Fix #1512

Changes (Desktop only)

Following the precedent of the Multi-Training tab, a new tab has been added to the Dive Desktop navigation bar, called "Pipeline."

This tab contains an interface similar to that of the multi-training interface, but with the ability to run a specified pipeline over multiple datasets.

Intended Workflow

Step 1. Choose a pipeline type

This is handled by a simple dropdown menu. This must be done first to enable the pipeline select to become enabled.

Limitation: The pipeline types "2-cam" and "3-cam" are filtered out for now to reduce the complexity of selecting datasets.

Step 2. Choose a pipeline

Once a pipeline type is chosen, the "Pipeline" dropdown will allow users to select a specific pipeline of that type. Once this is done, the available datasets appear for selection.

Because the pipeline/type pipeline are chosen first, we have additional control over what datasets are presented for staging. For example, measurement pipelines only allow users to select from stereo datasets.

Step 3. Stage datasets

Much like how datasets are staged/unstaged for multi-training, users can add from the available datasets table, and remove from the staged datasets table as needed.

Step 4. Run pipelines

Clicking the button will create a request for each selected dataset. If all requests come back to the frontend with no issues, the user will be moved to the Jobs tab to monitor the jobs that they just started. If any requests come back with issues, the user will be shown a prompt informing them that some of the pipelines did not start.