0.1.0-a3

Added

• TextSequence and DigitalSequence representing a Sequence in a given mode.
• E-value properties to Hit and Domain.
• TopHits now stores a reference to the pipeline it was obtained from.
• Pipeline.Z and Pipeline.domZ properties.
• Experimental pickling support to Alphabet.
• Experimental freelist to Sequence class to avoid allocation bottlenecks when iterating on a SequenceFile without recycling sequence buffers.

Changed

• Made Sequence an abstract base class.
• Additional Pipeline parameters can be passed as keyword arguments to pyhmmer.hmmsearch.
• SequenceFile.read can now be configured to skip reading the metadata or the content of a sequence.

Removed

• Redundant SequenceFile methods.

Fixed

• doctest loader crashing on Python 3.5.
• TopHits.threshold segfaulting when being called without prior Tophits.sort call
• Unknown format argument to SequenceFile constructor not raising the right error.