- Enables the option to draw ALStructure plots from the
plotmode.
- Clarifies that ALStructure does not currently support Best K estimation.
- Moves the population/individual names on the HTML plot back to the bottom. Somewhere aling the way, plotly must have changed the default to above.
- Improves readability and clarifies inline help text for the
plotoption.
- Corrects a bug with
setup.pynot working correctly when installing on Windows platforms.
- Adds a corner case to the STRUCTURE results parser for when the option
#define ANCESTDISTis set to1.
- Improves the shebang on
alstructure_wrapper.Rto improve conda compatibility.
- Forces plotly to version >= 4.1.1 due to a deprecation warning.
- Makes the VCF conversion script handle tri-alellic (and more) SNPs the same way that PLINK does - discard them as NA. I am under the impression that this could be better handled, but for now it will do.
Big thanks to Yamila Paula Cardoso for providing me with a dataset where this bug showed up!
- Fixes a Plotly deprecation warning:
/usr/local/lib/python3.6/dist-packages/plotly/tools.py:465: DeprecationWarning:
plotly.tools.make_subplots is deprecated, please use plotly.subplots.make_subplots instead
- Structure_threader can now supports passing a VCF file for ALStructure. Not even upstream ALStructure can do this. =-)
- Structure_threader can now wrap ALStructure!!! This is a major feature and warranted the release of version 1.3.0.
- Fixed a bug that occurred when multiple arguments were passed to
extra_optswhen wrapping fastStructure. Thank you to @oviscanadensis for the superb bug report.
- Fixed a bug when parsing yet another variant of STRUCTURE output.
- Fixed a bug that caused a crash when MavericK alpha and alphaPropSD parameter values were split with both a comma and a white-space. Thank you to Sophie Gresham for reporting it.
- Adds infrastructure to use Gitlab's CI server parallel to Travis CI.
- The field tests coding was improved to drop some hard-coded paths into dynamic ones.
- Marks
numpy>=1.12.1dependency. The key issue here being in the version number.
- Corrects a bug with plotting MavericK "Qmatrix" files that did not have a "population identifier" field.
- Using the
plotfunction on MavericK results now works as intended. - Improved and clarified how the
plotinput files should be indicated.
- Structure_threader now always makes a sanity check for
mainparamsandextraparams.
- New option introduced:
--seed, which allows the user to define a random seed value. It is not mandatory, and defaults to "1235813". - Structure_threader now checks if the
RANDOMIZEoption is set in theextraparamsfile. If it is, it gets disabled since random seeds are now handled internally.
- Corrected a bug when sorting plot order using an "indfile" with more than 10 populations.
- Improved unit and field tests.
- Removed a leftover debugging line.
- Allows a random seed to be passed to STRUCTURE. This seed will be used to generate N seeds, where N is the number of runs. Each will be attached to the appropriate CLI.
- Added an extra sanity check for write permissions to the output directory.
- The "skeletons" module was not being packaged in Pypi. This is now fixed.
- Added more underflow protection.
- Added an underflow protection.
- Added a fail-safe in case of MavericK using a single "MainRepeats" parameter.
- Improved some text strings.
- Corrected a few typos in some of the printed messages.
- Fixed a bug with a nonsensical leading "/" on the windows version.
- Corrected a bug where MavericK execution stopped due to multiple
alphaandalphaPropSDvalues being set. - K plots visualization is now improved when plotting a large number of Ks.
- Added option to generate skeleton parameter files (
mainparamsandextraparams) for STRUCTURE. - If wrapping MavericK with TI option turned off, bestK tests are now skipped.
- Vastly improved unit test suite.
- New presentation of "bestK" evidence when wrapping MavericK.
- Fixed a broken link in the docs (Thanks to @briantrice for finding it).
- Fixed a bug with the file prefixes for drawing plots that required the user to be in the directory where the meanQ files were located.
- Corrected a bug with the way "fastStructure" was spelled
- Corrected a bug with population delimiters.
- Sanity checks errors and warnings are now also colored.
- Fixed issue when best K is 1
- Fixed missing population separators on static
.svgplots when using --use-ind-labels - Changed a variable name to avoid conflicts, despite being in a different namespace.
- Fixed missing population separators on static .svg plots when using --use-ind-labels
- Fixed key error when providing faststructure format in plot mode
- Allow the program to draw the
.svgplots only on gray scale (with '-bw' option) - Allow the program to show individual label names even when population information is provided (with '--show-ind-labels' option)
- Fixed Qvalue sorting when providing an --ind file with two columns