small fix + bump to pyproject.toml ver. for bug fixed release on pypi

[jhnwu3]

This pull request includes a bug fix to the quantile calculation in the conformal prediction set base code, along with version bumps in the project configuration files to reflect the update.

Bug fix:

• Changed the quantile calculation in _query_quantile in pyhealth/calib/predictionset/base_conformal/__init__.py from using np.ceil to np.floor to correct the method for selecting the quantile index, ensuring proper coverage in conformal prediction sets.

Project configuration:

• Updated the project version from 2.0.0 to 2.0.1 in pyproject.toml to reflect the bug fix.
• Updated the pixi-build-python backend version from 2.0.0 to 2.0.1 in pyproject.toml for consistency with the project version.

[siddharthal]

Actually, neither of these are correct. The correct approach is this -- For each test sample, you should add the non-conformity score to the calibration list of non-conformity scores, and query percentile on this set (i.e total set size = N+1; N for calibration samples and 1 test sample). np.ceil makes more sense after you do that.