Rnalysis

------------------
This version introduces improvements to the usability and clarity of the graphic interface,
new methods for automatic estimation of the number of clusters in a dataset,
and various bug fixes.

Added
******
* Added three new methods for automatic estimation of the number of clusters in a dataset: Callinski-Harabasz, Davies-Bouldin, and Bayesian Information Criterieon.
* Added a 'Close all Figures' actions to the 'View' menu of the *RNAlysis* graphic interface.
* Added an 'interactive' parameter to Volcano Plots (DESeqFilter.volcano_plot) and 'Scatter Sample Vs Sample' (CountFilter.scatter_sample_vs_sample), allowing user to label data points interactively by clicking on them.
* Added more optional plotting parameters to Volcano Plots (DESeqFilter.volcano_plot) and 'Scatter Sample Vs Sample' (CountFilter.scatter_sample_vs_sample).

Changed
********
* Progress bars are now integrated into the main *RNAlysis* window instead of opening as a dialog box.
* Information about running proccesses and functions is now displayed in the main *RNAlysis* window.
* It is now possible to cancel queued jobs through the *RNAlysis* graphic interface.
* When loading multiple data tables at the same time, it is now possible to change the table type of all data tables at once, instead of one-by-one.

Fixed
******
* RNAlysis KEGG enrichment should now match the new KEGG annotation format from March 1st 2023.
* Fixed bug where importing *RNAlysis* would raise ImportError when cutadapt is not installed.
* Fixed bug where the 'Run' button in the Enrichment Analysis window would grey out whenever the enrichment dataset is changed.
* Fixed bug where the *RNAlysis* stand-alone versions were unable to export Figures in specific formats (e.g. PDF, SVG).
* Fixed bug where functions that depend on R scripts (such as DESeq2 and limma) would sometimes fail to run on MacOS (thanks to Matthias Wilm and `sandyl27 <https://github.com/sandyl27>`_ in `#12 <https://github.com/GuyTeichman/RNAlysis/issues/12>`_).
* Fixed bug where running limma-voom with a design matrix whose column names contained spaces or special characterse would raise an error.
* Fixed bug where the 'highlight' parameter of CountFilter.scatter_sample_vs_sample would not work when used through the graphic interface.
* Fixed bug where enrichment analysis would sometimes fail to run when 'exclude_unannotated_genes' is set to False.
* Fixed bug where translate_gene_ids() would fail for RankedSet objects.
* Fixed bug where filtering gene sets by user-defined attributes (FeatureSet.filter_by_attribute()) would occasionally fail to run.

New Contributors
*****************
* `sandyl27`_ in `12`_

------------------
This version includes bug fixes for a few quality-of-life issues which were introduced in version RNAlysis 3.5.0.

Changed
********
* It is now possible to change the parallel backend of performance-intensive functions such as clustering an enrichment analysis in non-standalone versions of RNAlysis.
* Expanded the Frequently Asked Questions page.
* Added Perl as a dependency for Windows users on the How to Install page.
* Automatic row colours in column-picking tables should no longer mismatch with font colors in a way that obscures visibility.

Fixed
*****
* Fixed bug where occasionally newly-created tabs would open twice instead of once.
* Fixed bug where the 'Add Function' button of the Pipeline window would appear in the wrong location.
* Fixed bug where RNAlysis sessions saved after version 3.5.0 which contain gene sets would raise an error when loaded.
* Fixed typos in the RNAlysis tutorial.

------------------
This version introduces minor bug fixes.

Fixed
*****
* Fixed bug where the *RNAlysis* stand-alone app would sometimes fail to run CutAdapt (thanks to Matthias Wilm).
* Fixed bug where the User Guide action in the graphical interface would point to the Tutorial, and vice versa.
* The X and Y axis labels on volcano plots should now format the 'log' in the label correctly.

------------------
This version introduces new features such as differential expression analysis with the Limma-Voom pipeline,
customizable databases for the quick-search function, basic filtering and procrssing functions for gene sets,
improved progammatic API for FeatureSet and RankedSet objects, and RPKM and TPM read count normalization.
Several changes have been made to improve the user experience, including updated documentation,
improved clarity of function tooltips, clearer output formats, and faster download speeds for tutorial videos.

Added
*******
* Added differential expression analysis with the Limma-Voom pipelines (CountFilter.differential_expression_limma_voom)
* You can now select which databases to display in the right-click quick-search menu, using the settings menu.
* Gene sets now support some basic operations (filtering, gene ID translating, etc) through the graphical interface.
* enrichment.FeatureSet and enrichment.RankedSet now support some filtering operations from the filtering module (such as filtering by user-defined attributes, GO terms, or KEGG pathways).
* Added reads-per-kilobase-million (RPKM) and transcripts-per-million (TPM) normalization methods (CountFilter.normalize_to_rpkm() and CountFilter.normalize_to_tpm()).

Changed
********
* Classes enrichment.FeatureSet and enrichment.RankedSet now inherit from Python base-class set, and can be interacted with like other Python sets. The old API and attributes of these classes were maintained as they were.
* Improved documentation for some functions.
* Function selection tooltips should now display information more clearly.
* Pipelines that contain consecutive clustering/splitting functions will now return their outputs in a clearer format.
* Enrichment bar-plots should now adjust the x-axis limits more tightly to fit the displayed data.
* Improved clarity of automatic titles in enrichment plots.
* Download/update speed of tutorial videos has improved significantly.

Fixed
******
* Fixed bug where Pipelines would not always properly run 'translate_gene_ids'

------------------
This version introduces minor bug fixes.

Fixed
******
* Fixed bug where updating RNAlysis through the graphical interface would not update some of the optional dependencies.
* Fixed typos in the documentation.

------------------
From this release forward, *RNAlysis* is made available as a stand-alone app for Windows and MacOS. You can download these stand-alone versions from the GitHub Releases page.
In addition, new features were added, including new plots, filtering functions, integration of the external tools bowtie2 and featureCounts, and the ability to generate Gene Ontology Graphs and KEGG Pathway Graphs without running enrichment analysis from scratch.

Added
******
* Added a Scree Plot (explained variance per PC) to Principle Component Analysis
* Added CountFilter.split_by_principal_component(), allowing users to filter genes based on their contributions (loadings) to PCA Principal Components.
* Added Filter.drop_columns
* Added support for the Sharpened Cosine distance metric in clustering analyses
* KEGG enrichment can now generate KEGG pathway graphs for pathways that were found to be statistically significant
* Added functions to the enrichment module that can generate KEGG Pathway or Gene Ontology plots based on previously-generated enrichment results
* You can now clear the *RNAlysis* cache from the *RNAlysis* GUI, or through the general module.
* Added bowtie2 alignment to the fastq module.
* Added FeatureCounts feature-counting to the fastq module.
* You can now choose whether or not to discard genes from enrichment analysis if they have no annotations associated with them.
* When right-clicking on a specific cell in a table or line in a gene set view, a context menu will open, allowing you to copy the associated value, or look it up in one of few common biology databases.
* Added sections to the programmatic user guide about the `fastq` module.

Changed
********
* Replaced the 'parallel' parameter in enrichment functions with the 'parallel_backend' parameter, allowing users to choose which parallel backend (if any) will be used in the function.
* Added 'parallel_backend' parameter to all clustering functions under the filtering module.
* When generating Gene Ontology/KEGG Pathway graphs, users can choose whether or not to generate the figure in an additional separate file.
* Updated type annotations of some functions to be more precise and helpful (for example, setting a lower bound on some int function parameters).
* The colorbar ticks in enrichment bar plots now match the axis ticks on the main axis.
* Slight improvements in GUI performance, stability, and looks.
* Slight improvements in performance of enrichment analysis when examining a small number of attributes.
* enrichment.plot_enrichment() was replaced by enrichment.enrichment_bar_plot().
* CountFilter.differential_expression() has new optional parameter `output_folder`, which allows users to save the generated data tables and the R script that generated them into a specified folder.

Fixed
******
* In CountFilter.differential_expression_deseq2(), fixed a bug where design matrix files with non-comma delimiters would cause an error (thanks to `Mintxoklet <https://github.com/Mintxoklet>`_ in `#7 <https://github.com/GuyTeichman/RNAlysis/issues/7>`_)
* Fixed bug where setup.py would install a directory named tests into site-packages folder (thanks to `Bipin Kumar <https://github.com/kbipinkumar>`_ in `#9 <https://github.com/GuyTeichman/RNAlysis/issues/9>`_)
* Fixed bug where the windows of some functions (differential expression, adapter trimming, etc) did not show a link to the function's documentation page.
* Fixed typos in some parts of the *RNAlysis* documentation
* When filtering a table by a single user-defined attribute, the automatic table name will now be more informative about the operation applied.
* Fixed bug where occasionally a Pipeline or Function would generate multiple tables of the same name, but only one of them will appear in the GUI.
* Fixed bug where occasionally significance asterisks on enrichment bar-plots would appear in the wrong location.
* Fixed bug where fastq.create_kallisto_index() (Create Kallisto Index) would not make use of the `make_unique` parameter (thanks to Matthias Wilm)

Removed
********
* Removed the previously-deprecated functions `enrichment.enrich_randomization()` and `enrichment.enrich_hypergeometric()`.



New Contributors
*****************
* `Mintxoklet`_ in `7`_
* `Bipin Kumar`_ in `9`_
* Matthias Wilm