using context managers to write files

[shimwell]

This PR changes the write statements to use context managers. This removes the need for a .close() when the file writing is done.

Context managers do however require another indentation which means the lines are shorter for the code below. The formatter has therefore decided that there is a more optimal layout for these lines, now that they have less space to stay below the line limit. So we have a bigger diff than might be expected when starting the PR

Requested in issue #111

closes #111

@alberto743 are you ale to review

tests passing locally

[alberto743]

@shimwell Yes, I can review by next Monday 6th.

[alberto743]

Hi @shimwell . Thank you for the development, which is generally fine for me.
Here are a few comments related to a problem already present in the lines you have changed.

1. the self.inpfile should be transformed to a local variable, which needs to be passed cross function calls, because logically the other private methods of the MONTECARLOCODENAMEInput class should not depend upon an attribute not initialized by the __init__ method. Also, trailing underscores in methods should be avoided. The issue was indeed already present in the previous code.
2. testing/test.py should be changed at lines 55-57 as well
3. geouned/GEOUNED/Utils/BasicFunctions_part2.py has the same problem at line 18, but maybe in this case it would be better to open a separate logging instance instead of writing to a file
4. geouned/GEOUNED/__init__.py presents the issue in function print_warning_solids, but maybe this one as well may be transformed to a logging instance
5. geouned/GEPOReverse/Modules/Parser/parser.py still has dangling open calls between lines 1214-1253, but maybe this section related to Python 2.7 shall be simply stripped away since it is dead code because of line 13

[shimwell]

Thanks alberto

I shall get this branch merge compatible with dev and get started on your recommendations. Much appreciated

[shimwell]

@alberto743 I've done 4 out of the 5 suggestions. I'm just holding back on the first suggestion regarding self.inpfile which was present prior to this PR. The reasons I'm seeing a few methods such as write_xml and write_py both write files it gets a bit messy. I'm thinking it might be best to have each of these smaller functions in the OpenMCFormat return a string or list of stings and then combine the strings and write it all in a single write function instead of passing around an open file handle and continually writing to it. I would prefer to do this in another PR if that is ok.

[shimwell]

Hi @shimwell . Thank you for the development, which is generally fine for me. Here are a few comments related to a problem already present in the lines you have changed.

1. the self.inpfile should be transformed to a local variable, which needs to be passed cross function calls, because logically the other private methods of the MONTECARLOCODENAMEInput class should not depend upon an attribute not initialized by the __init__ method. Also, trailing underscores in methods should be avoided. The issue was indeed already present in the previous code.

Not done yet but would like to fix this in a follow up PR. self.inpfile was also not introduced by this PR so perhaps a bit of topic to fix it now.

2. testing/test.py should be changed at lines 55-57 as well

Done

3. geouned/GEOUNED/Utils/BasicFunctions_part2.py has the same problem at line 18, but maybe in this case it would be better to open a separate logging instance instead of writing to a file

Done as a logging instance

4. geouned/GEOUNED/__init__.py presents the issue in function print_warning_solids, but maybe this one as well may be transformed to a logging instance

Done as a logging instance

5. geouned/GEPOReverse/Modules/Parser/parser.py still has dangling open calls between lines 1214-1253, but maybe this section related to Python 2.7 shall be simply stripped away since it is dead code because of line 13

Done by removing old python 2.7 code

[alberto743]

@alberto743 I've done 4 out of the 5 suggestions. I'm just holding back on the first suggestion regarding self.inpfile which was present prior to this PR. The reasons I'm seeing a few methods such as write_xml and write_py both write files it gets a bit messy. I'm thinking it might be best to have each of these smaller functions in the OpenMCFormat return a string or list of stings and then combine the strings and write it all in a single write function instead of passing around an open file handle and continually writing to it. I would prefer to do this in another PR if that is ok.

Fully agree!

[shimwell]

closing and replacing with PR #160 as I accidently made some commits to this branch and messed it up