* Remove pre-commit from github actions

pre-commit.ci will handle running our hooks.

As a result, it's no longer useful to have `tox -e lint` either.
Update `make lint` to run pre-commit directly.

* Bump self-version in pre-commit config

updates:
- [github.com/pre-commit/pre-commit-hooks.git: v3.4.0 → v4.1.0](https://github.com/pre-commit/pre-commit-hooks.git/compare/v3.4.0...v4.1.0)
- https://github.com/python/black → https://github.com/psf/black
- [github.com/psf/black: 21.5b0 → 22.1.0](psf/black@21.5b0...22.1.0)
- https://gitlab.com/pycqa/flake8 → https://github.com/PyCQA/flake8
- [github.com/PyCQA/flake8: 3.9.1 → 4.0.1](PyCQA/flake8@3.9.1...4.0.1)
- https://github.com/timothycrosley/isort → https://github.com/PyCQA/isort
- [github.com/PyCQA/isort: 5.8.0 → 5.10.1](PyCQA/isort@5.8.0...5.10.1)

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

This lets us use `responses` for testing as well.

Request retries are included in the behavior, but done manually rather
than via urllib3 Retry configurations.

closes #34

Co-authored-by: sirosen <sirosen@users.noreply.github.com>

Co-authored-by: sirosen <sirosen@users.noreply.github.com>

Remove `--schemafile` argument from hooks, and switch from
`--failover-builtin-schema` to `--builtin-schema`. In order to
continue to have a vendoring workflow, `vendor-schemas.py` needs a new
source from whence it can pull the schema names and URLs. The mapping
in question now lives in a dedicated module,
`check_jsonschema.catalog`, in a dict. `vendor-schemas.py` imports
this module as its data source.

resolves #38

In order to run these without subprocs, a minor adjustment is needed
to how exits are handled. Inspired by click's "standalone mode", give
the main entrypoint an `exit=False` option.

This is a flag to make the checker print all of the validation errors
encountered, including both the 'best_match' and a walk of all errors.

Move error printing into the utils module, to keep the SchemaChecker
simpler.

Add tests to cover the error printing functions.

This feature is built to allow us to define transforms which are
applied to any instancefile after loading, and before validation.

In the case of azure-pipelines, we have a reference implementation in
the azure-pipelines-language-server to look at. The `transforms`
subpackage contains a reimplementation of that same "parser" in
python. For now, only named transforms are supported, and the only
implemented transform is "azure-pipelines".

Update the hook for pre-commit to use
`--data-transform azure-pipelines`.

Include a test pulled from the pipelines-language-server (with
license) to confirm the result.

If `err.context` is populated, always print the best match error, even
when the `--show-all-validation-errors` flag has not been provided.

Update changelog with notes on anyOf/oneOf error handling -- the cases
which populate `err.context`

Add a new bump-version.py script for updating the version number in
setup.cfg, readme, and changelog.