Add newest RDKit version and remove oldest and tested with pytest

[dessygil]

Checklist:

• [x] Was this PR discussed in a issue? It is recommended to first discuss a new feature into a GitHub issue before opening a PR.
• [x] Add tests to cover the fixed bug(s) or the new introduced feature(s) (if appropriate).
• [x] Update the API documentation is a new function is added or an existing one is deleted.
• [x] Added a news entry.
  • copy news/TEMPLATE.rst to news/my-feature-or-branch.rst) and edit it.

---

[codecov[bot]]

Codecov Report

Merging #189 (869fcfa) into main (56f716d) will decrease coverage by 0.22%. The diff coverage is n/a.

@@            Coverage Diff             @@
##             main     #189      +/-   ##
==========================================
- Coverage   91.66%   91.45%   -0.22%     
==========================================
  Files          46       46              
  Lines        3661     3661              
==========================================
- Hits         3356     3348       -8     
- Misses        305      313       +8     

Flag	Coverage Δ
unittests	91.45% <ø> (-0.22%)	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

see 4 files with indirect coverage changes

:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more

[cwognum]

Out of curiosity: Why do we remove 2022.03? It seems like there's no changes to Datamol that break compatibility with 2022.03, right?

[hadim]

Out of curiosity: Why do we remove 2022.03? It seems like there's no changes to Datamol that break compatibility with 2022.03, right?

Good question. The answer is to not blow up the CI matrix. The current strategy is to test only on the two latest rdkit versions and the 3 last python ones and all the 3 main platforms (win, linux, osx).

For that reason, the doc limits the "compatibility promise" to only the two latest rdkit versions, but it does not mean it won't work. Simply that it's not tested anymore.

We have the below in the doc:

See below the associated versions of Python and RDKit, for which a minor version of Datamol has been tested during its whole lifecycle. It does not mean other combinations does not work but that those are not tested.

[hadim]

Thanks, @dessygil, for the help here!

[maclandrol]

@hadim, the new rdkit canvas based drawing seems nice. I think we should try to eventually refactor some existing viz like radialscope and circlegrid for latent embeddings, as well as think about new ones (e.g progression of a series or scaffold during optimization)

[hadim]

@hadim, the new rdkit canvas based drawing seems nice. I think we should try to eventually refactor some existing viz like radialscope and circlegrid for latent embeddings, as well as think about new ones (e.g progression of a series or scaffold during optimization)

All for it. I always wanted to refactor the viz module to make it more modular and easy to setup complex viz. Maybe an OO approach would be suited here.

[dessygil]

Hey @maclandrol and @hadim, Could I take this issue as well?

[cwognum]

I always wanted to refactor the viz module

Maybe we can open an issue to plan a little? I haven't looked into this in detail, but it seems like a big refactoring. @dessygil could take the lead?

[hadim]

Hey @maclandrol and @hadim, Could I take this issue as well?

That one might be a bit tricky simply because we don't know yet what we want to do exactly. @dessygil feel free to open an issue, so we can brainstorm there. I am all good if you want to start exploring this in a PR, but I can't guarantee it will end up being merged as it's too early at the moment.

[maclandrol]

We will likely need to explore the new functionality first. I haven't hard the time to try that and see the limitation, but looking at the snippet in the rdkit blog, it looks a lot more principled compared to dealing with PIL or composing an SVG. We should brainstorm first IMO.

[dessygil]

For sure, is there anything else that I could take on in the meantime?