Use f-strings

[cmaureir]

After #403 is merged, there will be no reason to use
the old string styling from Python2, and since
f-strings are available since Python 3.6 it could
be adopted as the default style.

[SylvainCorlay]

Thanks @cmaureir. This looks good to me.

We may be stuck with Python 3.5 for a few more months (december).

[cmaureir]

Np @SylvainCorlay I can do the refactoring once 3.5 is out of the picture

[jtpio]

Thanks @cmaureir for taking the time to do this.

I'm also waiting for the day when we can drop everything below 3.5 :)

[maartenbreddels]

Happy to merge this in the future with the f-strings on lines that do not use the logger. However, using log.<topic>(formatstring, *args) does not do any formatting if the log msg does not go anywhere while the f-strings will always will be evaluated, leading to possible performance issues. We should go over this PR and made sure we don't merge those lines.

[jtpio]

#403 has been merged, and Python 3.5 shouldn't be a requirement anymore.

So we should now be able to switch to using f-strings (omitting the changes to the logger) 👍

[cmaureir]

#403 has been merged, and Python 3.5 shouldn't be a requirement anymore.

So we should now be able to switch to using f-strings (omitting the changes to the logger) +1

not sure if I touched something related to the logger while solving all the conflicts, but let me know in case I did it 👯

[cmaureir]

thanks @jtpio, let's see now.

[jtpio]

Looking good :)

[jtpio]

@trungleduc mind providing more context for closing this PR?

Probably because this was handled in a separate PR, for example after switching to black? (not sure we use ruff here yet)

[trungleduc]

Thanks @jtpio for the double-check! I closed the old PRs thinking it was not worth the rebasing effort. Actually, the conflicts in this PR are quite easy to fix, reopening it.

[jtpio]

OK nice thanks for checking @trungleduc 👍

Yeah I remember this PR was created during the Voila Sprint at some previous PyData Berlin conference. Maybe we can help Cristián wrap it up so it can be merged :)

[cmaureir]

oh, super old PR sorry for not reacting before 😆 I tried to solve all the merge conflicts, but I'm certain some styling checks might fail. Let's see what the CI needs to tell us 👯

[trungleduc]

Thanks @cmaureir!

[jtpio]

super old PR sorry for not reacting before 😆

haha thanks @cmaureir for your patience! 🙏

[cmaureir]

I found new ones!
I think after these, we are good to go.

Please notice that "... %r ..." % var cases, were replaced to f"...{var!r}...",
but also there were some cases where %i was used, which doesn't have an equivalent to use in f-strings, so I could use {var:.0f} in order to bypass it, but it adds a float-conversion, but I decided to left them like f"{var}..." only, because the variables was most certainly a string already,.

[maartenbreddels]

However, using log.<topic>(formatstring, *args) does not do any formatting if the log msg does not go anywhere while the f-strings will always will be evaluated, leading to possible performance issues. We should go over this PR and made sure we don't merge those lines.

Note that this still applies. All lines calling the logger should be reverted if we do not want worse performance

[cmaureir]

oh, that's unfortunate, will try to amend the change then

[cmaureir]

I hope everything is in order now @maartenbreddels