Consider adding Bisections as part of the data model

[r-c-n]

I haven't seen bisections mentioned in the code or the documentation so I don't know where they're supposed to fit in the final design.

In the legacy KernelCI, bisections are handled completely separately and away from the main data and we've seen the disadvantages of that, as it's not always trivial to link a bisection to the related regression/s, and it'll be really helpful to have this bi-directional link available at all times to know which regressions a bisection points to and whether a specific regression has been bisected.

[gctucker]

Bisections are going to be part of the post-regression checks work which is in the "advanced features" category. This hasn't been detailed yet as we need to have the base parts of the system implemented first.

The API might not need to know about bisections or regressions, this could all be stored in node.data with different node types. That's the ideal case, but we'll see how feasible it is while making the actual design for it.