Improve documentation of algorithm options

[janosg]

In this PR, we improve the documentation of algorithm options. Primarily, we:

1. Use a modern sphinx version and remove deprecated sphinx extensions from the RTD environment
2. Now only call .. autoclass:: optimagic.optimizers... without adding any description that is not in the code itself (except for a small how-to guide for using the algorithm)
3. We give an example of how optimagic's algorithm classes need to be documented internally such that they are rendered nicely with sphinx
4. We extend the mini "how-to" guide present for each optimizer on how to use this algorithm for minimization. The idea is that it could be very confusing to see that the class is called, e.g., ScipyLBFGSB, but that you are supposed to call it differently. To circumvent any confusion and user-frustration, two mini code-snippets show the user how to call each algorithm.
5. We add a typing page to the documentation which documents all objects in the optimagic.typing module
6. We fix some wrong indentations in the pounders docs

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Files with missing lines	Coverage Δ
src/optimagic/optimizers/_pounders/bntr.py	86.71% <ø> (ø)
...timagic/optimizers/_pounders/pounders_auxiliary.py	97.32% <ø> (ø)
src/optimagic/optimizers/pounders.py	92.56% <ø> (ø)
src/optimagic/optimizers/scipy_optimizers.py	96.90% <100.00%> (+0.06%)	⬆️
src/optimagic/typing.py	88.23% <100.00%> (+1.56%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.