCatchup). Then you have to implement the two required methods, can_handle_url (a classmethod) and _get_streams.

can_handle_url is super simple, you just return True/False if this plugin can handle the supplied URL. Normally this is done wit a regex, (r"http://(?:www\.)?tvcatchup.com/watch/\w+").

_get_streams is also pretty simple, you have either return a list, iterable, or dict containing the streams. The dict must have the quality name (eg. 540p) as the key and an instance of Stream as the value, likewise with the list/iterable it must be a tuple of (quality, stream). There are a number of different Stream classes in the stream module. The most common one now is probably HLSStream (as in tvcatchup). There is a special global http which is a requests.Session instance (with some extra methods for json/xml), which has been set up with the users preferences, you should use this to do all http requests. You can see in the tvcatchup plugin, the page (self.url) is requested and an m3u8 URL is searched for, if it is found then it is parsed with HLSStream.parse_variant_playlist if the URL contains _adp. HLSStream.parse_variant_playlist returns a properly formatted dict with all the streams. In the other case (_adp not in the URL) it assumes it’s 576p and returns a dict with one HLSStream instance.

If you fire up a python shell and do

import streamlink
streamlink.streams("http://tvcatchup.com/watch/bbcone")

you will see what the plugin returns 😃 In this case the quality keys are in bitrate (550k, etc). Streamlink sorts the qualities returned by the Plugin (see stream_weight to see how the weights are compared).

To write your own plugin you need to work out how the site works, and how it gets it stream urls (m3u8/f4m/etc.). Then it’s just a case of getting your Plugin to replicate it 😃

[beardypig]

I have created a PR with an updated version of the plugin, would be good if you could test it out @sophof 😃