…morpheus#1111)

This PR changes the stub generation to be committed into the repo instead of as part of the build. This will allow CI steps that need the PYI files to not require a GPU as well.

This works by always copying the PYI files back to the source tree so they will show up as diffs. Any diffs during the test phase will be reported as an error.

Closes nv-morpheus#766

Authors:
  - Michael Demoret (https://github.com/mdemoret-nv)
  - David Gardner (https://github.com/dagardner-nv)

Approvers:
  - Christopher Harris (https://github.com/cwharris)
  - David Gardner (https://github.com/dagardner-nv)

URL: nv-morpheus#1111

* Add a `--viz_direction` flag to CLI
* Cherry-pick'd [9532bfe](nv-morpheus@9532bfe) from @mdemoret-nv 

fixes nv-morpheus#643

Authors:
  - David Gardner (https://github.com/dagardner-nv)
  - Michael Demoret (https://github.com/mdemoret-nv)

Approvers:
  - Christopher Harris (https://github.com/cwharris)
  - Michael Demoret (https://github.com/mdemoret-nv)

URL: nv-morpheus#1119

* Publish artifact urls in job summaries fixes nv-morpheus#1105
* Remove redundant apt installs for packages that exist in the conda env fixes nv-morpheus#1106
* Checks now performs a build first. This allows pylint to accurately warn for undefined members. fixes nv-morpheus#956
* Custom pylint checker to validate correct importing of `morpheus._lib` fixes nv-morpheus#727

Authors:
  - David Gardner (https://github.com/dagardner-nv)

Approvers:
  - Michael Demoret (https://github.com/mdemoret-nv)

URL: nv-morpheus#1112

…heus#1088)

The `gitutils.py` script often fails when you are working in a PR, or if there has been a commit to your target branch since you created your current branch.

This improves the utility by using the Github CLI to query the PR information when possible and get the true target branch for the current PR.

If the Github CLI is not available, or the branch isnt in a PR, it falls back to the previous functionality.

Authors:
  - Michael Demoret (https://github.com/mdemoret-nv)

Approvers:
  - David Gardner (https://github.com/dagardner-nv)

URL: nv-morpheus#1088

…morpheus#1113)

* Call the `manual_seed` method from within the subprocess, this ensures the subprocess runs deterministically.
* Add a sleep to a busy-loop in `morpheus/models/dfencoder/multiprocessing.py`
* Misc pylint fixes

fixes nv-morpheus#1021

Authors:
  - David Gardner (https://github.com/dagardner-nv)

Approvers:
  - Christopher Harris (https://github.com/cwharris)
  - Michael Demoret (https://github.com/mdemoret-nv)

URL: nv-morpheus#1113

…hen S3 is not in use (nv-morpheus#1120)

* `DFPFileToDataFrameStage` is no longer an S3 specific stage, but some of the log messages and the `unique-name` still refer to S3.

Fixes nv-morpheus#451

Authors:
  - David Gardner (https://github.com/dagardner-nv)

Approvers:
  - Christopher Harris (https://github.com/cwharris)
  - Michael Demoret (https://github.com/mdemoret-nv)

URL: nv-morpheus#1120

The test reorganization resulted in a couple of problems. First, not all tests were being run due to a mistake I made in handling cmake function arguments. Second, the `test_cuda` test was producing a segfault inside of triton client while initializing curl, but only in debug builds. The fix was to the segfault was not obvious, but apparently only occured when the test utilities were compiled as a separate library. Interestingly, the other tests were not affected. Perhaps there is some global state that needs to be managed that is not managed in the simple `test_cuda` test, which only calls a Matx function (also important, the Matx function result is not inspected, so the test itself is not comprehensive, as Matx does not produce an error if the kernel is not available for the given architecture).

Depends on nv-morpheus#1110

Authors:
  - Christopher Harris (https://github.com/cwharris)

Approvers:
  - Michael Demoret (https://github.com/mdemoret-nv)

URL: nv-morpheus#1118

Update the devcontainer base to 23.10 (this has no relation to cudf versioning).

While there are other improvements, the main reason I'm upgrading is due to a newer version of git, which allows inline public key signing of commits - something we need to be able to commit changes while inside the container without having to manage separate container-specific signing keys.

Example of signing key usage (from host, inherited by devcontainer):
```
ssh-keygen -t ed25519 -C "example@nvidia.com"
git config --global user.signingkey "key::$(cat ~/.ssh/id_ed25519.pub)"
git config --global commit.gpgsign true
git config --global gpg.format ssh
mkdir -p ~/.config/git
git config --global gpg.ssh.allowedSignersFile ~/.config/git/allowed_signers
awk 'BEGIN {FS=" "; OFS=" "} {print $3, $1, $2}' <(ssh-add -L) > ~/.config/git/allowed_signers
```

Authors:
  - Christopher Harris (https://github.com/cwharris)

Approvers:
  - Michael Demoret (https://github.com/mdemoret-nv)

URL: nv-morpheus#1116

…rpheus#1137)

Closes nv-morpheus#1132

Authors:
  - Christopher Harris (https://github.com/cwharris)

Approvers:
  - Michael Demoret (https://github.com/mdemoret-nv)

URL: nv-morpheus#1137