class ReadFromBGZF(beam.PTransform):
"""Reads variants from BGZF."""
def __init__(self,
input_files,
representative_header_lines,
allow_malformed_records,
pre_infer_headers,
sample_name_encoding=SampleNameEncoding.WITHOUT_FILE_PATH,
use_1_based_coordinate=False
):
# type: (List[str], List[str], bool, bool, int, bool) -> None
"""Initializes the transform.
Args:
input_files: The BGZF file paths to read from.
representative_header_lines: Header definitions to be used for parsing
VCF files.
allow_malformed_records: If true, malformed records from VCF files will be
returned as `MalformedVcfRecord` instead of failing the pipeline.
pre_infer_headers: If true, drop headers and make sure PySam return the
exact data for variants and calls, without type matching.
sample_name_encoding: specify how we want to encode sample_name mainly
to deal with same sample_name used across multiple VCF files.
use_1_based_coordinate: specify whether the coordinates should be stored
in BQ using 0 based exclusive (default) or 1 based inclusive coordinate.
"""
self._input_files = input_files
self._representative_header_lines = representative_header_lines
self._allow_malformed_records = allow_malformed_records
self._pre_infer_headers = pre_infer_headers
self._sample_name_encoding = sample_name_encoding
self._use_1_based_coordinate = use_1_based_coordinate
def _read_records(self, (file_path, block)):
# type: (Tuple[str, Block]) -> Iterable(Variant)
"""Reads records from `file_path` in `block`."""
record_iterator = vcf_parser.PySamParser(
file_path,
block,
filesystems.CompressionTypes.GZIP,
self._allow_malformed_records,
representative_header_lines=self._representative_header_lines,
splittable_bgzf=True,
pre_infer_headers=self._pre_infer_headers,
sample_name_encoding=self._sample_name_encoding,
use_1_based_coordinate=self._use_1_based_coordinate)
for record in record_iterator:
yield record
def _read_records(self, file_path_and_block_tuple):
# type: (Tuple[str, Block]) -> Iterable(Variant)
"""Reads records from `file_path` in `block`."""
(file_path, block) = file_path_and_block_tuple
record_iterator = vcf_parser.PySamParser(
file_path,
block,
filesystems.CompressionTypes.GZIP,
self._allow_malformed_records,
representative_header_lines=self._representative_header_lines,
splittable_bgzf=True,
pre_infer_headers=self._pre_infer_headers,
sample_name_encoding=self._sample_name_encoding,
use_1_based_coordinate=self._use_1_based_coordinate)
for record in record_iterator:
yield record
def read_records(self,
file_name, # type: str
range_tracker # type: range_trackers.OffsetRangeTracker
):
# type: (...) -> Iterable[MalformedVcfRecord]
record_iterator = vcf_parser.PySamParser(
file_name,
range_tracker,
self._compression_type,
self._allow_malformed_records,
file_pattern=self._pattern,
representative_header_lines=self._representative_header_lines,
pre_infer_headers=self._pre_infer_headers,
sample_name_encoding=self._sample_name_encoding,
use_1_based_coordinate=self._use_1_based_coordinate,
buffer_size=self._buffer_size,
skip_header_lines=0)
# Convert iterator to generator to abstract behavior
for record in record_iterator:
yield record