Commit a0f1ba8b authored by Ivan Vanat's avatar Ivan Vanat
Browse files

documentation updates

parent aefd50d6

nested/core/generator/generator.py

+3 −2
Original line number Diff line number Diff line
@@ -4,14 +4,15 @@ import random 
from Bio import SeqIO

from Bio.SeqRecord import SeqRecord



from nested.core.gene import Gene

from nested.core.tools.nester.gene import Gene

from nested.core.generator.te_generator import TEGenerator

from nested.core.generator.nested_element import NestedElement

from nested.utils import intervals





class Generator(object):

    """Class used to generate artificial nested elements for testing purposes

    """

    Class used to generate artificial nested elements for testing purposes



    Attributes:

        source_db (str): path to source database of TE's in fasta format

nested/core/tools/ltr/discovery_tool.py

+1 −0
Original line number Diff line number Diff line
@@ -5,6 +5,7 @@ from enum import Enum 
class DiscoveryTool(Enum):

    LTR_finder = 1

    LTRharvest = 2



    # add aliases with same value as the tool above

    finder = 1

    harvest = 2

nested/core/tools/ltr/te_finder.py

+0 −9
Original line number Diff line number Diff line
@@ -18,15 +18,6 @@ class TeFinder(DiscoveryToolInterface): 
        super().__init__()



    def run(self, sequence_id: str, sequence: Seq) -> List[TE]:

        """Run LTR finder on sequence and return list of transposons



                Arguments:

                    sequence_id (str): sequence id

                    sequence (Bio.Seq.Seq): sequence



                Returns:

                    list[TE]: list of found ltr pairs as a TE class

                """

        transposons = []



        if not os.path.exists("/tmp/nested"):

nested/core/tools/nester/gene.py

+11 −8
Original line number Diff line number Diff line
@@ -11,15 +11,17 @@ from nested.entities.te import TE 




class Gene(object):

    """Class representing a genereic gene in the program. Used in recursive calls, contain information about the gene

     domains and found TE's

    """

    Class representing one iteration of nested-nester module recursive call.



    Attributes:

        seqid (str): id of sequence

        sequence (Bio.Seq.Seq): sequence

        te_list (list[TE]): list of pairs of LTR's found from LTR finder

        domain_list (list[Domain]): list of found domains

        scores (list[float]): list of evaluations of each TE

        seqid: id of sequence

        sequence: sequence

        discovery_tool: implementation of DiscoveryToolInterface currently used

        te_list: list of pairs of LTR's found from LTR finder

        domain_finder: class used for identification of domains

        domain_list: list of found domains

        scores: list of evaluations of each TE



    """

    copia_pattern = ["ltr_left", "pbs", "GAG", "PROT", "INT", "RT", "RH", "ppt", "ltr_right"]
@@ -47,7 +49,8 @@ class Gene(object): 
        return '\n'.join(lines)



    def get_best_candidate(self) -> Optional[TE]:

        """Evaluate all LTR pairs and return best scored pair, or None if no pair found

        """

        Evaluate all LTR pairs and return best scored pair, or None if no pair found



        Returns:

            TE: best evaluated pair

nested/core/tools/nester/nester.py

+0 −13
Original line number Diff line number Diff line
@@ -7,30 +7,18 @@ from Bio.Seq import Seq 
from nested.utils import intervals

from nested.core.tools.nester.gene import Gene

from nested.core.tools.solo_ltrs import SoloLtrs

from nested.logging.logger import NesterLogger

from nested.config.config import config

from nested.core.tools.tool_interface import ToolInterface

from nested.entities.te import TE

from nested.utils.dependency_resolver import DependencyResolver

from nested.output.output_objects.nester_output_object import NesterOutputObject

from nested.core.module_management.module import Module





class Nester(ToolInterface):

    """Class represent nesting in sequence, recursively find best evaluated transposon and crop it until no new transposons found



    Attributes:

        seqid (str): id of sequence

        sequence (Bio.Seq.Seq): sequence

        nested_element (NestedElement): nested element

    """



    def __init__(self, sequence: SeqRecord, threshold: float, multiplier: float,

                 discovery_tool: str):

        super().__init__(sequence)

        self.solo_ltrs: SoloLtrs = None

        self._iteration: int = 0

        #self._logger: NesterLogger = NesterLogger('{}/{}'.format(config['logdir'], self.seqId))

        self.threshold = threshold

        self.multiplier = multiplier

        self.discovery_tool = discovery_tool
@@ -43,7 +31,6 @@ class Nester(ToolInterface): 
        return nested_list



    def _get_unexpanded_transposon_list(self, sequence: Seq, threshold: float) -> List[TE]:

        # recursivelly find and crop best evaluated transposon, return unexpanded list of found transposons

        self._iteration += 1

        gene = Gene(self.discovery_tool, self.seqId, sequence)

        candidates = gene.get_candidates_above_threshold(threshold)