Added efficiency boost to fetch_all_by_external_name and sub-methods. Wildcards now escaped in positions 1 and 3 unless user supplies override flag.

POD updated accordingly.

Xref table overhaul. Uniqueness enforced, coupled with insert-then-query execution order in DBEntryAdaptor.

Data patching is slow.

Adding the no_alias check to the get_species_and_object_type() method to avoid warning about the lack of a lookup DB

load_registry_from_db - load the stable ids lookup database if found; get_species_and_object_type - use the stable id lookup database if found

ENSCORESW-159. Fix for issue found with coding_region_start which was using the wrong method in the transcript object.

bringing the example into line with the new 'get_adaptor' method, rather than the old 'db adaptor' way

Altering the feature iteration code to while and shift rather than foreaching the array. It's far more memory efficient as we are dumping data once we are finished with it

ENSCORESW-140. Issue caused by 1 length sequences. Fixed by changing the logic in fasta serialiser. Test cases now assert for length 1 seqs and for other potential scenarios.

ENSCORESW-137. Code could not handle empty lines nor attributes with multiple entries. Hope this isn't going to mess up anything else

[ENSCORESW-136]. Adding count methods to the API. Also triggered bug fixing of SqlHelper which was not behaving as expected. Test cases added

fix for ChainedAssemblyMapper to map using the destination slice if one is provided http://www.ebi.ac.uk/panda/jira/browse/ENSCORESW-108