allow to append feature files

[audiofeature]

if audio features are extracted in multiple stages (e.g. from different folders, one after another) allow to append to existing feature files (instead of overwriting them). (take care of relative path issues though)

[audiofeature]

as an intermediate solution, we now added a function to read multiple feature files into one array, in case they were extracted in different steps:

from rp_feature_io import read_multiple_feature_files

[audiofeature]

this is working now already for CSV files (not HDF5)

[audiofeature]

works for HDF5 too now