maintenance release 0.2

This release captures the development / fixes until the release date.
A new development cycle might introduce backwards-incompatible changes.
Depracation will only be considered in the following release (0.3).

What's Changed

• fix warnings + make it a proper test by @agramfort in #1
• Create unit_tests.yml by @agramfort in #2
• remove specific sklearn dependence by @dengemann in #12
• add support for epochs by @apmellot in #3
• ENH/API: first wave of renaming by @dengemann in #13
• Small changes by @DavidSabbagh in #16
• [ENH]: start refactoring pipelines module by @dengemann in #14
• Features selection and renaming by @apmellot in #18
• FIX/ENH: improve API tests / misc / better log computation by @dengemann in #19
• FIX: make filter bank classifier available at top level by @dengemann in #22
• fix for one sample + test by @apmellot in #25
• Changing frequency bands default to avoid including alpha band if user provides a value by @hubertjb in #33
• Changing default Riemann projection scale to auto by @hubertjb in #32
• FIX: fix for latest MNE version that does not allow overwriting info['sfreq'] by @dengemann in #34
• fix PSD computation with new Spectrum class by @agramfort in #39
• Add support of numpy inputs to covariance_transformers by @apmellot in #37

New Contributors

• @agramfort made their first contribution in #1
• @dengemann made their first contribution in #12
• @apmellot made their first contribution in #3
• @DavidSabbagh made their first contribution in #16
• @hubertjb made their first contribution in #33

Full Changelog: https://github.com/coffeine-labs/coffeine/commits/0.2