moonbase59 / azuracast_xmltv Goto Github PK

In "gap_filler":

• {playlists}* -- comma-separated list of playlist names that make up the general rotation

{playlists} returns just one playlist. "Rock und Deutsch-Rock" which is the default one. Should it not return all (public) playlists?

I have two playlists called Hard Rock:

The first one (weight 1) is in my General Rotation and not in any schedule, while the second (weight 3) is scheduled to play on Fridays at 18:00–20:00 hours.

Having more than one playlist with the same name is perfectly legal in AzuraCast, and can help making nice names for the EPG and RSS feed. The only culprit is that AzuraCast still doesn’t show a playlist being on schedule or not.

You would, for instance, use this if you had genre-based playlists and a general rotation made up of a weighted mixture of several playlists (I have "Classic Rock", "Folk Rock" and "Hard Rock"). If you play 24/7, and auto-generate gap filler programmes in azuracast_xmltv, you would want it to generate nice names to be used with the {playlists} variable. In my case, it would generate Classic Rock, Folk Rock & Hard Rock as a variable that I can use in the subtitle and description of a programme.

Thus the need for occasionally using duplicate playlist names. And the need for a "Scheduled" tag in the UI, please…

Anyway, if using an API key with azuracast_xmltv, I can get much more information from the API, like whether listener requests are allowed or not, and generate much nicer programme information.

The API doesn’t give me the ID of the playlist involved in a scheduled item, only its name. This "name-only lookup" occasionally leads to using the wrong playlist’s data, and, for example, not showing the "requests enabled" text.

In my case, it would occasionally find the "Hard Rock" playlist that’s in General Rotation (and doesn’t allow requests) instead of the "Hard Rock" playlist that’s on the Friday schedule (and allows requests).

TODO: Find a better way to identify and use the correct playlist for a schedule item, if more than one playlist has the same name and an API key is used (allowing me to access the playlist info, which isn’t possible without an API key).

When using a port number in your AzuraCast server URL, it will be taken into the generated channel id and XML output file name. This is fine on Linux, but might be a problem for Mac and Windows users.

Since it also breaks the RFC2838-compliance for tv channel ids, I’ll fix it.

The replacer logic (using {variable}) produces an error when a field like Website-URL in the station is empty (JSON null) and you try to use this variable (ex: {station_website}):

TypeError: replace() argument 2 must be str, not None

It should instead return an empty string for the variable in question.

Will fix this ASAP.

I use the default rss_feed_description template with azuracast_xmltv v0.12.2:

rss_feed_description = """<h2>{title}</h2>
<h3>{subtitle}</h3>
<p><strong>{airdate}&emsp;{airtime}</strong> ({airtime_length})&emsp;{station_name} ({category})</p>
<p>{desc}</p>"""

Now for schedule items with syndicated content (remote stream), the subtitle is empty, effectively generating a <h3></h3> in the HTML. This apparently makes the Liferea RSS feed reader (v1.13.7) choke:

Interestingly enough, when I modify the template to add a blank after {subtitle}, as in

<h3>{subtitle} </h3>

everything gets displayed as it should:

Other readers aren’t affected by this change, and don’t display the <h3> line even when it contains a blank. (A blank, as opposed to a non-breaking space  , is considered whitespace and thus doesn’t get extra rendering.)

TODO: Change default RSS template to include an extra blank, just to be on the safe side.