matrix.py 文件源码

def build_from_mol_counter(molecule_counter, subsample_rate=1.0,
                               subsample_result=None):
        """ Construct a GeneBCMatrices object from a MoleculeCounter.
            Args: subsample_result (dict) - Return some metrics results into this dict. """

        # Reconstruct all barcode sequences in the original matrices
        barcode_whitelist = cr_utils.load_barcode_whitelist(molecule_counter.get_barcode_whitelist())
        barcode_length = molecule_counter.get_barcode_length() or len(barcode_whitelist[0])

        gem_groups = molecule_counter.get_gem_groups()
        barcode_seqs = cr_utils.format_barcode_seqs(barcode_whitelist, gem_groups)

        # Reconstruct Gene tuples from the molecule info ref columns
        gene_ids = molecule_counter.get_ref_column('gene_ids')
        genome_ids = molecule_counter.get_ref_column('genome_ids')
        gene_names = molecule_counter.get_ref_column('gene_names')
        gene_tuples = [cr_constants.Gene(gid, gname, None, None, None) for (gid, gname) in itertools.izip(gene_ids, gene_names)]
        genes = cr_utils.split_genes_by_genomes(gene_tuples, genome_ids)

        matrices = GeneBCMatrices(genome_ids, genes, barcode_seqs)

        # Track results of subsampling
        reads = 0

        for mol in molecule_counter.get_molecule_iter(barcode_length, subsample_rate=subsample_rate):
            matrices.add(mol.genome, mol.gene_id, mol.barcode)
            reads += mol.reads

        if subsample_result is not None:
            subsample_result['mapped_reads'] = reads

        return matrices