Transcript-transformer

- Update package dependencies to avoid conflicts.

- Add multiQC reports
Reports are added to the output folder that are picked up by multiQC.

Make inclusion of canonical CDSs default behavior when outputting gtf files.
rename flag `include_annotated` with flag `exclude_annotated`.

New features:
- A GTF file featuring the top predictions is now also generated when compiling results.
- `--include_annotated` : include annotated CDSs to generated GTF file

- New flags for filtering behavior of the result output table:
- `--include_invalid_TTS` : includes predicted ORFs with no valid in-frame stop codon in result table.
- `--min_ORF_len` (default=15) : minimum nucleotide length of predicted translated ORFs included in result table.
- `--start_codons` (default="*TG") : regular expression denoting valid start codon to be included in result table.

changed behavior:
- IDs of merged samples now are concatenated using "&" instead of "". This might require new data to be parsed again.
- dropped `--factor` flag from `riboformer.py`

- streamline output table tags: acronym characters are uppercase (e.g., cds_idxs -> CDS_idxs)
- new output table property: 'is_CDS' : does ORF sequence equal sequence of annotated CDS.



zenodo_release

Fix a couple of regressions. Change some behavior.


**Note**:

- the columns of the output table `support_lvl`, `tag` and `biotype` have been renamed to `tr_support_lvl`, `tr_tag`, and `tr_biotype`. `orf_id` has been added. Additionally, `rpb_in_ORF`, `rpb_out_ORF` have been replaced in favor of `reads_in_ORF` and `reads_out_ORF`.
- from now on, if using `correction` flag. ORFs that are duplicate (possible after correction step) are removed. This behavior can be altered through the `--keep_duplicates` flag.

If you were already working on results from the output tables, it might be good to rerun `riboformer yaml.yml --results` to unify existing tables with new ones.