bcf compatibility

[danielecook]

I have been playing around with pyVCF and found I was able to read in bcf files by adding the following lines to the reader object:

    if fsock:
        self._reader = fsock
        if filename is None and hasattr(fsock, 'name'):
            filename = fsock.name
            compressed = compressed or filename.endswith('.gz')
        # <-- New Lines --> 
        # if filetype is bcf, read it in using subprocess.
        elif fsock.endswith(".bcf"):
            self._reader = subprocess.Popen(['bcftools','view', fsock],stdout=subprocess.PIPE).stdout
        # <-- End addition --> 
    elif filename:
        compressed = compressed or filename.endswith('.gz')
        self._reader = open(filename, 'rb' if compressed else 'rt')
    self.filename = filename
    if compressed:
        self._reader = gzip.GzipFile(fileobj=self._reader)
        if sys.version > '3':
            self._reader = codecs.getreader('ascii')(self._reader)

bcftools is required to get this to work. What are peoples thoughts on using bcftools to read in bcf format data? A few things that could be added:

• Checking whether bcftools is available
• Ensuring the bcf file is indexed

[keoughkath]

I think this would be a great improvement! More and more people are moving to BCF, and I think bcftools is common enough (and easy to install) that it wouldn't present much of a barrier.

In the meantime, thanks for providing this workaround!