py_chai_lab: New package

[LydDeb]

Add dependencies:
* py_antipickle
* py_gemmi
* py_ihm
* py_modelcif
* py_pandera
* py_tmtools

Update packages:
* py_beartype: add version '0.21.0'
* py_biopython: add versions '1.83', '1.84', '1.85'
* py_typer: add version '0.12.5'
* rdkit: add versions '2024_09_5', '2024_09_6'

Fix the dependence to boost in rdkit:
* 'package boost_numpy is considered to be NOT FOUND' => add variant +numpy
* 'undefined reference to boost::program_options' => add variant +program_option

[RMeli]

I started looking at some of the packages and left some comments. If you don't need multiple versions for a new package, I'd suggest to only add the latest one; it makes things easier. If you add multiple versions, please do check if the dependencies have changed across versions and try not to pin dependencies to a single version if not necessary.

[LydDeb]

Thank you for your review @RMeli

[RMeli]

Another round of review, where I looked at all the packages. I left a comment where I saw that a more recent version is available (apologies if this PR sat here too long...); up to you if you want to update to the latest version or not. If not, just resolve the conversation.

[RMeli]

Suggested change

	depends_on("py-pybind11@2.10.4:", type=("build", "run"))
	depends_on("py-pybind11@2.10.4:2.10", type="build")

I see pybind11~=2.10.4 in the pyproject.toml. And isn't it only a build-time dependency?

[RMeli]

Do you need both versions or worth simplifying the package for the latest one? There is also 0.7.4 now.

[LydDeb]

Indeed, py-chai-lab only needs py-gemmi@0.6.7:

    depends_on("py-gemmi@0.6.3:0.6", type=("build", "run"))