It would be great if we could have something like:

::

    >>> from mediagoblin.util import parse_app_config
    >>> parse_app_config(path_to_config_file, app_section='app:mediagoblin')
    (app_config, global_config)

where the former is the app config (a dictionary of all values you
see in [app:mediagoblin] in mediagoblin.ini) and the latter is the
global config (a dictionary of dictionaries for all sections).

This tool kind of already exists in
mediagoblin/celery\_setup/**init**.py, you can see a section like:

::

        parser = NicerConfigParser(mgoblin_conf_file)
        parser.read(mgoblin_conf_file)
        parser._defaults.setdefault(
            'here', os.path.dirname(os.path.abspath(mgoblin_conf_file)))
        parser._defaults.setdefault(
            '__file__', os.path.abspath(mgoblin_conf_file))
    
        mgoblin_section = dict(parser.items(mgoblin_section))
        mgoblin_conf = dict(
            [(section_name, dict(parser.items(section_name)))
             for section_name in parser.sections()])

This should be:


-  separated out into an independent utility
-  given tests in mediagoblin/tests/test\_util.py
-  able to take the specific section as the app\_section keyword
   argument (defaults to 'app:mediagoblin')
-  bonus feature? If 'app:mediagoblin' is not found maybe we could
   find the section by iterating through until we find the one that
   has 'use': 'egg:mediagoblin#app'

We need to be able to support translations! This is a multi-step
process:


-  we need to be able to put
   `translations in the templates <http://jinja.pocoo.org/docs/extensions/#i18n-extension>`_
-  we need to be able to put them in normal python code
-  we need to be able to extract translations
-  we need to put translations up somewhere where people can start
   translating them (Transifex?)

We have a storage backend that supports multiple media types, but
it's possible that certain types of media might require local files
to be processed through locally. For example, maybe we'll need to
have a local copy of a movie before we do transcoding. Etc.

Maybe it would be helpful to create a BasicFileStorage instance
that writes to a directory either made in /tmp/ or user specified
somewhere else (could default to
./user\_dev/media/operating\_table/ in the config, but if
completely not found in the config, could make a temporary
directory to dump things in?), but which keeps non-permanent copies
of files and local instances that will allow the files to be
"operated" on while moving from the queue\_storage ->
public\_storage during processing.