Transcript-transformer

Fix: only parse samples listed by --samples into database. Important for parallel jobs.

Features

- `--pretrain` Support for pre-training a new models using ribosome profiling data. **It is now possible to apply RiboTIE on new species**
- `--parallel` : Support for parallel processing of ribo-seq samples for easy deployment on HPC servers.
- Ability to process multiple separate config files. As such, it is possible to group settings together. e.g. `ribotie base.yml samples.yml pretrained_model.yml`
- `--version` flag

Changes

- refactor : `--ribo` to `--samples`
- updated config file handling
- multiple configs
- all arguments
- allow overwriting (yaml parsed before bash arguments)

Fix typo

Alter transcript_id of gff files to reflect ORF_id from csv files.

Revert minimum version of python down to 3.10 to allow use of singularity containers.

- Update package dependencies to avoid conflicts.