EPG Problem

Forums › Home Theatre (including HTPC and EPG Collector) › EPG Problem

bot

17 posts

Geek

#20774 5-Apr-2008 16:46

A while ago I was having some problems with the epg files downloaded from http://epg.pvr.geek.nz/epg/. These are described in the topic http://www.geekzone.co.nz/forums.asp?ForumId=34&TopicId=19844.

After downloading a new file the problems were resolved. But over the last few days I seem to be having the same problem.

This shows up as a patchy tv guide in VMC with mostly "No data available". The programme data that does show is clearly not right e.g. Nigella Feasts on Sky Movies Greats.

I have checked the contents of http://epg.pvr.geek.nz/epg/ which shows up as:

Index of /epg

Name Last modified Size Description

Parent Directory -

icons/ 12-Mar-2008 09:26 -

listings-all.xml.gz 05-Apr-2008 13:12 228K

listings-freeview.xml.gz 05-Apr-2008 13:12 133K

listings-sky.xml.gz 05-Apr-2008 13:12 92K

listings.xml.gz 05-Apr-2008 13:12 120K

I then looked at a couple of the files and immediately saw some inconsistencies. For example:

- Some oddities from the listings-all file:

<title lang="eng">Te Karere</title>

<desc>Te Karere</desc>

<category>tvshow</category>

</programme>

...

<title lang="eng">Shortland Street</title>

<desc>Tuesday decides to reinvent herself. Sarah is driven to extremes. Libby uses Gerald to put off Sunil.</desc>

<category>series</category>

<category>Drama</category>

</rating>

</programme>

- Some more from the listings-sky file:

<title lang="eng">ONE News Tonight</title>

<desc>ONE News Tonight</desc>

<category>tvshow</category>

</programme>

<title lang="eng">Tagata Pasifika</title>

<desc>Tagata Pasifika</desc>

<category>tvshow</category>

</programme>

I am not complaining (the folks that make these files available obviously go to a good deal of trouble for us all), but I wonder whether we can help to resolve this.

Thanks.

openmedia

3449 posts

Uber Geek
+1 received by user: 878

Trusted

#121373 5-Apr-2008 19:26

The problem is that the EIT data from freeview is off by an hour from Sunday as they didn't take into account daylight savings stopping.

Chances are the data will correct itself on Sunday or Monday.

Steve

Generally known online as OpenMedia, now working for Red Hat APAC as a Technology Evangelist and Portfolio Architect. Still playing with MythTV and digital media on the side.

CYaBro

4708 posts

Uber Geek
+1 received by user: 1182

ID Verified

Trusted

#121396 5-Apr-2008 21:52

That doesn't explain why ONE News is showing under Sky Movie Greats.

Opinions are my own and not the views of my employer.

bot

17 posts

Geek

#121490 6-Apr-2008 12:19

I had a look at the files this morning and they look healthier. Partially because the files are back to their previously (large) sizes:

Index of /epg

Name Last modified Size Description

Parent Directory -

icons/ 12-Mar-2008 09:26 -

listings-all.xml.gz 06-Apr-2008 02:09 1.0M

listings-freeview.xml.gz 06-Apr-2008 02:09 143K

listings-sky.xml.gz 06-Apr-2008 02:09 867K

listings.xml.gz 06-Apr-2008 02:09 876K

Also, I can't know find any obvious anomalies with the file contents.

Whilst I am happy we are back working, I am still keen to understand what is happening.

File size seems to be a related factor (last time this happened I recall there were file size anomalies). Listings-all.xml.gzip was 228K yesterday (with the dodgy content) and is 1MB today (with apparently healthy content).

So I started looking back at my wget logs for previous downloads and found a pattern of smaller files:

20080220wget.log:Length: 999,010 [application/x-gzip]

20080222wget.log:Length: 1,000,222 [application/x-gzip]

20080224wget.log:Length: 945,636 [application/x-gzip]

20080228wget.log:Length: 1,038,758 [application/x-gzip]

20080301wget.log:Length: 62,098 [application/x-gzip]

20080304wget.log:Length: 1,248,900 [application/x-gzip]

20080305wget.log:Length: 1,223,994 [application/x-gzip]

20080307wget.log:Length: 1,253,960 [application/x-gzip]

20080309wget.log:Length: 1,239,481 [application/x-gzip]

20080311wget.log:Length: 1,245,327 [application/x-gzip]

20080313wget.log:Length: 1,265,271 [application/x-gzip]

20080315wget.log:Length: 189,730 [application/x-gzip]

20080316wget.log:Length: 1,183,180 [application/x-gzip]

20080317wget.log:Length: 1,224,892 [application/x-gzip]

20080323wget.log:Length: 1,186,994 [application/x-gzip]

20080325wget.log:Length: 1,192,287 [application/x-gzip]

20080327wget.log:Length: 1,242,113 [application/x-gzip]

20080329wget.log:Length: 193,119 [application/x-gzip]

20080331wget.log:Length: 1,184,034 [application/x-gzip]

20080331wget.log:Length: 1,207,164 [application/x-gzip]

20080402wget.log:Length: 1,199,319 [application/x-gzip]

20080404wget.log:Length: 1,210,480 [application/x-gzip]

20080405wget.log:Length: 233,813 [application/x-gzip]

20080406wget.log:Length: 1,072,914 [application/x-gzip]

The days with apparently dodgy files are all Saturdays...

I got the impression from somewhere that hosting of http://epg.pvr.geek.nz/epg/ is shared between several hosts and rotates between them. I wonder if the server hosting the files on Saturdays is dodgy.

Then again maybe I'm on a wild goose chase :-)

bot

17 posts

Geek

#123210 13-Apr-2008 16:10

The tradition of weird files on Saturdays (~one fifth their normal size and with programme entries that are obviously wrong) continues...

Recent file sizes:

20080408wget.log:Length: 1,220,494 [application/x-gzip]

20080412wget.log:Length: 209,231 [application/x-gzip]

20080413wget.log:Length: 1,232,914 [application/x-gzip]

Some dodgy programmes (not likely to feature on sky highlights channel I don't think):

<title lang="eng">Ugly Betty</title>

<desc>Henry reluctantly agrees to go on a double date with Betty and an unlikely couple - Hilda and Gio. Meanwhile, Wilhelmina's attempt to change her image as the 'queen of mean' fails miserably.</desc>

<category>series</category>

<category>Comedy</category>

</rating>

</programme>

<title lang="eng">Grey's Anatomy</title>

<desc>Part two of two. Meredith furiously works to save the life of a paramedic trapped inside an ambulance, while Derek must rely on the help of a nurse to save a patient's life. Guest star Seth Green.</desc>

<category>series</category>

<category>Drama</category>

</programme>

</tv>

hads

392 posts

Ultimate Geek
+1 received by user: 12

Trusted

#123353 14-Apr-2008 09:50

Very interesting. Well spotted that it only happens on Saturday, perhaps there is something conflicting with the EPG collection run on the host that's doing it on Saturdays.

So just to be clear, what time does your cron job download the data? From the directory listing in the first post it looks like it's after the afternoon run (3pm), and the data appears to be back to normal after the night run (3am).

nicegear.co.nz - VoIP, DVB and other Linux compatible hardware

bot

17 posts

Geek

#123858 15-Apr-2008 22:33

Sorry for the delayed reaction...

The scheduled job runs at 9.25pm every two days. So you are right about the Saturday file being downloaded after 3pm, the timestamp in the epg file document element shows the file to be generated at 15:09:41.

The non-dodgy file I downloaded on Sunday would most likely have been generated at 3pm again since I downloaded it after 4pm. So I am not sure if the 3am Saturday night version was good or not.