[Biojava-l] Added Blast-like SAX -> SeqSimilaritySearch objects

[Keith James]

The package org.biojava.bio.program.ssbind now contains some classes
for converting Blast-like SAX events into SeqSimilaritySearch*
objects. This should work for Fasta and supported Blast versions. The
SeqSimilaritySearch* classes were not designed with HMMER in mind, so
it's not supported in this case.

Steps for using the package are pretty simple:

Create a SAX parser:

    XMLReader parser = (XMLReader) new FastaSearchSAXParser();
or
    XMLReader parser = (XMLReader) new BlastLikeSAXParser();

Create a SAX handler which converts SAX events into calls a
SearchContentHandler:

    SeqSimilarityAdapter adapter = new SeqSimilarityAdapter();

Create the SearchContentHandler itself and a List for it to write the
SeqSimilaritySearchResults into:

    List results = new ArrayList();
    BlastLikeSearchBuilder builder = new BlastLikeSearchBuilder(results);

Set the builder's source of query sequences and databases:

    builder.setQuerySeqHolder(queryDB);
    builder.setSubjectDBInstallation(dbInstallation);

Set up the chain and start the parse:

adapter.setSearchContentHandler(builder);

parser.setContentHandler(adapter);
parser.parse(new InputSource(foo));

Making unit tests for all the delegate SAX handlers would have got
complex, so for now there are tests for the builder specific for each
supported Blast version + Fasta. If people are finding some of the
tests take a while, I could swap out the blast output files for some
smaller ones (esp. tblastn/x).

At some point I'll write a HomolgyFeatureAdapter to make
HomolgyFeatures from the same data sources (plus HMMER).

Keith

-- 

-= Keith James - kdj@sanger.ac.uk - http://www.sanger.ac.uk/Users/kdj =-
The Sanger Centre, Wellcome Trust Genome Campus, Hinxton, Cambs CB10 1SA