Mythlink Woes

[Steve Berrett]

Hi Marc
 
I'm afraid I've falllen at the first hurdle with your mythlink script and need your help please.
 
The script connects to the database and finds the data it is looking for but only manages to complete one full cycle (ie the first recording) and then fails. I have confirmed this by deleting the first recording and then the failure occurs on the following recording.
 
I have attached a debug dump in the hope you can work out what is going on.
 
I can confirm that the first recording (Gok) has been processed properly with all the associated nfo and txt files created in an appropriately named folder under the XBMC sub directory and that the symlink from the _ts file back to the .mpg file works as expected. However the folder for the following recording (Britians Best Dish) has been created but not populated with anything.
 
I am running the script with defaults and no options on a Lucid Linux server with python 2.6.
 
Myth recordings are on a dedicated 1TB drive mounted at ./MEDIA/Recordings/ with the intended XBMC symlinks at /MEDIA/Recordings/XBMC/
 
I intend to use the output with both Popcorn hour A110 and XBMC frontends, however I noticed my latest Dharma copy XBMC has the buttons for TV Shows greyed out in the settings. Am I missing something or will these menu options magically appear when I install the advancedsettings.xml file ?
 
Thanks in advance for your help.
 
Rgds
Steve

[me]

Hi Steve, I see the problem but I'll need more info to sort it out.
The target path for the symlink is coming up as"/MEDIA/Recordings/XBMC/
Britain_s Best Dish/Britain_s Best Dish.S00Em/63.2010-08-19
17_00_00-11000-2010-08-19 17_00_00.ts". The problematic portion of
this path is ".S00Em/63." That part of the string is supposed to be of
the form "S00Ennnn", where "nnnn"is the episode number. I obtain the
episode number from the "programid" field of the "recorded" table in
the mythtv database. My experience to date has been that the 4
characters beginning at character number 10 are a 4-digit episode
number. It also looks like the first two characters of your programid
contain "ww", which is a type of haven't dealt with before.

If you email me a dump of the follow columns from your recorded table
I should be able to figure out what is going on & provide you a
modified version of the scripts that deal with this type of recording:
chanid, starttime, title, subtitle, programid

It's pure luck that your first recording worked -- it came through
with an episode number of "4.c.".It's the slash that was fatal on the
second recording, but I suspect you will have other down-stream
problems even with the first one once you try to use XBMC.

By the way, I have some scripts for the popcorn hour that will
generate listings of recordings that include episode title, synopsis,
recording date, etc for the popcorn.. They run off the same NFO files
I generate for XBMC. Once we get this stuff working we can move on to
those scripts if you are interested.

Marc

>  debug_dump.txt
> 3KViewDownload

[Steve Berrett]

Hi Marc

 

Extract from Recorded table attached as requested.

 

The programid field is being populated by EIT. (dvb-s Freesat) Looking at the data I guess the last 3 digits are episode numbers? 

 

The PCH has been a huge disappointment in terms of interface and functionality; anything to give it a new lease of life would be appreciated.

 

Rgds

 

Steve 

[me]

Hi Steve,

Looks like I made some incorrect assumptions about the format of the
programid field -- that's the risk of "reverse engineering" the
behavior from a single data point (my system). I'll figure something
out -- need to sort out what to do for episodes like "murder she
wrote" that has a program id of "fp.bbc.co.uk/4j5dsj". I think it
would be helpful if you sent me a full extract of the "recorded" table
-- I probably need to find another field to key off of. Any guesses as
to what "fp" stands for?

In terms of the popcorn hour interface -- if you want to see what I've
built, take a look at http://mythtv.mlaronson.com/networked-media-tank/screenshots

It's not fancy, but it does the job.

Marc

[Steve Berrett]

 

Hi Marc,

 

Full extract of "recorded" table attached. Not much good news here; the "seriesid" field contains a similar mix of data. 

 

I was wondering, rather than reinventing the wheel with complicated scripting, does the EPG feed you use cover Freesat?  It is on my to-do list to configure a more comprehensive guide.

 

This is all I can find on the 'fp' front:

 

http://mythtv.org/pipermail/mythtv-dev/2007-January/052824.html

I guess as an alternative approach to program name matching,  the default authority may be useful to construct a unique key for testing for repeats in the event that one or more broadcasters use similar series / episode strings over time. However it only serves as noise for this exercise.  

 

Good work on the PCH scripts, just the job.

 

Rgds

 

Steve

[me]

Hi Steve, I use http://www.schedulesdirect.org for my guide data -- I
think they are US only but you can check them out to see if the cover
Freesat.

I've uploaded a new tarball with an updated script and helper script
at http://mythtv.mlaronson.com/myth-xbmc-connect/downloads -- it shows
up as V1.4 on of "xbmc_mythlink.tgz". I already had logic for using
starttime instead of episode number of generic episodes, as generic
episodes all have the same episode number of "0000". I changed it so
that it should now do this for your recordings. Let me know how it
works.

Marc

On Sep 8, 2:40 am, Steve Berrett <sberr...@googlemail.com> wrote:
> Hi Marc,
>
> Full extract of "recorded" table attached. Not much good news here;
> the "seriesid" field contains a similar mix of data.
>
> I was wondering, rather than reinventing the wheel with complicated
> scripting, does the EPG feed you use cover Freesat?  It is on my to-do list
> to configure a more comprehensive guide.
>
> This is all I can find on the 'fp' front:
>
> http://mythtv.org/pipermail/mythtv-dev/2007-January/052824.html
>
> I guess as an alternative approach to program name matching,  the default
> authority may be useful to construct a unique key for testing for repeats in
> the event that one or more broadcasters use similar series / episode strings
> over time. However it only serves as noise for this exercise.
>
> Good work on the PCH scripts, just the job.
>
> Rgds
>
> Steve
>

> On 8 September 2010 07:20, me <marc.aron...@gmail.com> wrote:
>
> > Hi Steve,
>
> > Looks like I made some incorrect assumptions about the format of the
> > programid field -- that's the risk of "reverse engineering" the
> > behavior from a single data point (my system). I'll figure something
> > out -- need to sort out what to do for episodes like "murder she
> > wrote" that has a program id of  "fp.bbc.co.uk/4j5dsj". I think it
> > would be helpful if you sent me a full extract of the "recorded" table
> > -- I probably need to find another field to key off of. Any guesses as
> > to what "fp" stands for?
>
> > In terms of the popcorn hour interface -- if you want to see what I've
> > built, take a look at
> >http://mythtv.mlaronson.com/networked-media-tank/screenshots
>
> > It's not fancy, but it does the job.
>
> > Marc
>
>
>

>  recordedfulldump.xls
> 53KViewDownload

[Steve Berrett]

Marc

 

Many thanks for this; I'll give it a go at the w/e and let you know.

 

Cheers

Steve