Source code for galaxy.datatypes.annotation

import logging
import tarfile
from typing import TYPE_CHECKING

from galaxy.datatypes.binary import CompressedArchive
from galaxy.datatypes.data import (
    get_file_peek,
    Text,
)
from galaxy.datatypes.sniff import (
    build_sniff_from_prefix,
    FilePrefix,
)
from galaxy.util import nice_size

if TYPE_CHECKING:
    from galaxy.model import DatasetInstance

log = logging.getLogger(__name__)


[docs]@build_sniff_from_prefix
class SnapHmm(Text):
    file_ext = "snaphmm"
    edam_data = "data_1364"

[docs]    def set_peek(self, dataset: "DatasetInstance", **kwd) -> None:
        if not dataset.dataset.purged:
            dataset.peek = get_file_peek(dataset.file_name)
            dataset.blurb = "SNAP HMM model"
        else:
            dataset.peek = "file does not exist"
            dataset.blurb = "file purged from disc"

[docs]    def display_peek(self, dataset: "DatasetInstance") -> str:
        try:
            return dataset.peek
        except Exception:
            return f"SNAP HMM model ({nice_size(dataset.get_size())})"

[docs]    def sniff_prefix(self, file_prefix: FilePrefix) -> bool:
        """
        SNAP model files start with zoeHMM
        """
        return file_prefix.startswith("zoeHMM")


[docs]class Augustus(CompressedArchive):
    """
    Class describing an Augustus prediction model
    """

    file_ext = "augustus"
    edam_data = "data_0950"
    compressed = True

[docs]    def set_peek(self, dataset: "DatasetInstance", **kwd) -> None:
        if not dataset.dataset.purged:
            dataset.peek = "Augustus model"
            dataset.blurb = nice_size(dataset.get_size())
        else:
            dataset.peek = "file does not exist"
            dataset.blurb = "file purged from disk"

[docs]    def display_peek(self, dataset: "DatasetInstance") -> str:
        try:
            return dataset.peek
        except Exception:
            return f"Augustus model ({nice_size(dataset.get_size())})"

[docs]    def sniff(self, filename: str) -> bool:
        """
        Augustus archives always contain the same files
        """
        if filename and tarfile.is_tarfile(filename):
            with tarfile.open(filename, "r") as temptar:
                for f in temptar:
                    if not f.isfile():
                        continue
                    if (
                        f.name.endswith("_exon_probs.pbl")
                        or f.name.endswith("_igenic_probs.pbl")
                        or f.name.endswith("_intron_probs.pbl")
                        or f.name.endswith("_metapars.cfg")
                        or f.name.endswith("_metapars.utr.cfg")
                        or f.name.endswith("_parameters.cfg")
                        or f.name.endswith("_parameters.cgp.cfg")
                        or f.name.endswith("_utr_probs.pbl")
                        or f.name.endswith("_weightmatrix.txt")
                    ):
                        return True
                    else:
                        return False
        return False