Prevent tests from failing randomly

[andrea-pasquale]

Originally posted by @andrea-pasquale in https://github.com/qiboteam/qibocal/issues/731#issuecomment-1967436824

[alecandido]

In general, it is good to fix the seed for tests, because they should be reproducible. If you want more randomness, fix many seeds.

Moreover, for tests that depends on the input, sometimes you may want to apply fuzzing. Hypothesis works quite well, and it is integrated with Pytest (not only).

https://hypothesis.readthedocs.io/en/latest/