A few hints needed for MediaTitles.xml

[davidk21770]

I'm trying to write some match strings and I'm missing what the input and output strings SHOULD look like? What am I matching?

For example, why is there a leading and trailing test for \ or /?

Thanks!

[stuckless]

Originally Posted by davidk21770

I'm trying to write some match strings and I'm missing what the input and output strings SHOULD look like? What am I matching?

For example, why is there a leading and trailing test for \ or /?

Thanks!

David, perhaps you can post some examples, and I can help you fill in the blanks.

MediaTitles.xml accepts a complete file path as the input. Let me break down the sample...

	<match>
		<regex>[\\/]Babylon\s*5[\\/]</regex>
		<title>Babylon 5</title>
		<year>1993</year>
		<metadata type="tv" name="tvdb">7072</metadata>
	</match>

The about matcher uses a regular expression against the filename to force any files that are in a directory called Babylon 5 to have the series id 7072.

In the <regex>, [\\/] means match a forward slash or backslah in the path. In regex speak the [] is a special command that means match any of these characters. The backslash is doubled \\ because a single backslash is a regex escape char, so if you mean a literal backslash, then you use \\. Because I tend to want to support any platform then I need to use \\ and / to support both windows \\ and linux/mac /. So, in the above regex, it will match a section of the path that looks like
\Babylon 5\ or /Babylon 5/, and because of the optional whitespace (\s*) it will also match \Babylon5\ or \Babylon 5\, etc.

The <title> and <year> are not required in this case, since you are specifically providing <metadata tv="tvdb"> tag to specify the series id. I tend to put them in, just for "documentation". But if you didn't specify a series, then you could simply provide the <title> and optionally the <year> and bmt will dynamically look it up.

The MediaTitles is good for forcing an entire directory to a specific series, or something if you don't want to rename your files to include a better title/year, then you can force it the media titles xml.

Personally, I tend to name my stuff better.... ie, for Movies, I make sure that I include the year in the filename and for TV, I also try to name them how they are named in TVDB. Sometimes just renaming the files, is better than messing with the MediaTitles.xml.

Movie Title (Year).avi
Series Name S##E##.avi

If you have something not in a standard naming format, then the other option is to create your own scraper that will scrape your files correctly.

[davidk21770]

Ok. I missed the "file path" part.
So, if I'm matching (say) CSI: Crime Scene Investigation, it normally comes in something like:
CSICrimeSceneInvestigation-HouseofHoarders-8688341-0.mpg
So, my match'd be
<regex>[\\/]CSICrimeSceneInvestigation-*</regex>

Edit:
Now, if I'm getting an imported file of the form Sanctuary.US.xxxx do I match to "Sanctuary US" or "Sanctuary.US"?

[stuckless]

Originally Posted by davidk21770

Ok. I missed the "file path" part.
So, if I'm matching (say) CSI: Crime Scene Investigation, it normally comes in something like:
CSICrimeSceneInvestigation-HouseofHoarders-8688341-0.mpg
So, my match'd be
<regex>[\\/]CSICrimeSceneInvestigation-*</regex>

That looks correct.

Edit:
Now, if I'm getting an imported file of the form Sanctuary.US.xxxx do I match to "Sanctuary US" or "Sanctuary.US"?

<regex>[\\/]Sanctuary.US.*</regex>
<title>Sanctuary</title>

[davidk21770]

Thanks!

[ohpleaseno]

So, just so I understand this correctly, any show that BMT is having difficulty matching, we can add additional lookups in this one file with the show title in the regex field and tvdb number in the metadata field.

          <match>
		<regex>[\\/]This Old House[\\/]</regex>
		<metadata type="tv" name="tvdb">WHATEVER THE TVDB ID IS HERE</metadata>
	  </match>

          <match>
		<regex>[\\/]New Yankee Workshop[\\/]</regex>
		<metadata type="tv" name="tvdb">WHATEVER THE TVDB ID IS HERE</metadata>
	</match>

[stuckless]

That is correct. the Regular Expression matches a portion of the filename (and path) to whatever you provide. So in your case, it will force the lookup to be tv (with the given tvdb id) if the file is in a directory that exactly matches, "This Old House".

[ohpleaseno]

When you say "in a directory" does that mean it only works for imported? I'm looking to get the scan working for some PBS shows that don't show up properly for some reason. Will this fix that as well if the filename from recorded TV contains This Old House?

[stuckless]

if you are talking about recordings, then they are usally named something like,

ThisOldHouse-EpisodeTitle-0000000-0.ts

So, if you wanted to match on that, then you could use a starts with regex...

<regex>[\\/]ThisOldHouse-.*</regex>

That basically means match /ThisOldHouse- or \ThisOldHouse- as part of the filename path.

So if the full recording was

c:\tv\recordings\ThisOldHouse-SomeEpisode-000000-0.ts, then it would match. But if the recording was c:\tv\recordings\This Old House.ts then it would not match (because of the spacing and missing dash)

[ohpleaseno]

does sage service need to be restarted for the change in this file to be recognized?

[jusjoken]

Originally Posted by ohpleaseno

does sage service need to be restarted for the change in this file to be recognized?

Yes... at least in older versions it did and I haven't seen any word that it changed.

k

[stuckless]

I think you can reload them from the bmt web ui. It's an option under the Configuration sections.

[ohpleaseno]

Is there a log file that will tell me if things are properly hitting MediaTitles.xml?

[stuckless]

Originally Posted by ohpleaseno

Is there a log file that will tell me if things are properly hitting MediaTitles.xml?

The phoenix.log log file will tell you if it's using the MediaTitles... if you open the file and search for MediaTitles, i think you'll see some messages related to it.

[phelme]

Can we add season and episode elements to MediaTitles.xml to better target the query? A show like Jeopardy! fails because there is no season 1/episode 1 entry, which is the default to search.

2010-12-08 11:39:27,342 [2438065@qtp-18376755-6] INFO  sagex.phoenix.metadata.provider.tvdb.TVDBItemParser - TVDB Episode: http://www.thetvdb.com/api/5645B594A3F32D27/series/77075/default/1/1/en.xml
2010-12-08 11:39:27,451 [2438065@qtp-18376755-6] WARN  sagex.phoenix.metadata.provider.tvdb.TVDBItemParser - Failed to get season/episode specific information for 77075; Season: 1; episode: 1

java.io.FileNotFoundException: http://www.thetvdb.com/api/5645B594A3F32D27/series/77075/default/1/1/en.xml
	at sun.net.www.protocol.http.HttpURLConnection.getInputStream(Unknown Source)

But there is a 24/165 which would work.

Hmm, alternatively, maybe I'll see if I can put in the first episode in thetvdb.com...

Edit: Added entry to thetvdb.com for 1/1. We shall see if that takes care of it.

[stuckless]

Originally Posted by phelme

Can we add season and episode elements to MediaTitles.xml to better target the query? A show like Jeopardy! fails because there is no season 1/episode 1 entry, which is the default to search.

2010-12-08 11:39:27,342 [2438065@qtp-18376755-6] INFO  sagex.phoenix.metadata.provider.tvdb.TVDBItemParser - TVDB Episode: http://www.thetvdb.com/api/5645B594A3F32D27/series/77075/default/1/1/en.xml
2010-12-08 11:39:27,451 [2438065@qtp-18376755-6] WARN  sagex.phoenix.metadata.provider.tvdb.TVDBItemParser - Failed to get season/episode specific information for 77075; Season: 1; episode: 1

java.io.FileNotFoundException: http://www.thetvdb.com/api/5645B594A3F32D27/series/77075/default/1/1/en.xml
	at sun.net.www.protocol.http.HttpURLConnection.getInputStream(Unknown Source)

But there is a 24/165 which would work.

Hmm, alternatively, maybe I'll see if I can put in the first episode in thetvdb.com...

Edit: Added entry to thetvdb.com for 1/1. We shall see if that takes care of it.

It would only search for season 1 episode 1, if it determined that the episode was, in fact, season 1 and episode 1 by scraping the file. BMT has no default season and episode values, they either exist, or the values are -1 (which means they don't exist). But in seeing a values of 1, means that the scrapers pulled that from somewhere.

[phelme]

So basically if there is no date match or season/episode identifier match (as in my Jeopardy! example, where there usually is no exact entry for every show), I'm out of luck at extracting any good data from thetvdb.com? Because currently I simply get metadata lookup failures every night.

[stuckless]

Originally Posted by phelme

So basically if there is no date match or season/episode identifier match (as in my Jeopardy! example, where there usually is no exact entry for every show), I'm out of luck at extracting any good data from thetvdb.com? Because currently I simply get metadata lookup failures every night.

Well, the data has to be in tvdb at the time your do the lookup. Let's say that you record Jeopardy at 8pm but tvdb has nothing for the episode, so at 9pm somone goes in the adds the data... you missed it because the lookup happened before there was data. I've seen this happen many times for some shows.

Recordings are looked up by aired date and episode title (it will try both). So, in the case of jeopardy, if it's not finding, then either the episode does not exist... or the aired date doesn't match, or the episode name doesn't match.

If you use the web ui, the next day (or a couple of days later) to do manually do a lookup, does it find it then? Have you confirmed that tvdb has the data and the data does in fact match your aired date, or episode title?

[phelme]

Originally Posted by stuckless

If you use the web ui, the next day (or a couple of days later) to do manually do a lookup, does it find it then? Have you confirmed that tvdb has the data and the data does in fact match your aired date, or episode title?

Nope, the only dates that have anything are either the celebrity matches or some tournaments that people enter manually. I get that since its a daily game show, there isn't likely to be any reason for it to be updated on a regular basis.

Hmm. Is there anyway I can tell BMT to ignore Jeopardy! so I don't get the daily failure messages?

[stuckless]

Originally Posted by phelme

Nope, the only dates that have anything are either the celebrity matches or some tournaments that people enter manually. I get that since its a daily game show, there isn't likely to be any reason for it to be updated on a regular basis.

Hmm. Is there anyway I can tell BMT to ignore Jeopardy! so I don't get the daily failure messages?

You can add an exclude filter, I'd try something like...

[/\\]Jeopardy-

See the Phoenix Configuration section of the BMT User Guide for more information (specifically the table entry for the exlude filter at the end of the section)