Nice one :)
works fine for me (tested on http://liris.cnrs.fr/~pchampin/tmp/ns).
An interesting thing, though, is that
http://champin.net/tmp/ns/voc1
does *not* redirect with a 303 to voc1.html or voc1.rdf, but instead
return a 200, but with a "Content-Location" header field containing the
appropriate location (i.e. voc1.html or voc2.html). This is therefore
not complying with the letter [HttpRange-14].
Note that I do believe that it is complying with the spirit of
[HttpRange-14], though. I interpret Content-Location as a kind of
shortcut-redirection, which spares the client and the server the burden
of a 2nd query whenever the server is able to produce the queried
content anyway. So it can be considered in a sense as an optimized 303
redirection. (I actually discussed that very matter a month ago with
Harry Halpin at ESWC).
So, is this actually a common practice in the LOD community? Does this
look like a good practice (it does to me)? If so, should HttpRange-14 be
amended to acknowledge that?
pa
[HttpRange-14]
http://www.w3.org/2001/tag/doc/httpRange-14/2007-05-31/HttpRange-14
Le 05/07/2009 16:16, Hugh Glaser a écrit :
OK, I'll have a go :-)
Why did I think this would be fun to do on a sunny Sunday morning that has
turned into afternoon?
Here are the instructions:
1. Create a web-accessible directory, let's say foobar, with all your .rdf,
.ttl, .ntriples and .html files in it.
2. Copy lodpub.php and path.php into it.
3. Access path.php from your web server.
4. Follow the instruction to paste that text into .htaccess
5. You can remove path.php if you like, it was only there to help you get
the .htaccess right.
That should be it.
The above text and files are at
http://www.rkbexplorer.com/blog/?p=11
Of course, I expect that you can tell me all sorts of problems/better ways, but
I am hoping it works for many.
Some explanation:
We use a different method, and I have tried to extract the essence, and keep
the code very simple.
We trap all 404 (File not Found) in the directory, and then any requests coming
in for non-existent files will generate a 303 with an extension added,
depending on the Accept header.
Note that you probably need the leading "/" followed by the full path from the domain
root, otherwise it will just print out the text "lodpub.php";
(That is not what the apache specs seem to say, but it is what seems to happen).
If you get "Additionally, a 404 Not Found error was encountered while trying to use
an ErrorDocument to handle the request.", then it means that web server is not
finding your ErrorDocument .
Put the file path.php in the same directory and point your browser at it - this
will tell you what the path should be.
Note that the httpd.conf (in /etc/httpd/conf) may not let your override, if
your admins have tied things down really tight.
Mine says:
AllowOverride All
Finally, at the moment, note that I think that apache default does not put the
correct MIME type on rdf files, but that is a separate issue, and it makes no
difference that the 303 happened.
Best
Hugh
On 05/07/2009 01:52, "Pierre-Antoine Champin"<[email protected]>
wrote:
Le 03/07/2009 15:14, Danny Ayers a écrit :
2009/7/2 Bill Roberts<[email protected]>:
I thought I'd give the .htaccess approach a try, to see what's involved in
actually setting it up. I'm no expert on Apache, but I know the basics of
how it works, I've got full access to a web server and I can read the online
Apache documentation as well as the next person.
I've tried similar, even stuff using PURLs - incredibly difficult to
get right. (My downtime overrides all, so I'm not even sure if I got
it right in the end)
I really think we need a (copy& paste) cheat sheet.
Volunteers?
(raising my hand) :)*
Here is a quick python script that makes it easier (if not completely
immediate). It may still requires a one-liner .htaccess, but one that (I
think) is authorized by most webmasters.
I guess a PHP version would not even require that .htaccess, but sorry,
I'm not fluent in PHP ;)
So, assuming you want to publish a vocabulary with an RDF and an HTML
description at http://example.com/mydir/myvoc, you need to:
1. Make `myvoc` a directory at the place where your HTTP server will
serve it at the desired URI.
2. Copy the script in this directory as 'index.cgi' (or 'index.wsgi' if
your server as WSGI support).
3. In the same directory, put two files named 'index.html' and
'index.rdf'
If it does not work now (it didn't for me),you have to tell your HTTP
server that the directory index is index.wsgi. In apache, this is done
by creating (if not present) a `.htaccess` file in the `myvoc`
diractory, and adding the following line::
DirectoryIndex index.cgi
(or `index.wsgi`, accordingly)
There is more docs in the script itself. I think the more recipes
(including for other httpds) we can provide with the script, the more
useful it will be. So feel free to propose other ones.
enjoy
pa
