Improve coverage of dodiscover to 95% in time for v0.1 release

[adam2392]

https://app.codecov.io/gh/py-why/dodiscover/tree/main/dodiscover shows the coverage by our unit-testing. A robust software package should at least aim for >90% test coverage.

This is feasible for us given that many of the API and algorithms have expected usage. The main components that we DEFINITELY should improve upon are:

1. Context/ContextBuilder: If Context is a central ingredient for our API, then we should have full test coverage of the expected errors and edge cases
2. CI tests
3. Constraint-based algorithms

The first two should be relatively straightforward upon inspection of the codecov files. However, adding additional tests to the causal discovery algos may require some more thinking since these need to test different parametrizations of the algorithms.