Use Jinja2 to generate job scripts

[wtbarnes]

Currently, each job script is generated by a combination of old-style (i.e. not f-strings) string interpolation and joining of lists of strings. Using a templating language like Jinja would simplify this approach and offload the string formatting to a separate file.

It looks like the possibility of using jinja2 was discussed in #321 and #133. Is there a consensus on whether this is a good idea? It would certainly clean up some of the code, but maybe at the cost of readability as it would offload a lot of the logic (e.g. whether to include certain flags/options) to a different file.

If there's interest in this approach, I'd be happy to take a crack at it.

[lesteve]

I think the consensus is that there is definitely an interest in using Jinja2 and we would certainly welcome an attempt to do it.

I would advise waiting for #307 (refactoring to use SpecCluster) to be merged before having a go at it. I would expect this is a matter of days before we can merge this in master.

I am going to close this issue in favour of #133. If you can post a comment in #133 to say you are interested to work on this, it would be great.

Side-comment: we are still supporting Python 3.5 so we can not use f-strings.

[wtbarnes]

Thanks for the feedback @lesteve.

I would advise waiting for #307 (refactoring to use SpecCluster) to be merged before having a go at it. I would expect this is a matter of days before we can merge this in master.

For some reason I thought this had already been merged. I'll definitely wait on that. I'm not in a huge hurry!

I am going to close this issue in favour of #133. If you can post a comment in #133 to say you are interested to work on this, it would be great.

Will do

Side-comment: we are still supporting Python 3.5 so we can not use f-strings.

I figured as much.

[lesteve]

I would advise waiting for #307 (refactoring to use SpecCluster) to be merged before having a go at it. I would expect this is a matter of days before we can merge this in master.

FYI #307 has been merged. If you want to consider working on this again, that'd be great.

If you are looking for smaller issues to work on, #332 could be a good one to get started. Other than this #122 would be a great one to get to the bottom of. In particular having a snippet to reproduce the problem would be a great start (see #122 (comment) for what I would try).