Add a check if files exist to klaas_split_data

[jebuss]

if one of the files you create with klaas_split_data, you might get the following error:

klaas_split_data build/data/gamma_precuts.hdf5 build/data/gamma \
        -f 0.25 -n train_sep \
        -f 0.5 -n train_reg \
        -f 0.25 -n test \
        -i events
INFO:root:Found a total of 692836 single-telescope events in the file
INFO:root:Writing 173209 telescope-array events to: build/data/gamma_train_sep.hdf5
Traceback (most recent call last):
  File "/home/jbuss/.conda/envs/gamma/bin/klaas_split_data", line 11, in <module>
    load_entry_point('klaas', 'console_scripts', 'klaas_split_data')()
  File "/home/jbuss/.conda/envs/gamma/lib/python3.5/site-packages/click/core.py", line 716, in __call__
    return self.main(*args, **kwargs)
  File "/home/jbuss/.conda/envs/gamma/lib/python3.5/site-packages/click/core.py", line 696, in main
    rv = self.invoke(ctx)
  File "/home/jbuss/.conda/envs/gamma/lib/python3.5/site-packages/click/core.py", line 889, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/home/jbuss/.conda/envs/gamma/lib/python3.5/site-packages/click/core.py", line 534, in invoke
    return callback(*args, **kwargs)
  File "/home/jbuss/software/classifier-tools/klaas/scripts/split_data.py", line 111, in main
    write_data(selected_data, path, key=key, use_hp5y=not use_pandas)
  File "/home/jbuss/software/pyfact/fact/io.py", line 34, in write_data
    to_h5py(file_path, df, key=key, **kwargs)
  File "/home/jbuss/software/pyfact/fact/io.py", line 267, in to_h5py
    append_to_h5py(f, array, key=key)
  File "/home/jbuss/software/pyfact/fact/io.py", line 428, in append_to_h5py
    raise KeyError('No such dataset {}'.format(dataset))
KeyError: 'No such dataset None'
make: *** [build/data/gamma_train_reg.hdf5] Fehler 1

I suggest, that klaas_split_data checks if files excist before and gives a warning

[jebuss]

maybe even a prompt if files should be overwritten