Jenkins integration

[dagardner-nv]

Updates conda scripts to uses sccache (ccache work-alike with an S3 backend)
Adds 4 Jenkins stages:

1. check
2. build
3. test
4. docs build

Results are publised into S3 with a 30 day retention police and are available via:
https://downloads.rapids.ai/ci/morpheus/pull-request/<pull-req id>/<git commit>/<arch>/
ex:
https://downloads.rapids.ai/ci/morpheus/pull-request/80/5161d30673f4d5fb34452df5a8ac6f68b8462685/x86_64/

Intermediate Artifacts pushed to S3:

• conda_env.tar.gz - Morpheus conda environment used by test & documentation stages
• workspace.tar.bz - Morpheus workspace used by test & documentation stages

Results pushed to S3:

• docs.tar.bz - Sphinx HTML documentation
• report_pytest.xml - Pytest junit report
• report_pytest_coverage.xml - Pytest coverage report
• conda_build.tar.bz - Morpheus conda build (not currently working)

Conda builds for cudf are cached seperately (even with sscache this step was still took 17 minutes).
S3 cache location at:
https://downloads.rapids.ai/ci/morpheus/cudf/<cuda ver>/<python ver>/<rapids ver>/<latest git commit to ci/conda>/<arch>cudf_conda.tar.bz
ex:
https://downloads.rapids.ai/ci/morpheus/cudf/11.4/3.8/21.10/b7f728e6db4fc6124970935f461a6791818bfb1d/x86_64/cudf_conda.tar.bz

TODO List:

• Need a test stage
• Need to run style checks
• Need to add sccache
• Need to cache cudf conda packages
• Need to be able to build without using a gpu
• Output junit style xml
• Output docs build
• Add conda build step
• Need to figure out a way to cache lfs

Closes #102
Closes #104

[mdemoret-nv]

Changes looking much better. Approving so we can start running this against PRs.

[mdemoret-nv]

@gpucibot merge

[mdemoret-nv]

@gpucibot merge