Machine-readable dump of the defined sessions

[ppentchev]

How would this feature be useful?

Hi,

First of all, thanks a lot for writing and maintaining Nox!

What do you think about a possible enhancement: add a couple of options to nox --list to make it output the session data (at least the names and the tags, maybe the function names and/or signatures) in a machine-readable format? My main use case would be something like the nox-stages utility that I just wrote - https://gitlab.com/ppentchev/nox-dump or more precisely https://gitlab.com/ppentchev/nox-dump/-/blob/main/src/nox_dump/stages/main.py and its sibling files - so that I can do something like nox-stages @check @tests and have it only run the time-consuming unit tests if the static checkers like mypy and pylint are happy with my changes. Right now, nox-stages obviously uses the nox-dump module that shamelessly sticks its grubby mitts into the internals of nox.registry, and I can imagine how that might not be seen as a viable long-term-supportable API-chasing approach, although I personally would be happy to commit to doing that (including the chasing of API and code internals changes) if you have no strong objections.

Describe the solution you'd like

Still... I wonder if it might not be better to implement this in Nox itself. Would you like me to try to come up with a proof of concept that adds, say, a --toml or --json or -f toml or similar option to nox --list? If so, here are a couple of questions, since you might not want the kind of overengineering that I sometimes do to future-proof the very first version of my projects :)

• what fields would it be a good idea to include in the output? nox-dump currently goes for a lot, but it's fine if you do not really share my view of what is necessary and what is still internal and subject to change
• what do you think about the idea of including a format: version: major: 1, minor: 0 part to handle future added fields (bumping the minor format version) or changed fields (bumping the major one)?
• would a dependency on the toml library be okay for TOML output?
• would some kind of human-readable output, like TSV or CSV or something, be desired, likely with a name like "csv1" to allow for future formatting changes?
• if so, what should it include - function name, description, maybe the tags? maybe the function signatures?
• and how should it be formatted - there are various not-really-compatible CSV-like formats out there, with various ideas about quoting, allowed special characters, etc?

Describe alternatives you've considered

Of course, a "nah, it's too early yet, nothing is stable enough to commit to a stable output format" answer would be just fine, in which case I can maintain nox-dump externally if that's okay.

...and if it turns out that I have missed some Nox command-line option and that it can already do what I want... ooooooof :)) yeah, that would be just like me :))

Anything else?

Thanks again, and keep up the great work!

G'luck, Peter

[ppentchev]

Hi,

I'd like to apologize if my chosen wording came off as condescending or offensive. When I asked if you'd like me to put together a proof of concept, what I meant was, it would be great if something like that were implemented and - if you want - you can offload some of the work on me. Of course I know you're more than capable of implementing things like that yourself if you want to.

So once again, apologies, and thanks for writing Nox!

G'luck, Peter

[ssbarnea]

This issue is currently a blocking for https://github.com/pycontribs/mk/issues/141

[FollowTheProcess]

FYI this is already merged https://github.com/wntrblm/nox/pull/665 and will be in the next release