qAlgorithms Beta Workflow - now includes componentisation (incorrect results)

This and all preceding releases contain some catastrophic errors which we are still in the process of finding and fixing. Please do not use qAlgorithms unless you are willing to build a more up-to-date version from source yourself or until we publish a fixed release. Results produced with this or the old version will be incorrect!

This release adds a tenuously validated version of the new qPattern algorithm to the qAlgorithms toolchain. Additional features include numerous bugfixes in the rest of the program and a few new command line options.

qAlgorithms Beta Workflow

This release incorporates multiple bugfixes to the previous release as well as a first working prototype of our (as of now unpublished) componentisation strategy. The feature list will now include component numbers for features with a highly similar peak shape.
Issue #25 has been addressed through the stopgap solution of printing a warning instead of exiting the program until we find a complete solution.

This release no longer has any external dependencies, so none of the .dll files linked in the readme are required.

qAlgorithms Beta Workflow

There is still a critical bug in the code, results from this release and earlier are unreliable.

This release fixes the broken -printcentroids option. It should also no longer require the omp dll, but i haven't tested it.
Note: Centroids are always printed in the old version, even if no output directory was specified. In that case, the result file is written in the same directory as the input file.

qAlgorithms Beta Workflow

This release corrects some errors which had a low impact on result correctness and massively improves on performance. Starting at this release, centroiding has been disabled due to low certainty regarding the mass accuracy. Another addition is the inclusion of the -px flag for printing all centroids that are part of a feature, including the regression information. For the time being, all output has been set to verbose.

Remeber to also download the required .dll files linked in our readme.

qAlgorithms Beta Workflow

Addendum 2024.01.06: The currently produced results are wrong in a significant amount of cases. This and all previous releases are affected. A new release will be published once that is no longer the case.

fixed an oversight that made the CLI unuseable, program no longer exits immediately if an error occurs in interactive mode.

qAlgorithms Beta Workflow

Re-added the interactive workflow when double-clicking the .exe (starting it with no arguments).
Behaviour of the command line utility is not affected.

Note: source code is one commit out of date; the correct version is commit 9688080.

Patch to the previous release

Fixed a memory leak during centroiding / peakfinding.

qAgorithms Beta Release

This release only corrects inconsistent user input options and removes the requirement to print a summary when printing the bins.

qAlgorithms Beta Experimental Workflow

Warning: There is a memory leak in qAlgorithms, which will cause the program to crash if too many files are processed at once. Keep your process queue below 40 for now.

This release adds multiple bugfixes and oversights which were discovered since the last, for a full lst refer to the release notes. Notably, a problem where scan numbers caused integer overflow has been fixed.

Major changes to the previos version are that now, the prediction interval is used to describe centroid errors. The code has been error-tested using noise data to prevent crashes when no features can be found. The "-pp / -printpeaks" option has been renamed to "-pf / -printfeatures". "-e / -extended" do not produce different output from "-pf" for the time being. A further change to the output is that now, if a bin encompasses the entire measurement, it has the DQSB -1. This condition is fulfilled if the bin spans from at least the third scan to the third-to-last scan.

For processing ToF data, a hardcoded intensity threshold of 750 is set. This is broadly the minimum before the binning algorithm breaks. We are working on finding a parameter-independent solution in this region, too.

Patch: Correctness check for the first regression coeff being not 0 has been removed. This should be fine.

v0.1.3.beta

This version contains multiple bugfixes which led to incorrect results when returning the regression limits.
Use the -e option to print an extended table containing the (zero-indexed) limits of all peaks respective to their source bin. (Example: A regression has a binIdxStart of 0 and binIdxEnd of 12. This means that all points from the first to the thirteenth in the bin obtained with the -pb option are included in the regression.) The correct bin is given through the binID.

The retention times of these points are already included in the output under lowestRetentionTime and highestRetentionTime.