Skip to content

Latest commit

 

History

History
216 lines (130 loc) · 7.57 KB

changelog.md

File metadata and controls

216 lines (130 loc) · 7.57 KB

Changes in release "0.8.0"

Contributors

  • @rmcar17 - Python simulate_alignment implementation, documentation, project maintanence and Python processing of additional cli options.
  • @thomaskf - IQ-TREE library interface maintanence and enhancements (including processing additional cli options!).
  • @trongnhanuit & @The-CodeBreakeR - AliSim implementation.
  • @YapengLang - processing UNREST model parameters and storing additional values on returned tree objects.

ENH

  • AliSim can now be used through simulate_alignment.
  • Additional command line options can now be sent to build_tree, fit_tree and model_finder with the other_options parameter.
  • UNREST model parameters are now populated on generated trees.
  • More specific type information is now used across a suite of functions.
  • The used model name is now stored on trees generated by build_tree.

API

  • A seed of 0 is now deterministic across all IQ-TREE functions. None must be used (the default) for non-determinism.

DOC

  • Various enhancements to grammar/spelling.
  • Add sections for simulate_alignment.

Changes in release "0.7.0"

Contributors

  • @rmcar17, fixed a jupyter bug on windows and unwanted behaviour when taxon names contained underscores.
  • @thomaskf and @rmcar17 implemented the ability to fix certain branch lengths when using fit_tree.

ENH

  • Add an option to fix branch lengths when using fit_tree called bl_fixed.

BUG

  • piqtree now works in jupyter notebooks on windows platforms, like it already does for the others.
  • Fix unexpected behaviour when taxon names contain underscores when using fit_tree.

Changes in release "0.6.1"

Contributors

  • @GavinHuttley docs
  • @rmcar17 dependencies

BUG

  • Pin build dependencies for reproducible builds and use new delvewheel get_all_needed API.

DOC

  • Update examples to the latest Cogent3 API.

Changes in release "0.6.0"

Contributors

  • @rmcar17 and @thomaskf enabled windows support for piqtree!

  • @rmcar17 ensured nj_tree raises an appropriate error on NaN input.

  • @rmcar17 reworked random_trees and the related app to become random_tree, returning a single tree.

  • @rmcar17 and @thomaskf integrated IQ-TREE 3 into piqtree.

  • @rmcar17 added support for Lie Markov Model prefixes.

  • @rmcar17 added the STRSYM and UNREST models.

  • @rmcar17 added source tracking for the inputs of apps.

  • @rmcar17 and @thomaskf integrated consensus tree contruction into piqtree.

  • @rmcar17 added support for the parameterisation of currently supported models, base frequencies, and rate heterogeneity.

ENH

  • Added windows support!

  • Add random_tree and piqtree_random_tree. The function and app have the same parameters as the random_trees and piqtree_random_trees versions, except the num_trees argument has been removed.

  • piqtree now uses IQ-TREE 3 for its internal library.

  • Add support for Lie Markov Model prefixes.

  • Add the STRSYM and UNREST models.

  • piqtree apps now track the sources of their inputs.

  • Add a method for building consensus trees with specifiable support parameters.

  • Add a consensus tree app.

  • Substitution models, base frequencies, and models for rateheterogeneity can now be parameterised.

BUG

  • An error is now raised when calling nj_tree with NaN distances instead of failing silently.

API

  • Split Lie Markov Models out of the standard dna enum.

  • Remove unused rand_seed from fit_tree.

  • All apps have been renamed to take the form piq_<func_name>.

  • random_trees and piqtree_random_trees have been removed in favour of random_tree and piq_random_tree. Looping can be used if multiple trees are needed.

Changes in release "0.5.0"

Contributors

  • @GavinHuttley added nj_tree as a hook for cogent3.Alignment.quick_tree.
  • @YapengLang handled negative branch lengths from the rapidNJ tree.
  • @rmcar17, @thomaskf general maintanence on the piqtree/IQ-TREE sides including work on windows behind the scenes.

ENH

  • Add support for Python 3.13, remove support for 3.10
  • IQ-TREE's rapidNJ implementation can be used as a hook for quick_tree on cogent3 alignment objects. Try Alignment.quick_tree(use_hook="piqtree").
  • nj_tree now by default does not allow negative branch lengths. Set allow_negative=True if this behaviour is desired.
  • Allow str to be used for model in build_tree and fit_tree. The value is automatically coerced into the Model class.

API

  • Simplify API for piqtree_phylo and piqtree_fit apps. Both now take a single parameter for the model, matching the parameter for model in build_tree and fit_tree.

Changes in release "0.4.0"

Contributors

  • @rmcar17 integrated new functions from IQ-TREE in Python, added multithreading support, and wrote the API refererence and quickstart.
  • @thomaskf exposed all new functions from IQ-TREE to be available from Python, and worked on multithreading support.
  • @GavinHuttley worked on setting up and writing the documentation and associated devtools, ModelFinder, and integration with cogent3 apps.
  • @YapengLang worked on bootstrapping support and extracting model parameters
  • @KatherineCaley worked on processing the ModelFinder results.

ENH

  • piqtree2 renamed piqtree to support future major releases of IQ-TREE.
  • piqtree now supports multithreading!
  • New function nj_tree constructs a rapid neighbour-joining tree from a pairwise distance matrix.
  • New function model_finder finds the best model for a given alignment.
  • New function jc_distances constructs a pairwise JC distance matrix from an alignment.
  • New function make_model allows converting an IQ-TREE string representation of a model to a Model class.
  • API for random_trees has changed - new order (num_trees, num_taxa, tree_mode, then rand_seed).
  • API for robinson_foulds has changed - now accepts a Sequence of trees.
  • Model parameters are now extracted from IQ-TREE where for now possible.
  • build_tree now supports ultrafast bootstrapping.
  • Model creation is now more robust.
  • Use piqtree.__iqtree_version__ to see what version of piqtree is being used.
  • See what can now be done in our new documentation!

DOC

Changes in release "0.3.1"

ENH

  • Add support for Lie Markov Models.
  • Base frequencies default to None (specified by model).

BUG

  • piqtree2 apps are now pickleable (they can now be run with parallel=True in the cogent3 app infrastructure)

Changes in release "0.3.0"

Contributors

  • @rmcar17 Added new classes to enhance model specification when calling build_tree and fit_tree.
  • @thomaskf fixed a bug in IQ-TREE resulting in segmentation faults on some invalid arguments.

ENH

  • build_tree and fit_tree now allow specifying base frequencies, invariable sites and rate heterogeneity options.

BUG

  • Fixed a segmentation fault on repetitive calls to IQ-TREE with particular arguments.

Changes in release "0.2.0"

Contributors

  • Richard Morris
  • Robert McArthur

ENH

  • build_tree and fit_tree now use enums for the substitution model.

BUG

  • Fixed an issue where calling build_tree or fit_tree twice, then another function with invalid input resulted in a segmentation fault.

DOC

  • Implement scriv as a tool to manage collection of changes, and automated collation into the changelog