Stability Update

[Alexander-Sol]

This PR is intended to address longstanding instability in MetaMorpheus results. Specifically, Calibration can produce different outputs with identical inputs.

Unfortunately, even with this PR, variable results from Calibration were observed

Changes

• Fixed locks in DataPointAcquisitionEngine

  • I believe this is the main issue. Previously, improper locking led to race conditions when adding to the lists of LabeledDataPoint

• Changed MsDataFileManager to use ConcurrentDictionary to eliminate warnings about race conditions caused by asynchronous loading

  • (To be clear, I don't think race conditions could have arisen with how FileManagers are currently used, but better safe than sorry)

[codecov]

Codecov Report

Attention: Patch coverage is 93.75000% with 1 line in your changes missing coverage. Please review.

Project coverage is 93.93%. Comparing base (15f8a1c) to head (bfec019).
Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
MetaMorpheus/TaskLayer/MetaMorpheusTask.cs	0.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #2455   +/-   ##
=======================================
  Coverage   93.93%   93.93%           
=======================================
  Files         148      148           
  Lines       22457    22455    -2     
  Branches     3083     3082    -1     
=======================================
- Hits        21094    21093    -1     
  Misses        911      911           
+ Partials      452      451    -1     

Files with missing lines	Coverage Δ
...ineLayer/Calibration/DataPointAcquisitionEngine.cs	94.84% <100.00%> (ø)
...heus/EngineLayer/FdrAnalysis/SpectralMatchGroup.cs	86.66% <100.00%> (+0.95%)	⬆️
MetaMorpheus/TaskLayer/MyFileManager.cs	96.55% <100.00%> (+0.89%)	⬆️
MetaMorpheus/TaskLayer/SearchTask/SearchTask.cs	95.92% <100.00%> (+<0.01%)	⬆️
MetaMorpheus/TaskLayer/MetaMorpheusTask.cs	89.02% <0.00%> (ø)

[nbollis]

Why do we need a new resolve and changes to crosslink?

[Alexander-Sol]

Why do we need a new resolve and changes to crosslink?

CrosslinkSpectralMatch keeps the legacy version of ResolveAllAmbiguities because I don't understand how XL search works and the new method broke things

[nbollis]

Why do we need a new resolve and changes to crosslink?

CrosslinkSpectralMatch keeps the legacy version of ResolveAllAmbiguities because I don't understand how XL search works and the new method broke things

Let me rephrase. Why do we need a new list based resolve method? (Which in turn forces you to do that implementation of cross-link)

[Alexander-Sol]

Why do we need a new resolve and changes to crosslink?

CrosslinkSpectralMatch keeps the legacy version of ResolveAllAmbiguities because I don't understand how XL search works and the new method broke things

Let me rephrase. Why do we need a new list based resolve method? (Which in turn forces you to do that implementation of cross-link)

Calling ResolveAllAmbiguities previously set the MatchedFragmentIons property by getting the first element of the BioPolymersWithSetModsToMatchingFragments dictionary. Dictionaries aren't intrinsically ordered. Creating a resolve method guarantees consistent behavior when setting MatchedFragmentIons