GitLab

stable_id_version setter/getter for core object types, sets stable_id with a . delimited version or fetches stable_id with .version concatonated (if version is set)

- Added fetch_by_stable_id_version to TranslationAdaptor for consistency
- Related tests for new function

- Ensire the "version" if tested for is digits only

- Added support in the Registry for version ids (\.\d+) support in stable_id when determining idenfitier type
- Added support to the ArchiveStableIdAdaptor for versions in stable_ids on object creation
- Test cases for identifiers with and without versions when doing lookups

- handlers in Exon, Gene, Operon, OperonTranscript and Transcript Adaptors for a versions fetch_by_stable_id call and new handler for fetch_by_stable_id_version
- Translation is handled implicitly through calls to Transcript
- Test cases for the new handlers and functionality

- Corresponding test to test functionality

Provide some accessors for people that don't have Ensembl objects but still require the SO term mappings for things.

Seems by passing @_ onwards I broke later initalisation code. Thanks
Travis for picking that up :)

Following on from Alessandro's comments from the pull request.
SequenceAdaptor can now control the cache size and FileFaidx's POD is
now correct.

For trans-spliced transcripts, augment the CDS IDs with the location, to prevent multi-row features having different seq_region names and/or strands

Seems that calling external code could assume the sequence cache is, in
fact, a hash. If the cache is replaced with the wrong data structure
(say a normal hash) then a memory leak will appear as seen in the xref
pipeline whilst dumping sequences.

This reverts commit 38b2e251.

Skipping trans-splicing transcripts and the gene if it's a single transcript gene with optimisation call.

[ENSCORESW-1451] Stop logger utility from terminating host process. error() can now be caught via try/catch or eval.

$self->get_DBEntryAdaptor() doesn't exist