BOINC Combined Statistics

Hello!

Wanted to first point out a typo in the stats URL given in the thread in Announcements. It points to http://boinc.netsoft-online.com/stats/l where it should be just http://boinc.netsoft-online.com/stats/

Secondly I was wondering of the possibility of an XML file aggregated by project? Something that gives the stats at the top of the page at http://boinc.netsoft-online.com/e107_plugins/boinc/bp.php?project=1 but in XML format. Even better would be to list the stats of all supported projects in a single XML file. My idea for this is that I'm looking to build a page of movers and shakers by project, and not having to process full user dumps for this would be ideal.

I actually have a question regarding the xml, the users.xml published on this website appears to contain mallformedness errors, for example, row 1090, column 15, a "rogue" ampersand (&) appears.

I'm trying to process the file, but none of the parsers I've tried can't get beyond that point (SAX/Xerces, XOM, StAX/Woodcrest). The environment I'm in is Java.

Is it possible to escape these or can you suggest an xml parser that can do this (and have it stream-able)?

An example user (the one above) is c5ec61c267a96cb37d499079c7895b4d

The same thing happens with individual users on this site (ie. http://boinc.netsoft-online.com/get_user.php?cpid=c5ec61c267a96cb37d499079c7895b4d).

But that user at for example http://einstein.phys.uwm.edu/ does have an escaped & but not on this site.

I hope this is can be fixed, since it isn't proper XML (otherwise) and I haven't been able to find an XML (stream) parser that can ignore part of the XML specifications.

*uber sigh* No, I'm not ignoring the comments, just very busy.

The url in the announcement should be fixed.

Escaping the xml chars in the get_user web rpc should be fixed.

user export file should now be fixed as well.

Ran out of time for tonight, but will work on your request AlphaLaser soon.

Escaping the xml chars in the get_user web rpc should be fixed.

James

What is that? PHP already has it built-in: htmlspecialchars.

I looked at what was stored in the name fields in my DB - sometimes the entries were already escaped and sometimes not (not entirely sure why), so I couldn't just use the php function or in some cases you end up with a real mess.

I looked at what was stored in the name fields in my DB - sometimes the entries were already escaped and sometimes not (not entirely sure why), so I couldn't just use the php function or in some cases you end up with a real mess.

James

function escape_xml_special_chars ($text) {
    $text = htmlspecialchars_decode($text);
    $text = htmlspecialchars($text);
    return $text;
}

Thanks James!

That saves me from writing my own xml parser (or maiming an existing one).

I've just downloaded it, and it seems to be partly resolved.

The name element of user c5ec61c267a96cb37d499079c7895b4d still has an unescaped & in it.
L&S IT Office

The project entry doesn't now.

But the url of user 93fd38421af91db4c3868f6128c9955e does have one in the Einstein@Home user_url element.
Is it possible that the import hasn't been done with the proper xml producing code yet?

[quote]

I looked at what was stored in the name fields in my DB - sometimes the entries were already escaped and sometimes not (not entirely sure why), so I couldn't just use the php function or in some cases you end up with a real mess.

James

function escape_xml_special_chars ($text) {
    $text = htmlspecialchars_decode($text);
    $text = htmlspecialchars($text);
    return $text;
}

[/quote1207230790]

That's fine for HTML, but not XML. The apostrophe is not translated correctly by the htmlspecialchars function (xml standard says it should end up as ' not #039 see http://www.w3.org/TR/REC-xml/ ). Looks like I mised the double quote yet. *sigh*