Automatic retrieval based on input

[kjappelbaum]

Needs discussion.

For some, especially smaller, files one could enhance the user experience by automatic retrieval and automatic parsing. I especially think of the the MSD and radial distribution functions.

For the RDF i currently use something like

        # ToDo: this is not really elegant, optimize it
        if 'ComputeRDF' in self.ctx.parameters['GeneralSettings'].keys():
            if self.ctx.parameters['GeneralSettings']['ComputeRDF'] in [
                    'yes', 'YES', 'Yes'
            ]:
                self.ctx.settings = ParameterData(
                    dict={
                        'additional_retrieve_list':
                        ['RadialDistributionFunctions/System_0/*'],
                    })

which of course would be needed to be generalized to multiple systems. It can get large if the user specifies a lot of different atom types.

In my case, I also parse the RDF. One could save space by saving r only once as it is the same for all files.

def parse_rdfs(rdf_file_list):
    rdf_dict = {}
    for file in rdf_file_list:
        name = Path(file).stem.strip('RDF_')
        df = pd.read_csv(file, comment='#', header=None, sep='\s+')
        r = df.iloc[:, 0].values
        rdf = df.iloc[:, 1].values
        rdf_dict[name] = {'r': r, 'rdf': rdf}
    return rdf_dict

Linking @danieleongari for notice.