Thanks a lot for your first contribution! Please check out our contributing guidelines and don't hesitate to ask whatever you need.
In the meantime, check out the #python channel on our Powertools for AWS Lambda Discord: Invite link

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarCloud

Hey @JP-Ellis thanks for working in this PR and creating a reproduicle outcome from what we can expect.

Let me split down my feedback to make easy we agree on the next steps.

pdoc3 vs mkdocstrings

We already have an API documentation created by pdoc3, which works well. Personally, I like it; however, I see much more value in using mkdocstrings because it integrates seamlessly with our documentation and offers greater potential to utilize features from MkDocs, which is fantastic. In addition, we already have an issue opened for this migration, which is a great step forward to agree with this kind of work.

Testing this PR

2 - I tested this PR locally and I was able to create the objects.inv file and render the new API tab, which is great to see this working and how it looks like.

Current limitations/problems

3 - Even if we agreeing to migrate from pdoc3 to mkdocstrings, I see some problems in accepting this PR as it is. To move forward, we need to rethink how to do this. In this PR, you are adding 55 new files and marking some to be ignored by markdown lint, which may have some undesirable effects. Also, I admit that our codebase is far from being suitable for mkdocstrings to render this perfectly - or even good. Here are some examples:

We should not render these attributes

The code is completely broken here

We should not have these special characters in the examples

Links without href tag

Making a smooth transition

I'm thinking about making a smooth transition from pdoc3 to mkdocstrings and splitting it by utility, rather than all utilities at once. I suggest this because we'll need to change a lot in our docstrings, comments, and others to have documentation that is actually usable by customers. Otherwise, we'll just introduce a new section that doesn't have much value and may generate doubts about how to use it.

Enabling a new Ruff rule

5 - We can use this opportunity to progressively enable a new Ruff linter rule: https://docs.astral.sh/ruff/rules/#pydocstyle-d. We didn't enable that because our docstrings and comments need to be revamped.

Next steps

So, based on what I explained before, I suggest the next steps are:

1 - Modify this PR to add just one utility and make all the changes we need to make to the mkdocstrings + docstrings configuration. I suggest we start with BatchProcessing as the codebase is not complex and we can do an excellent job on it that will serve as a base for the others.
2 - Enable the Ruff D rule on this PR and see which files Ruff complains about.
3 - Publish the first version with this new change and see how this looks like.

Please let me know what you think and if you would like to make these changes and go through some back-and-forth review process. If you don't have time to do this —we know everyone is very busy— I can handle it and ping you for reviews and feedback. Please let me know your thoughts.

Again, thanks a lot for kicking out this to make our documentation even better.

Amazing feedback!

I should have made it clear at the start that this is meant to be a first proof-of-concept and step towards MkDocstrings. I didn't want to invest too much time in case this wasn't going to be adopted. I'm glad to hear this is something you are keen on!

To answer some of your more specific feedback.

pdoc3 vs mkdocstrings

We already have an API documentation [and] we already have an issue opened for [migrating to MkDocs]

I'll be honest, I completely missed that. But I'm glad to hear this was in the pipeline already!

In this PR, you are adding 55 new files and marking some to be ignored by markdown lint, which may have some undesirable effects.

Completely agree. When I personally use this, I actually create a small Python scripts that is invoked by Mkdocs to automatically generate the Markdown files on-the-fly. There are a couple of pros and cons:

• Pro: The Markdown files are guaranteed to correspond to the Python files
• Pro: No need to commit a large number of mostly-the-same Markdown files
• Con: It may seem a bit 'magical' to someone unfamiliar
• Con: It doesn't allow for customisation of the Markdown file on a per-file basis (I would say that for API docs, this is typically not a big issue)
• Con: It adds yet another dependency, though only for building the docs.

Also, I admit that our codebase is far from being suitable for mkdocstrings to render this perfectly [...] we'll just introduce a new section that doesn't have much value and may generate doubts about how to use it.

Yeah, that's completely fair. A per-utility approach would certainly be possible. Since you already have the pdoc3 docs up, you could also put a disclaimer during the transition to the effect of

We can use this opportunity to progressively enable a new Ruff linter rule: https://docs.astral.sh/ruff/rules/#pydocstyle-d. We didn't enable that because our docstrings and comments need to be revamped.

Absolutely a fantastic thing to add. Just be aware that Ruff doesn't necessarily enforce all docstring rules all the time as it tries to be smart about which lints are applicable to which rules. As a concrete example in repos I maintain, Ruff will detect if there is a mismatch between the args documented and the args in the function; but it will not raise a lint if the docstring has no args at all. This can be problematic if the args are in fact documented but in a style which Ruff did not expect.

With the pace of Ruff's development, this may have changed over the past months, but that is something to be aware of.

1. Modify this PR to add just one utility and make all the changes we need to make to the mkdocstrings + docstrings configuration. I suggest we start with BatchProcessing as the codebase is not complex and we can do an excellent job on it that will serve as a base for the others.
2. Enable the Ruff D rule on this PR and see which files Ruff complains about.
3. Publish the first version with this new change and see how this looks like.

Sounds good! I won't get much time this week to contribute, but hopefully this weekend or next week I should get a bit of time :)

Hey @JP-Ellis! Thanks for the feedback and clarifying things about this PoC. I am writing additional feedback based on your answers.

Completely agree. When I personally use this, I actually create a small Python scripts that is invoked by Mkdocs to automatically generate the Markdown files on-the-fly. There are a couple of pros and cons:

Yes, we should avoid doing this. My observation about the number of files is because it makes it difficult to review the PR and we can lose attention with too many files. That's why I suggested dividing this into small PRs by utility, which will make our work easier.

Yeah, that's completely fair. A per-utility approach would certainly be possible. Since you already have the pdoc3 docs up, you could also put a disclaimer during the transition to the effect of

Perfect, this is a really good idea.

Absolutely a fantastic thing to add. Just be aware that Ruff doesn't necessarily enforce all docstring rules all the time as it tries to be smart about which lints are applicable to which rules. As a concrete example in repos I maintain, Ruff will detect if there is a mismatch between the args documented and the args in the function; but it will not raise a lint if the docstring has no args at all. This can be problematic if the args are in fact documented but in a style which Ruff did not expect.

With the pace of Ruff's development, this may have changed over the past months, but that is something to be aware of.

Hmmm, I didn't know that. I'll spend some time checking this out and see what limitations/caveats we should be aware of and deal with "manually" when doing this work.

1. Modify this PR to add just one utility and make all the changes we need to make to the mkdocstrings + docstrings configuration. I suggest we start with BatchProcessing as the codebase is not complex and we can do an excellent job on it that will serve as a base for the others.
2. Enable the Ruff D rule on this PR and see which files Ruff complains about.
3. Publish the first version with this new change and see how this looks like.

Sounds good! I won't get much time this week to contribute, but hopefully this weekend or next week I should get a bit of time :)

Perfect! Take your time, no rush at all and will be amazing to make this change and improve our documentation.

Thanks for taking the time to do this and let's make it happen 😄