GitLab

added order as an optional argument to list_dbIDs, mosy useful for genes and transcripts so caching is maximised

modified execute calls to make compatible with other databases. Need to use bind_param first, with the type of the parameter, and execute later

bug fix in Bio::EnsEMBL::Feature
added import of warning method into DnaAlignFeatureAdaptor

fixes RepeatFeatureAdaptor::fetch_by_dbID

fixed some memory leak problems, store Analysis object will now happen on feature store. That could slow down the Pipeline

Updates so slices can handle negative strand, not finished yet, and not tested

-Sped up general feature retrieval a little bit

Fixed a bug in _objs_from_sth where the hit start coordinate was accidentally being used as the hit score.

-'Flatening' FeaturePair object so that it no longer contains 2 SeqFeatures; It is now isa SeqFeature with some additional Hit attributes (object still bhaves as if it contained two features
-Adding 'fast' constructors to DnaDnaAlignFeatures and PepDnaAlignFeatures
-Some other minor tweaks such as using bind columns instead of fetchrow_hashref

    fetch_by_contig_id(int) -> fetch_by_Contig(RawContig)
  store(int, list of Features) -> store(list of Features with attached contigs)

and Updated DBConnection so that get_adaptor no longer checks the object type of adaptors

BaseFeatureAdaptor was returning list when it should have been returning scalar - fixed.  Added dbID to SimpleFeature, dbID of DnaDnaAlignFeatures and PepDnaAlignFeatures now set by their respective feature adaptors upon retrieval from the database

updated ProteinAlignFeatureAdaptor DnaAlignFeatureAdaptor to extend new BaseAlignFeatureAdaptor thereby eliminating quite a few lines of duplicated code

ExonAdaptor this now expects to store new style supporting features is, storing alignfeatures then appropriate idss in supporting feature table which is now a linkage table

warning all other code for supporting features will be broken if you are using a new style supporting feature table or the store method will be broken if you are using an old style supporting feature table this will hopefully be fixed soon so update your supporting feature tables!

AnalysisAdaptor - logic_name must be unique so fetch_by_logic_name now only returns one analysis object
DnaAlignFeatureAdaptor - formatting print statements etc
ProteinAlignFeatureAdaptor - Now follows the same format as DnaAlignFeatureAdaptor, has methods to fetch by contig, slice and assembly location and when constrainted by score
or percent id all methods can take an optional type which is logic name
RepeatFeatureAdaptor
PredictionTranscriptAdaptor all have methods which fetch by slice, contig_id and assembly location and each method has optional type which should be logic_name
SimpleFeatureAdaptor

added new methods to get things by contig_id, slice or assembly location and also using score and percent_id as constraints
all methods can take an optional type which is logic_name and then feature realting to that logic_name only will be returned

altered fetch_by_assembly_location so returns assembly coordinates rather than raw contig coordinates and so features not on golden path are skipped

The only modelue with any significant changes other than that is Transcript where moltype
has been changed to alphabet due to another change in Bioperl 1.0