Genewalk

- Minor bug fix that affected only the custom input network options `sif_annot` and `sif_full`: added them to the `--network_source` input argument choice list.

- Increased support for alternative organisms:
- Rat as an organism is now fully automated, similar to GeneWalk for mouse. We allow rat gene identifiers as input gene list with `--id_type rgd_id`. Rat gene ids are then automatically mapped to human ortholog genes with subsequent automated network assembly.
- For organisms other than human, mouse and rat, we now allow custom gene ids with `--id_type custom`: these gene ids are not attempted to be mapped to human orthologs. So the columns hgnc_symbol and hgnc_id are not applicable in the genewalk_results.csv output file and thus not included with `--id_type custom`. When genewalk is run with this option, it requires as additional input a user-provided network file specified with `--network_file` and `--network_source`.
- More options and support for `--network_source`: `edge_list, sif, sif_annot`, or `sif_full` to enable full network assembly when the end user either provides a gene network, gene + GO annotation networks or a full GeneWalk network (gene-gene interactions, GO annotations + GO ontology relations) as input network file in SIF or edge list format. See the README section Custom input networks for full details and which to options to use for other organisms.
- Bug fix: we now include in the genewalk_results.csv output genes without GO annotations. By default, this results file contains all (properly recognized) input genes. See the perform_statistics.GeneWalk class description for the reasons on the fringe cases that input genes could not be included in the results file.
- Download resource files only when needed, instead of previously all in advance.
- Remove INDRA as a default install dependency: INDRA can be separately installed when it is needed in case of using option `--network_source indra`
- Automated gene mapping is now done within genewalk through a stand alone GeneMapper instead of previous INDRA hgnc_client calling.
- Added automated testing for the run options that require a specified `--network_file` and `--id_type rgd_id` or `custom`.

- A new visualization stage has been added that is run by default as part of the whole program or alone when calling `--stage visual`.
- This stage automatically generates plots of all GeneWalk results in the project/figures folder.
- The file `index.html` in the figures folder contains all visualization and allows the end user to easily browse the GeneWalk results:
- Interactive scatter plots to identify the regulator and moonlighter genes and
- Bar plots with GO annotations ranked by their global_padj for each input gene.
- Each figure is also saved as separate .png and/or .pdf file.
- Automated testing has been improved to cover the whole CLI.

- New globally adjusted p-values (`global_padj`) reported in output along with the gene-level adjusted p-values (`gene_padj`), the latter was previously called `mean_padj`.
- New CLI option for changing the dimensionality of the embeddings vectors
- GeneWalk now supports Python 3.6+ and is compatible with INDRA 1.19.0
- Automated testing is now done on GitHub Actions

- Adapt to use INDRA v18.
- Update README documentation to include all v1.2.* features.

- Replace outdated HGNC gene symbols with their current ones in Pathway Commons v12 knowledge base.
- Handle outdated HGNC gene symbols from user input list.
- Add missing documentation sections for readthedocs manual.