Using Eversion as dynamic Jukebox - An Idea

[vaidyasr]

I have a basic high level idea to integrate Eversion with Oversight. It may be feasible, but not now.

Now Eversion is opening the static XML files, so if we can modify the eversion to either use a PHP script to create categories.xml on the fly or call the oversight.cgi to get the categories.

So the things required to achieve the above.
The oversight has to be edited to give the xml output when the URL oversight.cgi has been called, may be with arguments. This may be faster, as at the backend it interacts with OS level instead of php. But you may not able to use the actual features of oversight.cgi. So, instead of modding the oversight, we can probably write a separate program which uses the oversight libraries and can be used. The advantage would be no intermediate building of XML files. And the index is always updated one.

From the current DB, it is very time consuming process to fetch the categories available. So, if the oversight DB has been ported to SQLite3 possibly we can fetch the categories in seconds instead of iterating the whole movies collection.

[lordy]

Hi, this is interesting. I haven't played with the new integrations yet , but will start after the next stable Oversight release.

[vaidyasr]

I just did a first test, by editing the eversion source to call the php to get the categories.xml instead of static and it works fine with 2-3 seconds delay. The speed can be improved if we switch the backend from flatfile to SQlite3 db.

[accident]

You should have asked, there's settings for what you wanted to try, no source code changes needed

There's a config file setting for where the jukebox is located.. you could swap it from ./ to anything.. I use http:// for my jukebox files at home. It's faster, flash halts when it's touching the drive.

something liek http://pch/api.php?xmlfile= would be valid.. then it'l ljust ask for the xml filename at the end.

Only source code changes are needed if you can't work with the categories.xml and index filenames.

also it's gonig to look to that url for all artwork also

---

btw, this plan in my mind is a mistake and just throwing together crap to make it work.. make a proper api that could be used by tablets, future players when the chips can handle html5 browsers with ajax. Also swap the data to json it's so much faster than xml.

Then add proper stable support to eversion for it adding a new data class for this api.. you'll gain a lot more speed and much more options for what to do with the api. If you mimic the plex or xbmc api then you could even work with those systems as well as oversight work with xbmc programs like their tablet apps.

[vaidyasr]

Yes, I have already seen the config file, but I was testing by replacing the existing 'Categories.xml' call in the source code to full url and it works. Also, the additional XML's were loading from the Categories.xml file too. So, need to find a way to load them too via dynamic.

[vaidyasr]

By editing the config file, the Categories page is loading fine. But when I select any category, it says in a message box 'ERROR' with OK button. But I can able to find the argument to the php file is going and the output is also good when I run the php. But eversion is not taking it. Need to debug the program.

[accident]

what's the output of your categories.xml look like? the error message shouldn't be empty so I think your missing what it should be calling in categories

[vaidyasr]

The output of the Other_HD_1.xml is below

Code:

<?xml version='1.0' encoding='UTF-8'?>
<library>
  <category name="Other" current="true" count="1">
    <index name="HD" current="true" first="Other_HD_1" previous="Other_HD_1" next="Other_HD_2" last="Other_HD_3" currentIndex="1.xml" lastIndex="3">Other_HD_1</index>
  </category>
  <movies count="27" cols="9">
    <movie isExtra='false' isSet='false' isTV='false'>
      <id moviedb='imdb'>tt1521197</id>
      <id moviedb='themoviedb'>61891</id>
      <indexedDate>2012-01-21 02:29:00</indexedDate>
      <baseFilenameBase>details_1288</baseFilenameBase>
      <baseFilename>details_1288</baseFilename>
      <title>Anonymous_</title>
      <titleSort>Anonymous</titleSort>
      <originalTitle>Anonymous</originalTitle>
      <year Year='2010-12'>2011</year>
      <releaseDate>UNKNOWN</releaseDate>
      <rating>68</rating>
      <ratings>
         <rating moviedb='imdb'>68</rating>

I have truncated the output, as it is around 1800 lines. The XML is valid and tested in chrome and IE.

[accident]

currentIndex="1.xml" should be 1, not 1.xml.. it uses this to know what file it's in

[vaidyasr]

OK, thanks it works fine and fast too. I'll check some more and post the results.

---

It seems eversion doesn't use the detail.xml at all. It only uses only the categories.xml and All categories.xml and find the info about each movie. Am I right accident?.

But setting the YAMJ path to a php file, each jpg request also comes to the php script. I need to find a way to bypass if the request is an image.

[accident]

nope, it doesn't use details files.

same flash stops touching the drive applies to images.. if you could return the image with the correct header it would speed it up. everything without.xml on the end is an image request

[vaidyasr]

So if it doesn't use details files, possibly we can skip building the xml files in the current build.php. Yes, the image loading has been done now.

[accident]

(03-26-2012 10:22 AM)vaidyasr Wrote:  So if it doesn't use details files, possibly we can skip building the xml files in the current build.php. Yes, the image loading has been done now.

I don't use them.. don't know if you need them to insert into an index

[vaidyasr]

The page no. on eversion is required?. The conventional HTML skin may need to load every page, so the page number is required. For Eversion, it is stays on same page and doing a scolling. Is it required?.

[accident]

I don't understand what you mean by page number.. what data point?