Add Elasticsearch-DSL code

[miguelgrinberg]

This change adds the Elasticsearch-DSL code. Note that the documentation will be merged in a separate PR.

• The imported source tree has the same structure as in the elasticsearch_dsl package, except that its root is now elasticsearch.dsl. Users can replace the _ with a . in all imports.
• I had to make minor changes to some of the DSL's test fixtures, because this client uses function scope for the client fixtures and DSL used session scope. I have switched all the DSL fixtures to function scope.
• I had to make the DSL client fixtures work with https:// Elasticsearch endpoints, since that is what we are using in integration test runs.
• The unasync of the DSL classes remains in a separate run-unasync-dsl.py script. We can consider merging the two.
• The generator.py script that generates the query, aggregation and response source files was moved to this project as generator-dsl.py, along with its Jinja templates. We can consider moving this to the Python client generator project.

[github-actions]

A documentation preview will be available soon.

• 🔨 Buildkite builds
• 📚 HTML diff
• 📙 Preview page

Request a new doc build by commenting

• Rebuild this PR: run docs-build
• Rebuild this PR and all Elastic docs: run docs-build rebuild

_{run docs-build is much faster than run docs-build rebuild. A rebuild should only be needed in rare situations.}

_{If your PR continues to fail for an unknown reason, the doc build pipeline may be broken. Elastic employees can check the pipeline status here.}

[pquentin]

For what it's worth, this one makes me a bit sad, as it still pulls six to this day: https://inspector.pypi.io/project/python-dateutil/2.9.0.post0/packages/ec/57/56b9bcc3c9c6a792fcbaf139543cee77261f3651ca9da0c93f5c1221264b/python_dateutil-2.9.0.post0-py2.py3-none-any.whl/python_dateutil-2.9.0.post0.dist-info/METADATA

[miguelgrinberg]

Let me see if we can drop it.

[miguelgrinberg]

@pquentin The builds are currently not generating code coverage data, and I haven't been able to figure out why. And I get coverage data when I run the tests locally. I'll continue looking for this, but if you have any ideas why this is happening let me know. Other than this the PR is ready for a review.