[MrEx3cut0r]

I added query normalize function to the MPDS Retrieve class.

[MrEx3cut0r]

@blokhin after accepting this PR, normalize_query no more needed in mpds-api

[alinzh]

I recommend merging master into this branch after PR #22 is merged, or rewriting the imports to relative ones now so that the PR can be accepted independently.

I would also handle the case where the user submits an empty list; script will break now. It's better to add a check for case: _normalize_query({"elements": []}) .
With this inputs format, '-'.join([]) yields an empty string, which will most likely trigger an API error.

[alinzh]

Change _normalize_query to:

    def _normalize_query(self, query):
        out = dict(query)
        if 'elements' in out and isinstance(out['elements'], (list, tuple)):
            if out['elements']:
                out['elements'] = '-'.join(out['elements'])
            else:
                raise ValueError("'elements' list must not be empty")
        if 'classes' in out and isinstance(out['classes'], (list, tuple)):
            if out['classes']:
                out['classes'] = ','.join(out['classes'])
            else:
                raise ValueError("'classes' list must not be empty")
        return out 

to avoid the empty list problem.

[MrEx3cut0r]

@alinzh is elements and classes lists are required in any request?