o analyze all files in the repository, do something like this
    (note: this is not self-contained code)::

        >>> repos = np.lib._datasource.Repository('/home/user/data/dir/')
        >>> for filename in filelist:
        ...     fp = repos.open(filename)
        ...     fp.analyze()
        ...     fp.close()

    Similarly you could use a URL for a repository::

        >>> repos = np.lib._datasource.Repository('http://www.xyz.edu/data')

    c