Handle assembly edge case

Documented in examples/example_insertion_vs_circle.py

In an insertion assembly, when the two homology regions overlap, the same assembly could be interpreted as a circular one.

In the picture below, we could have:

1. An insertion where the bottom sequence gets inserted into the top
2. A circularisation of the bottom one, with the top acting as a "bridge"

This bridge thing could be not supported, but I suppose it can actually be quite useful for something like a gibson assembly.

Supporting this two cases makes it necessary to have the parameter is_insertion in assemble. Passing the same fragment twice for the insertion assembly would solve the problem and turned into a pydantic object, since the fragment representation would start with None and end with None locations, if the ids are the same it could be inferred that it's an insertion.

Start with examples/example_insertion_vs_circle.py and the assemble entry point described in the issue. Compare the insertion and circularisation cases when homology regions overlap, then determine how is_insertion, repeated fragments, and the proposed pydantic representation should distinguish them; done means both interpretations are supported or the bridge case is explicitly rejected.