Funannotate

Commands: compare Compare annotations to reference (GFF3 or GBK annotations)
tbl2gbk Convert TBL format to GenBank format
gbk2parts Convert GBK file to individual components
gff2proteins Convert GFF3 + FASTA files to protein FASTA
gff2tbl Convert GFF3 format to NCBI annotation table (tbl)

* bug fix for `funannotate remote` moving logfile
* bug fix for mapping proteins to genome where tmp folder wasn't being properly removed
* run GO enrichment in parallel in `funannotate compare`
* update colors in some graphs from `funannotate compare` to 24-pack Crayola colors
* add option to use `iqtree` to draw ML phylogeny in `funannotate compare`
* bug fix for `funannotate database` command where it was not displaying table correctly.

* bug fix for `funannotate setup` added missing shutil library import

* bug fix for weights being set for Augustus HiQ models in `funannotate predict`
* bug fix for download_buscos function
* bug fix for `funannotate annotate` where tbl file was occasionally not being parsed correctly --> re-write of parsing function
* fix bug in antiSMASH/MiBIG parsing
* add method to try to recover from failed GeneMark run
* several bug fixes for `funannotate update` related to UTRs and multiple transcripts per locus.
* added missing dependencies to `funannotate check`
* updated code to work with PASA > v2.3 - this is important PASA update that allows SQLite usage instead of MySQL
* improved terminal log output to tell user which files (with locations) are being re-used if they are found.

* v1.2.0 now supports multiple transcripts per gene locus. The funannotate pipeline will only generate multiple transcripts per locus if given evidence in the form of RNA-seq data, this is done in the `funannotate update` command. It should also now support input with multiple transcripts as well.
* move installation of busco models to `funannotate setup`
* added annotation edit distance (AED) to `funannotate update` to record the changes in annotation. As well the PASA annotation update text file is changed to incorporate these changes as well
* accessory script `util/compare2annotations.py` can compare multiple annotations in either GFF3 or GBK format to a reference, generating summary stats as well as individual gene stats (AED per mRNA and CDS)
* added a `--drop` option to `funannotate fix` that you can remove unwanted gene model annotations, to use pass a file containing locus_tag (1 per line) to the `--drop` parameter
* fix bug in finding high-quality Augustus predictions (HiQ) models in `funannotate predict`
* `funannotate predict` will now detect if a `training` folder exists in output directory, if it does it will find the correct PASA, BAM, and Trinity output and use automatically during the prediction step.

* fix for braker to work on docker. For some reason (I don't know why) the symlinks that braker tries to create cause an error when run on docker. The error references too many levels of symlinks essentially. To circumvent this, I modified `braker.pl` code to copy instead of symlink. Also fixed the `braker.pl --version` option which was broke in most recent release.
* Note for a "normal" system, v1.1.0 should work fine. The updated braker code was run on both docker and Mac native and runs fine on those, hopefully also working well on linux.

* bumping version to 1.1.0 to highlight that v1.0.X versions have a bug in the tbl annotation file and will not pass GenBank specs. This was derived from dropping GAG from funannotate I had the tbl spec wrong for adding transcript_id and protein_id to both CDS and mRNA features.
* fixes for `funannotate update` and properly filtering overlapping genes
* fix for `funannotate annotate` that was switching the 5' and 3' partial gene designations on crick orientated gene models, causing them to look correct after predict step and then become errors after annotate step
* added Braker 2.0.3 to funannotate.... this was necessary as `braker.pl --version` doesn't display the version number so I can't enforce a version requirement. The larger issue has to do with how the different versions of braker save the output data, there are at least 3 different behaviors in the last 4 or 5 versions which makes impossible for funannotate to determine where output will be.