Rnalysis

------------------
This patch contains bug fixes and improved functionality for enrichment lollipop plots,
as well as bug fixes for issues with the stand-alone version.

Changed
********
* Single-set enrichment result tables now contain observed/expected values based on the XL-mHG test cutoff.
* When loading a differential expression design matrix, RNAlysis now issues an error if the design matrix column names contain invalid characters.
* Updated the scaling of enrichment lollipop plots to make small 'observed' values easier to discern.

Fixed
*******
* Fixed bug where an error message would sometimes appear after RNAlysis finishes generating an automatic session report on the stand-alone app.
* Fixed bug where enrichment lollipop plots in horizontal mode would display the observed/expected values in reverse order.
* Fixed bug where enrichment lollipop plots and the 'show_exp' parameter would not work on single-set enrichment data.
* Fixed bug where sometimes the tab label of clustering/differential expression output tables would not match the name of the generated table.

------------------

Version 3.9.1 of RNAlysis introduces several improvements and fixes to further improve your analysis experience.
The release includes new optional parameters for single-set enrichment functions, compatibility improvements with newer Python versions,
improved error messaging for R scripts, and adresses minor issues related to enrichment analysis, documentation, plotting parameters, and Pipeline saving.

Added
*******
* Added new optional parameters to single-set enrichment functions, allowing users to determine the top and bottom cutoffs for the XL-mHG test ("X" and "L").

Changed
********
* RNAlysis single-set enrichment analysis using the XL-mHG test now supports Python versions >= 3.8.
* RNAlysis stand-alone app is now built on Python 3.11, improving overall performance.
* Error messages caused by running R tools such as DESeq2, Limma-Voom, and FeatureCounts will now clearly state the reason the script failed, making it easier to understand what went wrong.

Fixed
*******
* Fixed bug where enrichment analysis would raise an error when running enrichment analysis on a gene set with no relevant annotations, or a gene set that does not intersect at all with the background gene set.
* Added missing documentation for plotting parameters in some enrichment functions.
* Depracation Warning should no longer appear when generating a box-plot or enhanced box-plot with scatter=True (CountFilter.box_plot, CountFilter.enhanced_box_plot)
* Fixed bug in featureCounts single-end mode where the 'output_folder' parameter could appear as disabled.
* Fixed bug where RNAlysis would raise an error message after saving a Pipeline, even when the Pipeline was saved successfully.

------------------
Version 3.9.0 of *RNAlysis* introduces several enhancements and fixes to improve your experience.
The release includes additional enrichment plot styles, a new option for PCA plots,
the ability to load and save data tables in Parquet format, and new actions in the Help menu for reporting issues and suggesting improvements.
The update also improves the performance of various functions, ensures consistency in font and theme settings,
and addresses multiple bug fixes, including issues with automatic session reports and visualization functions.

Added
*******
* Added additional parameters to enrichment bar plots (enrichment.enrichment_bar_plot), including a new plot style ('lollipop') and observed/expected labels on the graph.
* Added a new parameter to Principal Component Analysis plots (CountFilter.pca) 'plot_grid', which can enable or disable adding a grid to PCA plots.
* RNAlysis can now load and save data tables in Parquet format (.parquet)
* Added new actions to the Help menu, allowing users to report issues, suggest issues, or open discussions.

Changed
********
* Functions in the FASTQ model are now added to automatic session reports.
* Many of the functions in RNAlysis should now run faster.
* Font type, size, and color for help tooltips should now match the global font settings.
* True/False toggle switches now scale with font size.
* When loading data tables into RNAlysis, you will now see only supported file formats by default.
* Clustering PCA plots are now plotted in proportion to the % variance explained by each PC.
* The legend in clustering PCA plots is now draggable.

Fixed
*******
* Fixed bug where data tables generated through the FASTQ model would not display properly in automatic session reports.
* Fixed bug where graphs generated through the Visualize Gene Sets window would not be added to automatic analysis report.
* When saving a file through the graphical interface, automatically-suggested filenames no longer contain illegal characters.
* Improved clarity of error message when R installation folder is not found.
* Fixed bug where some input parameter widgets in the RNAlysis graphical interface would not display properly.
* RNAlysis now provides a clearer warning message when attempting to run HDBSCAN clustering, if the hdbscan package is not installed.
* Label text in PCA plots and hierarchical clustergrams should no longer be cropped outside of the visible region of the plot.
* Fixed bug where some visualization functions, such as pair-plot (CountFilter.pairplot) would not display properly due to version mismatches between pandas and seaborn.
* Improved clarity of error messages when external apps' (kallisto, bowtie2, etc) installation folders are not found.
* Fixed bug where running the RNAlysis graphical interface on a new computer would sometimes raise an error (thanks to `NeuroRookie <https://github.com/NeuroRookie>`_ in `#25 <https://github.com/GuyTeichman/RNAlysis/issues/25>`_).
* Fixed a bug where the 'min_samples' parameter in HDBSCAN clustering could not be disabled.
* Fixed a bug where applying a function to a gene set with inplace=False would cause the new gene set to be called 'New Table'.
* Fixed a bug where RNAlysis would display the message "Pipeline saved successfully", even when the user cancels the save operation.

New Contributors
*****************
* `NeuroRookie`_ in `25`_

------------------
Version 3.8.0 of *RNAlysis* comes with several exciting new features, including the ability to generate interactive analysis reports automatically.
This feature allows users to create an interactive graph of all the datasets they loaded into RNAlysis, the functions applied, and the outputs generated during the analysis.
You can read more about this feature in the `Tutorial chapter <https://guyteichman.github.io/RNAlysis/build/tutorial.html#create-analysis-report>`_ and the `User Guide chapter <https://guyteichman.github.io/RNAlysis/build/user_guide_gui.html#rnalysis-interactive-analysis-reports>`_.

The new release also includes Pipelines for FASTQ functions, the ability to export normalization scaling factors, and other changes to improve the software's performance.
RNAlysis now supports Python 3.11, and many functions should now run faster. The software's graphic interface has also improved significantly, and users will now see a clearer error message when attempting to load unsupported file formats.
Lastly, the release also fixes several bugs.

Please note that, since Python 3.7 will be reaching end of life as of June 2023, new versions of *RNAlysis* will no longer support it.

Added
*******
* You can now generate interactive analysis reports automatically using the RNAlysis graphical interface. Read more about this feature `here <https://guyteichman.github.io/RNAlysis/build/user_guide_gui.html>`_.
* Added Pipelines for the FASTQ module (SingleEndPipeline and PairedEndPipeline), allowing you to apply a series of functions (adapter trimming, alignment, quantification, etc) to a batch of sequence/alignment files.
* Added new parameter 'return_scaling_factors' to normalization functions, that allows you to access the scaling factors calculated by RNAlysis.
* Added new parameter 'gzip_output' to CutAdapt adapter trimming (fastq.trim_adapters_single_end and fastq.trim_adapters_paired_end), allowing users to determine whether or not the output files will be gzipped.

Changed
*******
* RNAlysis now supports Python 3.11, and no longer supports Python 3.7.
* Many of the functions in RNAlysis should now run faster.
* The RNAlysis graphic interface should now boot up significantly faster.
* RNAlysis now shows an easier to understand error message when users attempt to load a table in an unsupported format (e.g. Excel files).
* CutAdapt adapter trimming (fastq.trim_adapters_single_end and fastq.trim_adapters_paired_end) now outputs non-gzipped files by default.
* Standardized all plotting functions in the filtering module to return a matplotlib Figure object, which can be further modified by users.

Fixed
*******
* RNAlysis failing to map gene IDs during GO enrichment analysis should no longer raise an error (thanks to `clockgene <https://github.com/clockgene>`_ in `#16 <https://github.com/GuyTeichman/RNAlysis/issues/16>`_).
* Fixed bug where the Command History window would not display history of the current tab immediately after clearing the current session.
* Fixed bug where adapter trimming would fail to run when using CutAdapt version >= 4.4.0.
* Fixed bug where 'Filter rows with duplicate names/IDs' (Filter.filter_duplicate_ids) would raise an error when applied to some tables.

New Contributors
*****************
* `clockgene`_ in `16`_

------------------
This version introduces small RNA read alignment using ShortStack, new filtering functions, a new optional parameter for Principal Component Analysis, improvements to the graphical interface, and bug fixes.

Added
*******
* Added small RNA read alignment using ShortStack (fastq.shortstack_align_smallrna).
* Added new filtering function 'Filter specific rows by name' (Filter.filter_by_row_name).
* Added new filtering function 'Filter rows with duplicate names/IDs' (Filter.filter_duplicate_ids).
* Function parameters in pop-up windows in the graphical interface can now be imported and exported.
* Added new parameter to Principal Component Analysis (CountFilter.pca) 'proportional_axes', that allows you to make the PCA projection axes proportional to the percentage of variance each PC explains.
* Improved clarity of error messages in the graphical interface.

Changed
*******
* Tables loaded into RNAlysis that use integer-indices will now be converted to use string-indices.
* Refactored CountFilter.from_folder into CountFilter.from_folder_htseqcount, and added a new CountFilter.from_folder method that accepts a folder of count files in any format.
* In the RNAlysis graphical interface, optional parameters that can be disabled will now display the hint "disable this parammeter?" instead of "disable input?".
* Added optional parameter 'ylim' to 'create enrichment bar-plot' function (enrichment.enrichment_bar_plot), allowing users to determine the Y-axis limits of the bar plot.
* Updated function signatures of 'Visualize gene ontology' and 'Visualize KEGG pathway' (enrichment.gene_ontology_graph and enrichment.kegg_pathway_graph) to make more sense.
* Removed parameter 'report_read_assignment_path' from featureCounts feature counting (fastq.featurecounts_single_end and fastq.featurecounts_paired_end).
* The RNAlysis graphical interface should now load more quickly.
* Progress bars in the graphical interface should now reflect elapsed/remaining time for tasks more accurately.

Fixed
*******
* Fixed bug in the function 'Split into Higly and Lowly expressed genes' (Filter.split_by_reads) where the two resulting tables would be named incorrectly (highly-expressed genes would be labeled 'belowXreads' and vice-versa).
* Fixed bug where the 'column' parameter of some functions ('Filter by percentile', 'Split by percentile', 'Filter with a number filter', 'Filter with a text filter') would not automatically detect column names in the graphical interface.
* Fixed bug where the 'numerator' and 'denominator' parameters of of the function 'Calculate fold change' would not automatically detect column names in the graphical interface.
* Fixed bug where tables with integer-indices could not be visualized properly through the graphical interface.
* Fixed bug in the function 'featureCounts single-end' (fastq.featurecounts_single_end) where setting parameter 'report_read_assignment' to any value other than None would raise an error.
* Functions that take column name/names as input (transform, filter_missing_values, filter_percentile, split_percentile) can now be applied to fold change tables (FoldCangeFilter objects).
* Fixed bug where the description for the 'n_bars' parameter of the 'create enrichment bar-plot' function (enrichment.enrichment_bar_plot) would not display properly.
* 'Visualize gene ontology' and 'Visualize KEGG pathway' (enrichment.gene_ontology_graph and enrichment.kegg_pathway_graph) now have proper parameter descriptions.
* Fixed bug where in-place intersection and difference in the filtering module would fail when using recent versions of pandas.
* Fixed bug where graphs generated through the Visualize Gene Sets window would not immediately display when using the RNAlysis stand-alone app.

------------------
This version introduces quality-of-life changes to the graphical interface, as well as bug fixes.

Added
*******
* Sample groupings in functions such as PCA, Pair-plot, etc., can now be imported and exported through the graphical interface.

Fixed
*******
* Fixed bug where the stand-alone Mac version of RNAlysis would sometimes fail to map gene IDs (directly or in enrichment analysis).