Improve parsability of output of ResultWriter

[HomesGH]

Description

Currently, there is a # in the output of the ResultWriter, which makes the header row technically a comment. This is a problem when trying to parse the whole file. It is much easier with simstep instead of # step. This can be parse in python just with:
df = pd.read_csv(path2File, delim_whitespace=True, comment="#", engine="python")

[cniethammer]

I use the following python pandas.read_csv command for result files:
data = pd.read_csv(inputfile, header=2, skipfooter=1)

Would this work for you, too?

I am hesitant to change this long-standing output format ...

[cniethammer]

Instead of modifying the output of the current ResultWriter your question is for a proper CSV file IMHO - feel free to have a look at #369

[HomesGH]

I use the following python pandas.read_csv command for result files: data = pd.read_csv(inputfile, header=2, skipfooter=1)

Would this work for you, too?

I am hesitant to change this long-standing output format ...

I am not sure how this could work for you. I tried it with the Argon example and you have to specify at least delim_whitespace=True otherwise you end up with just one column in your dataframe. But when setting this option, the resulting dataframe is not only wrong, it is also not too obvious since the column names are shifted by one compared to the data. E.g.

    #      step     time    U_pot     U_pot_avg             p     p_avg  beta_trans  beta_rot   c_v   N
0   0  0.000000 -2.09893 -2.09893 -2.669010e-07 -2.669010e-07  3.625780         1.0       0.0  2048 NaN
1   5  0.333758 -2.10937 -2.10316  6.642450e-07  2.621810e-07  1.001330         1.0       0.0  2048 NaN

If you don't have a close look at the data and just use e.g. data["U_pot"] you actually get the data of U_pot_avg.

The only way to parse the present result file correctly is by explicity specifying the column header names.

[HomesGH]

Instead of modifying the output of the current ResultWriter your question is for a proper CSV file IMHO - feel free to have a look at #369

This approach is IMHO the reason why there are so many plugins in ls1 right now. Instead of improving/extending an existing one, it is more convenient to just write a new plugin with a very similar functionality.
The differences between the ResultWriter and your new plugin are just:

• Documentation of the column data (nice!)
• Comma instead of whitespace as delimiter (as broad as long)
• simstep instead of # step (see this PR)