All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog. This project adheres to Semantic Versioning.
- PhenotypeManager: prevented Pearson calculation on constant or NaN-containing arrays
- GeneralParser: fixed error-handling on malformed input files.
- MS2DeepScoreNetworker: fixed MS2 spectra filtering for ms2deepscore algorithm
- Fixed bug in SummaryWriter: a nonexisting function was referenced, leading to premature exit of module.
- Versioning
- [Breaking change] Removed MS2Query de novo annotation after observation of process instability (unforeseen process termination by system with SIGKILL (9))
- Added exception for fail of Sample Scores calculation in case of lack of spectral similarity networking data (e.g. no MS/MS data provided)
- Corrected erroneous assignment of 'True' to 'module_passed' for FragmentAnnotator and NeutralLossAnnotator if no MS/MS information was provided
- Loosened typing restrictions for Feature and Sample object attributes: area and height (intensity) now accept float values.
- [Breaking change] Removed toggle 'nonbiological' from 'FragmentAnnotator' and from parameters file; 'nonbiological' fragment annotation is now performed automatically
- Instead of full file paths, only filenames are now written to the 'out.fermo.session.json' file
- [Breaking change] Parameter settings for
additional_modules/feature_filteringwere changed from a list of ranges to a dictionary with explicit values. - Reworked score assignment for qualitative phenotype data: phenotype-associated features now always receive a score of 1.0, and non-associated ones a score of 0.0.
- For all modules with runtime restriction, the 'maximum_runtime' parameter was set to a default of '0' (unlimited runtime). Therefore, runtime restriction must now be specified explicitly.
- Added a 'module_passed' parameter to all modules. This allows a more accurate description via the SummaryWriter (e.g. module was activated but timed out, and lack of e.g. annotation is due to premature ending and not because there were no hits).
- Removed 'phenotypes' as separate Feature attribute: write to Annotation object instead.
- Implemented sorting of annotation entries in descending order
- Fixed bug in "SummaryWriter": implemented error catching.
- Replaced global logger with logger specific for
main_cli().main()now needs an argumentlogger - Reworked output file naming: all output files now start with
out.fermo.and a suffix specifying their type - Removed output directory selection: the output directory is now always
resultslocated in the directory in which the peaktable resides. - Features now always have default result values (previously, some Features could have an empty dictionary)
- MS2Query assignment now uses temporary directories for data reading/writing
- Removed
sys.exit(0)in case of successful run to fix compatibility issue with Celery task manager (fermo_gui)
- Implemented logging verboseness parameter for CLI interface
- First public release of
fermo_core