Pyvcf

-------------

* Add strict whitespace mode for well formed VCFs with spaces
in sample names (thanks Marco)
* Ignore blank lines in files (thanks Martijn)
* Tweaks for handling missing data (thanks Sean)
* bcftools tests (thanks Martijn)
* record.FILTER is always a list

-------------

* Backwards incompatible change: _Call.data is now a
namedtuple (previously it was a dict)
* Optional cython version, much improved performance.
* Improvements to writer (thanks cmclean)
* Improvements to inheritance of classes (thanks lennax)

-------------

* VCF 4.1 support:
- support missing genotype 28 (thanks martijnvermaat)
- parseALT for svs 42, 48 (thanks dzerbino)
* `trim_common_suffix` method 22 (thanks martijnvermaat)
* Multiple metadata with the same key is stored (52)
* Writer improvements:
- A/G in Number INFO fields 53 (thanks lennax)
- Better output 55 (thanks cmclean)
* Allow malformed INFO fields 49 (thanks ilyaminkin)
* Added bayes factor error bias VCF filter
* Added docs on vcf_melt
* filters from libor-m (SNP only, depth per sample, avg depth per sample)
* change to the filter API, use docstring for filter description

-------------

* Performance improvements (47)
* Preserve order of INFO column (46)

-------------

* Support exponent syntax qual values (43, 44) (thanks martijnvermaat)
* Preserve order of header lines (45)

-------------

* Support whitespace in sample names
* SV work (thanks arq5x)
* Python 3 support via 2to3 (thanks marcelm)
* Improved filtering script, capable of importing local files