Fix for comment lines inside blocks

[pointedsphere]

Hi Jacob,

I'm Scott, one of Matt's PhD students that was looking at incorporating castep_outputs into some example scripts.

Found a small bug with the reading of a cell file where there are comments within a block that contain numbers. A bit of a niche thing, but the CASTEP output of final post geom-opt cells (with write_cell_structure) includes comments of this form.

This PR is to account for this possibility. Unfortunately cannot use re.split('[#!]',line)[0] to get the bit of the line without a comment as the end of the lines Kinetic eigenvalue # 1 = 0.19929544E+02 in CASTEP output files are also thrown away by this. As such, this just discards full comment lines.

Thanks,

Scott

[oerc0122]

This seems like a far-reaching and potentially harmful change to automatically ignore any/all #/! characters in any file parsed by castep_outputs.

I would suggest instead a separate strip_comments filter in the utilities.

comment_re = re.compile(r"^\s*#")
itertools.filterfalse(comment_re.match, my_data)

[pointedsphere]

I agree it is a bit drastic.

I'm not 100% on exactly how you were suggesting to add the filter. I've modified the PR to use itertools.filterfalse in the block reading to ignore full line comments. However, if I've missed the mark very happy to close the commit and leave it as an issue in the tracker.

[oerc0122]

Like I said, I don't think this should be applied generally in FileWrapper, but as a utility function when parsing chunks of data where comments should be stripped.

[oerc0122]

I agree it is a bit drastic.

I'm not 100% on exactly how you were suggesting to add the filter. I've modified the PR to use itertools.filterfalse in the block reading to ignore full line comments. However, if I've missed the mark very happy to close the commit and leave it as an issue in the tracker.

@pointedsphere c.f. #173 and check that meets your needs.