Feature/improve speed and limit memory (#11)

[sambenfredj]

Improve speed and limit memory consumption

• stream input files for inference
• add feature: skip deduplication
• add feature: ensemble model
• add feature: rescale input before inference with pre-trained models

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 89.75469% with 71 lines in your changes are missing coverage. Please review.

Project coverage is 83.21%. Comparing base (da2d545) to head (6726dea). Report is 3 commits behind head on main.

Files	Patch %	Lines
mokapot/confidence.py	88.10%	22 Missing :warning:
mokapot/dataset.py	87.15%	14 Missing :warning:
mokapot/mokapot.py	71.87%	9 Missing :warning:
mokapot/model.py	70.37%	8 Missing :warning:
mokapot/parsers/pin.py	90.69%	8 Missing :warning:
mokapot/aggregatePsmsToPeptides.py	89.28%	6 Missing :warning:
mokapot/brew.py	97.32%	3 Missing :warning:
mokapot/utils.py	98.63%	1 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #100 +/- ## ========================================== - Coverage 85.36% 83.21% -2.15% ========================================== Files 19 21 +2 Lines 1640 2032 +392 ========================================== + Hits 1400 1691 +291 - Misses 240 341 +101 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[gessulat]

Thanks @sambenfredj for taking care of the tests! Samia replaced some function working on data in-memory with streaming-based counterparts. Some of the "old" functions are still in the code base, but not used and the tests were removed, while the new functions should be covered. This still leads to the regressions reported by codecov.

We didn't want to remove the "old" functions just yet, before you have a look and are happy with the changes @wfondrie 😄