[mythtv-users] DVB re-scan and XMLTV EPG?

Mike Perkins mikep at randomtraveller.org.uk
Thu Jan 29 10:12:19 UTC 2015


On 29/01/15 09:39, Alistair Grant wrote:
> Hi Hika,
>
> On Thu, Jan 29, 2015 at 4:18 AM, Hika van den Hoven <hikavdh at gmail.com> wrote:
>> I've been thinking a bit about this thing. As you might have seen I am
>> currently developing on the Dutch XMLTV grabber.
>>
>> There are several identifiers for a channel. Of which several in
>> essence are free for the user to choose/define. What such a script
>> should do is save the linking between the three identifiers BEFORE you
>> delete the old lineup and then afterwards restore on the basis of the
>> saved info. So you need a two step script. (three if you include the
>> scan)
>>
>> You have
>> - chanid    Internal MythTV identifier, which you on first setup are
>>              free to choose and which links to all kind of other
>>              tables.
>
> chanid is the primary key for the channel table.  The user shouldn't
> have to know about it, and I have a hard time believing that it is OK
> to  change it.
>
chanid is also the first part of the generated filename for the recording, the 
rest being the date and time. It will be the only way to link older recordings 
back to the channel they came from, so changing it should be a last resort.

Personally, I would have steered clear of that format, since it is putting 
metadata in the filename. A simple unique integer,with perhaps an alpha prefix, 
could link the file back to the appropriate database records and avoid the 
problem I mention above.
>
>
>> - sourceid  The MythTV internal ID of the source of your EPG-data
>>              (confusingly called video-source)
>> - xmltvid   EPG-id defined by the XMLTV grabber/EPG-source.
>> - channum   Is the identifier from your provider/cable company, but I
>>              think only relevant during a scan and afterwards free to
>>              choose/change. Unless you use Schedules Direct, where it
>>              doubles as EPG-id.
>> - callsign  Is I think the naming of the channel by your provider/cable
>>              company, but again I think in essence free to choose/change
>>              afterwards.
Be aware that Callsign in the US is a short jumble of letters. In UK Freeview, 
it can be a long descriptive string, such as 'BBC ONE South'. This usually gets 
truncated, meaning matching can be a problem, even by eye.

>> - name      Is your name of choice
>
>
>> - freqid    On analog the frequency in kHz or the channel-number
>> - mplexid   and
>> - serviceid The identifiers of the actual digital channel
>> - iptvid    The identifier of an iptv channel
>
> The user shouldn't have to know about any of these, hopefully :-)
> (I'm not sure about setting up iptvid).
>
In the UK, freqid is the number of the digital channel as supplied by Freeview. 
This will be used by the tuner to locate and tune the channel.

It is possible that the other values are used instead, but if I type in '11' 
then the channel having freqid '11' is what is tuned.

>> What the script needs to do is to relink chanid, xmltvid, name and
>> maybe sourceid back to the found channels as identified by channum. A
>> problem comes if your provider changes channums, so you always will
>> need a personal check afterwards.
>
> I assumed that the script would:
>
> * Use the CallSign to link the tv_grab_xxx (EPG data) output to the
> MythTV channel data.
> ** This is where the fuzzy logic would be used to match callsigns.
> * Update the xmltvid's in the MythTV channel table.
> * Allow the user to:
> ** Modify the CallSign in MythTV (to match the EPG data)
> ** Provide mapping between EPG callsign and MythTV callsign (if they
> don't want to modify the CallSign for some reason).
> ** Modify ChanNum in MythTV
> ** Modify ChannelName in MythTV
>
> The script configuration will be in a file that will allow it to be
> re-run as a single step in the event that a re-scan is necessary.
>
> I'm not aware of the design decisions behind the structure of the
> Channel table, so some of the above may be incorrect.
>


-- 

Mike Perkins



More information about the mythtv-users mailing list