feat(Pyclient): use column metadata when processing data from CSV API

[YpeZ]

Issue

Currently the Pyclient implicitly converts CSV data from the get method to a pandas DataFrame. In this conversion pandas makes assumptions about the contents of each column. This can lead to unwanted results, such as the casting of integer values to string values, or string values to floats. The behaviour of dealing with NA values is often unpredictable too.

Solution

The Pyclient already contains functionality for working with column metadata. This metadata can then be used to ensure the conversion occurs in the way it is expected to.
A newly written parsing function must be implemented within the get method of the Pyclient.

Alternatives

No response

Additional context

When processing data from the National Node staging areas into BBMRI-ERIC tables, this unexpected behaviour was encountered (f.e. latitude values were seen as floats and zero's were removed from the values). To circumvent this, the pyclient's get method has been copied into the BBMRI-ERIC publish package and adjusted in such a way that no pandas DataFrame is included in the process. A second function has been added that resets the datatypes: reset_data_types