[mod_python] Accessing objects loaded with PythonImport

On 31/05/2005, at 6:51 AM, Nicolas Lehuen wrote:

> Hi Lee,
>
> Unfortunately the answer is not as simple as it seems.
>
> Try doing this in xmlhandler.py :
>
> from mod_python import apache
>
> from os.path import dirname
> this_directory = dirname(__file__)
> site_setup = apache.import_module("site_setup",path=[this_directory])
>
> check = str(dir())
>
> def handler(req):
>            ...
>            req.write(req.interpreter + '<br />')
>            req.write(check)
>            ...
>            return apache.OK
>
> We use apache.import_module to import the site_setup module, looking

Just to clarify Nicolas' answer a bit.

Although PythonImport imports a module at startup, it is still a  
distinct module
and thus why you don't see it within the globals of your handler. Even  
when you
use apache.import_module() as above, you still need to refer to the  
module to
get the data. Thus, to see your data, should change:

   check = str(dir())

to:

   check = str(dir(site_setup))

and reference data as:

   site_setup.magic_string

as appropriate.

In respect of Nicolas' comment in his other email, ie.,

> Plus, if
> you have many site_setup.py in different directories, they will
> collide with each other.

True, at least where we are talking about the same interpreter. As you  
are
talking about different interpreters for each virtual host, you could  
have more
than one. You would though need to ensure that the files were in  
different
directories and each interpreter set the Python path to only the one  
directory
appropriate to that interpreter.

If your intent is for this module only to be used for configuration, as  
Nicolas
pointed out, you may want to have a look at Vampire, which includes a  
configuration
mechanism based on use of Python ConfigParser module.

In your case with multiple sites, you appear to be reserving a "config"  
directory
for common stuff in which case with Vampire all you would need to do is  
create a
file called something like "site_setup.cf" in the "config" directory.  
It may for
example contain:

     [Settings]

     magic_string = Ooggaa-Booggaa!

In your handler you could then say:

     from mod_python import apache
     import vampire

     def handler(req):
         config = vampire.loadConfig(req,"config/site_setup.cf")
         magic_string = config.get("Settings","magic_string")
         req.content_type = "text/plain"
         req.send_http_header()
         req.write("magic_string = %s"%magic_string)
         return apache.OK

This will work whether or not you actually use Vampire for its  
mechanism for
dispatching to different handlers based on resource name. Ie., it  
should work
with a standard handler of your own creation.

Further, this handler could be in subdirectory, ie., not in the root  
directory
where "config" exists, as the configuration mechanism will search back  
up the
hierarchy to the root of where PythonHandler is defined (as necessary),  
looking
for "config/site_setup.cf". This means you do not need to hard code any  
paths
to anything in either the Apache configuration or your handlers.  
Because the
configuration file is looked up when required, you can change it  
whenever you
want and do not need to restart Apache for changes to be reflected.

There is a little bit of information on the configuration mechanism in  
Vampire
in the overview for it:

    
http://www.dscpl.com.au/projects/vampire/articles/vampire 
-001.html#configuration-mechanism

Hope this is of interest.

Graham