Unverified Commit 0aa5616b authored by Marius van den Beek's avatar Marius van den Beek Committed by GitHub
Browse files

Merge pull request #20820 from mvdbeek/limit_sam_metadata 

[25.0] Skip sam metadata if we have too many references
parents 7bcdd373 118bf715

lib/galaxy/datatypes/binary.py

+30 −13
Original line number Diff line number Diff line
@@ -552,22 +552,29 @@ class _BamOrSam: 
    Helper class to set the metadata common to sam and bam files

    """



    max_references = 100000



    def set_meta(self, dataset: DatasetProtocol, overwrite: bool = True, **kwd) -> None:

        try:

            bam_file = pysam.AlignmentFile(dataset.get_file_name(), mode="rb")

            with pysam.AlignmentFile(dataset.get_file_name(), mode="rb", check_sq=False) as bam_file:

                # TODO: Reference names, lengths, read_groups and headers can become very large, truncate when necessary

                if bam_file.nreferences <= self.max_references:

                    dataset.metadata.reference_names = list(bam_file.references)

                    dataset.metadata.reference_lengths = list(bam_file.lengths)

                    dataset.metadata.bam_header = dict(bam_file.header.items())  # type: ignore [attr-defined]

                    dataset.metadata.read_groups = [

                read_group["ID"] for read_group in dataset.metadata.bam_header.get("RG", []) if "ID" in read_group

                        read_group["ID"]

                        for read_group in dataset.metadata.bam_header.get("RG", [])

                        if "ID" in read_group

                    ]

            dataset.metadata.sort_order = dataset.metadata.bam_header.get("HD", {}).get("SO", None)

            dataset.metadata.bam_version = dataset.metadata.bam_header.get("HD", {}).get("VN", None)

                else:

                    dataset.metadata.metadata_incomplete = True

                dataset.metadata.sort_order = bam_file.header.get("HD", {}).get("SO", None)  # type: ignore [attr-defined]

                dataset.metadata.bam_version = bam_file.header.get("HD", {}).get("VN", None)  # type: ignore [attr-defined]

        except Exception:

            # Per Dan, don't log here because doing so will cause datasets that

            # fail metadata to end in the error state

            pass

            dataset.metadata.metadata_incomplete = True





class BamNative(CompressedArchive, _BamOrSam):
@@ -656,6 +663,16 @@ class BamNative(CompressedArchive, _BamOrSam): 
        optional=True,

        no_value={},

    )

    MetadataElement(

        name="metadata_incomplete",

        default=False,

        desc="Indicates if metadata is incomplete",

        param=MetadataParameter,

        readonly=True,

        visible=False,

        optional=True,

        no_value=False,

    )



    def set_meta(self, dataset: DatasetProtocol, overwrite: bool = True, **kwd) -> None:

        _BamOrSam().set_meta(dataset, overwrite=overwrite, **kwd)
@@ -1054,7 +1071,7 @@ class BamInputSorted(BamNative): 
        """

        # The best way to ensure that BAM files are coordinate-sorted and indexable

        # is to actually index them.

        with pysam.AlignmentFile(filename=file_name) as f:

        with pysam.AlignmentFile(filename=file_name, check_sq=False) as f:

            # The only sure thing we know here is that the sort order can't be coordinate

            return f.header.get("HD", {}).get("SO") == "coordinate"  # type: ignore[attr-defined]
@@ -1074,7 +1091,7 @@ class BamQuerynameSorted(BamInputSorted): 
        """

        # The best way to ensure that BAM files are coordinate-sorted and indexable

        # is to actually index them.

        with pysam.AlignmentFile(filename=file_name) as f:

        with pysam.AlignmentFile(filename=file_name, check_sq=False) as f:

            return f.header.get("HD", {}).get("SO") != "queryname"  # type: ignore[attr-defined]