Allow integrator to take repre files with a single file as string instead of list

[moonyuet]

Changelog Description

This PR is to allow integrator to take repre files with a single file as string instead of list
Fix #1836

Additional info

Marked it as draft for the discussion.

Testing notes:

1. Create Render
2. Publish both local/farm rendering

[iLLiCiTiT]

I do remember we were discussing this and the result was that we should not allow this and we should fix the source of the issue instead. Or am I wrong? Do you remember @BigRoy ?

[BigRoy]

I do remember we were discussing this and the result was that we should not allow this and we should fix the source of the issue instead. Or am I wrong? Do you remember @BigRoy ?

You are not wrong - but it was for the initial hotfix for a particular case which was fixed here: #1836

However, long-term I believe we do want to 'simplify' things if we can. The question was more, that if we provide a single file in a list, whether we should still assume it to be a sequence (or 1 frame?) by the fact of it being a list or should we enforce it to be a frame-less file (as this PR would make it do; but that's also what the list[str] -> str on representation.files would already be doing.) So it wouldn't technically differ currently.

Anyway, we held off from allowing this fix initially due to the worry of edge cases it may somehow affect. But if we can reasonable avoid having to make it a str and perhaps even rely on it always being a list[str] it'd make the variable repre["files"] during publishing much more consistent.

This 'issue' though describes that single frame in list[str] should be a single frame sequence - which does mean that the fix done here is incorrect.

Anyway, main question is:

• is there still reason for having files: str = "singlefile" and `files: list[str] = ["single_frame.exr"]"?
• can we actually safely detect always the frame number from a single file in the list e.g. sh001.exr might refer to the shot number, not the frame number?

[iLLiCiTiT]

I agree with the questions, but as you said this PR is not handling that and the question is out of scope of this PR (and the source issue) -> this is not the right place to ask the questions.

[antirotor]

But this will break single frame sequences, right? I think we should communicate it somehow if that is the solution (even temporary one)

[jakubjezek001]

I am not sure if this is related but there might be also case where single frames with slates will need to be published. It might be really rare use case. And it will in fact become two frame sequence.