Free XMLTV EPG scraper

[dneprrider]

In previous threads alternatives to the zap2it TV guide have been mentioned.

One very usable and easy to setup alternative is available at:

http://planetreplay.com/phpBB2/viewtopic.php?t=14314

I thought I would start a separate thread for this free XMLTV EPG scraper, as I believe a few other GBPVR users have tried this little utility. Some additional work is required to get it to work better with GBPVR.

First a few good points -
- It is free
- Does not required MCE or VISTA
- Supports Canadian users - including OTA & ExpressVu & Starchoice

Now a problem I am having

I can download program listings through the utility, and set the EPG source in config.
But in the GBPVR TV guide I can see all the channels - but the message (no show details) are displayed for every time slot. But I do see program/TV show details in the generated XML file.

After talking with drlava, it may be that gbpvr dosen't read in more than one <display-name> tag per channel. Or where GBPVR gets the channel number from - the ID, or the first <display-name> tag.

So there are a few questions that I am sure others here will be interested in helping out with.

[whurlston]

In previous threads alternatives to the zap2it TV guide have been mentioned.

One very usable and easy to setup alternative is available at:

http://planetreplay.com/phpBB2/viewtopic.php?t=14314

I thought I would start a separate thread for this free XMLTV EPG scraper, as I believe a few other GBPVR users have tried this little utility. Some additional work is required to get it to work better with GBPVR.

First a few good points -
- It is free
- Does not required MCE or VISTA
- Supports Canadian users - including OTA & ExpressVu & Starchoice

Now a problem I am having

I can download program listings through the utility, and set the EPG source in config.
But in the GBPVR TV guide I can see all the channels - but the message (no show details) are displayed for every time slot. But I do see program/TV show details in the generated XML file.

After talking with drlava, it may be that gbpvr dosen't read in more than one <display-name> tag per channel. Or where GBPVR gets the channel number from - the ID, or the first <display-name> tag.

So there are a few questions that I am sure others here will be interested in helping out with.

GB-PVR will only read the first display-name attribute. The channel= attribute in the <programme> tag should match the id= attribute in the <channel> tag. The display-name should have no effect on the programs.

Code:

<channel id="001">
    <display-name>LOOR001</display-name>
  </channel>
  <programme start="20070802020000 +0000" stop="20070802060000 +0000" channel="001">
    <title>Local Origination</title>
  </programme>

It sounds as if the output of that program is not truly xmltv compliant. Do you have an output file you can post? OTA would be great because it shouldn't be too big.

[whurlston]

I tested the program and it looks like the issue is the way they write the XML output. The write the output as:

Code:

<channel id="1">
<programme channel="1" />
<channel id="2">
<programme channel="2" />
<channel id="3">
<programme channel="3" />

Which is not the correct order. It should be all channel info first then all the program info:

Code:

<channel id="1">
<channel id="2">
<channel id="3">
<programme channel="1" />
<programme channel="2" />
<programme channel="3" />

Also, you would have to list the display-name that you want to use first. The first display-name that is listed only has the channel number.

sub would be able to verify if I am right about the required order.

[drlava]

Other programs, such as sageTV, do not assume that the channel ID is the channel number. They get the channel number from the fisrt numeric-only <display-name>. In fact, if you look at the official XMLTV description, the station ID they use in their example is the descriptive text ID, not the channel number.
http://xmltv.cvs.sourceforge.net/xmltv/x...iew=markup

The XMLTV document specifies that the first display name should perhaps be the channel number, and after that, order them in terms of obscurity.

Perhaps GBPVR could accept multiple display-names to accomodate the additional information they provide, such as affiliates?

It depends on how GBPVR is coded as to whether it can deal with the channel IDs not all at the beginning. Most XML COM interfaces wouldn't care the order, someone who has seen the GBPVR source could comment on this...

[whurlston]

Other programs, such as sageTV, do not assume that the channel ID is the channel number. They get the channel number from the fisrt numeric-only <display-name>. In fact, if you look at the official XMLTV description, the station ID they use in their example is the descriptive text ID, not the channel number.

Neither does GB-PVR. It will take the channel number from either of the first two display-name tags that you are outputting. The issue is that it ONLY reads the FIRST display-name tag which is only a channel number in your output. I was just lazy in my above examples.

The main issue is that GB-PVR expects that the channels will all be listed before the programs as per the XMLTV.DTD (shown at http://xmltv.cvs.sourceforge.net/xmltv/x...iew=markup)

It depends on how GBPVR is coded as to whether it can deal with the channel IDs not all at the beginning. Most XML COM interfaces wouldn't care the order, someone who has seen the GBPVR source could comment on this...

Only sub would be able to comment on the source. It would not be hard for me to write a program that would rearrange the order of the output file so that it is compatible GB-PVR. If I get some time this weekend I may do that. It would probably be the simplest solution.

[sgilani]

Wow, great find.. this gem looks very promising!

[whurlston]

Ok, I rearranged the output and it turns out that is not the issue after all. The issue is that there is not stop time in the programme tag and the timezone offset is only 3 digits.

Code:

<programme channel="28456298" start="20070804000000 -300">

It should be:

Code:

<programme channel="28456298" start="20070804000000 -0300" stop="20070804010000 -0300">

I can easily fix this but it will be a few days before I can get to it.

[drlava]

So from the sounds of it only two changes are necessary:?

I can change the time zone offset to be zero padded to four digits on the source, no problem. The stop is an optional attribute according to the XMLTV documentation, but that could be added in the original source, if GBPVR relies on it. The WiRNS reader uses top or duration to calculate the end time of the programme.

[sub]

It'll cope without the stop time, but of course means the last listing in each channel be dropped since the stop time for these cant be derived.

<channel id="1">
<programme channel="1" />
<channel id="2">
<programme channel="2" />
<channel id="3">
<programme channel="3" />

It should matter what order the channel or programme tags are supplied, as long as the referenced channel is supplied before the programme, so this example should be ok, though I've never seen an xmltv file in that format so never tested it.

[whurlston]

Thanks sub. I only changed one or two entries so that may be why it ddnt work for me without the stop time.