maint: Add additional linter-related pre-commit hooks

[Skylion007]

Description

• This PR adds a few pre-commit hooks that will ban a few bad practices in the repo. Mainly ban blanket linter ignores (all ignores need to be error specific and some common RST bugs.
• I also fixed all the found rst bugs in the repos (exclusively single backticks that should have been double for code).
• This PR also adds a pre-commit hook that removes all unnecessary NOQAs automatically while keeping the flake8 plugins in sync flake8 hook.
• I pushed upstream changes to the mypy hook that allows us to simplify our mypy hook config.

Suggested changelog entry:

* Add pre-commit hooks to ban blanket linter ignores 
* Add pre-commit hooks to detect common RST mistakes and fix all found instances of those.
* Add the yesqa pre-commit hook to automatically remove unused noqa comments

[henryiii]

What are the failures if you add the triple tick check?

[Skylion007]

@henryiii failures from the rst-backticks check

docs/advanced/classes.rst:1269:wish to directly manipulate the `PyHeapTypeObject` corresponding to a
docs/advanced/embedding.rst:43:all the functions and classes in pybind11. The RAII guard class `scoped_interpreter`
docs/advanced/embedding.rst:50:There are a few different ways to run Python code. One option is to use `eval`,
docs/advanced/embedding.rst:51:`exec` or `eval_file`, as explained in :ref:`eval`. Here is a quick example in
docs/advanced/embedding.rst:111:Python modules can be imported using `module_::import()`:
docs/advanced/embedding.rst:137:Modules can be reloaded using `module_::reload()` if the source is modified e.g.
docs/advanced/embedding.rst:147:Embedded binary modules can be added using the `PYBIND11_EMBEDDED_MODULE` macro.
docs/advanced/embedding.rst:173:`PYBIND11_EMBEDDED_MODULE` definitions (as long as they have unique names).
docs/advanced/embedding.rst:219:The Python interpreter shuts down when `scoped_interpreter` is destroyed. After
docs/advanced/embedding.rst:221:`initialize_interpreter` / `finalize_interpreter` pair of functions can be used
docs/advanced/embedding.rst:245:Creating multiple copies of `scoped_interpreter` is not possible because it
docs/advanced/exceptions.rst:99:parameter, a `handle`:
docs/advanced/exceptions.rst:106:Then `PyExp` can be caught both as `PyExp` and `RuntimeError`.
docs/advanced/exceptions.rst:110:The default base class is `PyExc_Exception`.
docs/advanced/functions.rst:235:scope guard will work. This is very useful in combination with `gil_scoped_release`.
docs/advanced/misc.rst:87:The ``call_go`` wrapper can also be simplified using the `call_guard` policy
docs/advanced/pycpp/utilities.rst:31:redirection. Replacing a library's printing with `py::print <print>` may not
docs/advanced/pycpp/utilities.rst:65:extra type, `py::scoped_estream_redirect <scoped_estream_redirect>`, is identical
docs/advanced/pycpp/utilities.rst:67:`py::call_guard`, which allows multiple items, but uses the default constructor:
docs/advanced/pycpp/utilities.rst:77:manager, using the `py::add_ostream_redirect() <add_ostream_redirect>` function:
docs/advanced/pycpp/utilities.rst:106:pybind11 provides the `eval`, `exec` and `eval_file` functions to evaluate
docs/changelog.rst:253:* ``py::str`` changed to exclusively hold `PyUnicodeObject`. Previously
docs/changelog.rst:254:  ``py::str`` could also hold `bytes`, which is probably surprising, was
docs/changelog.rst:1407:* Fixed implicit conversion of `py::enum_` to integer types on Python 2.7.

[henryiii]

Pretty sure those are all mistakes, using markdown syntax. If you really intended single-ticks, you should add a role, like

:py:`stuff`

[Skylion007]

@henryiii To clarify, all those errors should be rewritten as by being prepended with :py: ?

[henryiii]

No, they should use double tick marks. I think single tick marks use the default role, which is :py:, but I don't think any of those were trying to use it. If it's a fully qualified local name, then using that role is fine. But don't think any of those are, they are just code.