eds_scikit.biology.utils.process_concepts

ConceptsSet

ConceptsSet(name: str, concept_codes: List[str] = None)

Class defining the concepts-sets with 2 attributes:

• name: the name of the concepts-set
• concept_codes : the list of concepts codes included in the concepts-set

Source code in eds_scikit/biology/utils/process_concepts.py

def __init__(self, name: str, concept_codes: List[str] = None):
    self.name = name
    if concept_codes is None:
        self.concept_codes = fetch_concept_codes_from_name(name)
    else:
        self.concept_codes = concept_codes

fetch_all_concepts_set

fetch_all_concepts_set(concepts_sets_table_name: str = 'default_concepts_sets') -> List[ConceptsSet]

Returns a list of all the concepts-sets of the chosen tables. By default, the table is here.

PARAMETER	DESCRIPTION
concepts_sets_table_name	Name of the table to extract concepts-sets from TYPE: str, optional DEFAULT: 'default_concepts_sets'

RETURNS	DESCRIPTION
List[ConceptsSet]	The list of all concepts-sets in the selected table

Source code in eds_scikit/biology/utils/process_concepts.py

def fetch_all_concepts_set(
    concepts_sets_table_name: str = "default_concepts_sets",
) -> List[ConceptsSet]:
    """Returns a list of all the concepts-sets of the chosen tables. By default, the table is [here][concepts-sets].

    Parameters
    ----------
    concepts_sets_table_name : str, optional
        Name of the table to extract concepts-sets from

    Returns
    -------
    List[ConceptsSet]
        The list of all concepts-sets in the selected table
    """
    concepts_sets = []
    default_concepts_sets = getattr(datasets, concepts_sets_table_name)
    for concepts_set_name in default_concepts_sets.concepts_set_name:
        concepts_sets.append(ConceptsSet(concepts_set_name))
    logger.info("Fetch all concepts-sets from table {}", concepts_sets_table_name)
    return concepts_sets

get_concept_src_to_std

get_concept_src_to_std(concept: DataFrame, concept_relationship: DataFrame, concepts_sets: List[ConceptsSet], standard_concept_regex: dict = default_standard_concept_regex, standard_terminologies: List[str] = default_standard_terminologies) -> pd.DataFrame

Process Concept and Concept Relationship tables to obtain a wide DataFrame that gives for all concepts-sets the source code along with the standard concepts codes.

PARAMETER	DESCRIPTION
concept	Concept OMOP table TYPE: DataFrame
concept_relationship	Concept Relationship OMOP table TYPE: DataFrame
concepts_sets	List of concepts-sets to select TYPE: List[ConceptsSet]
standard_concept_regex	EXAMPLE: {"LOINC": "[0-9]{2,5}[-][0-9]","AnaBio": "[A-Z][0-9]{4}"} TYPE: dict, optional DEFAULT: default_standard_concept_regex
standard_terminologies	EXAMPLE: ["LOINC", "AnaBio"] TYPE: List[str], optional DEFAULT: default_standard_terminologies

RETURNS	DESCRIPTION
pd.DataFrame	DataFrame with a column for the source concepts codes and columns for the selected standard concepts codes

Source code in eds_scikit/biology/utils/process_concepts.py

def get_concept_src_to_std(
    concept: DataFrame,
    concept_relationship: DataFrame,
    concepts_sets: List[ConceptsSet],
    standard_concept_regex: dict = default_standard_concept_regex,
    standard_terminologies: List[str] = default_standard_terminologies,
) -> pd.DataFrame:
    """Process ``Concept`` and ``Concept Relationship`` tables to obtain a wide DataFrame that gives for all concepts-sets the source code along with the standard concepts codes.

    Parameters
    ----------
    concept : DataFrame
        [Concept](https://www.ohdsi.org/web/wiki/doku.php?id=documentation:cdm:concept) OMOP table
    concept_relationship : DataFrame
        [Concept Relationship](https://www.ohdsi.org/web/wiki/doku.php?id=documentation:cdm:concept_relationship) OMOP table
    concepts_sets : List[ConceptsSet]
        List of concepts-sets to select
    standard_concept_regex : dict, optional
        **EXAMPLE**: `{"LOINC": "[0-9]{2,5}[-][0-9]","AnaBio": "[A-Z][0-9]{4}"}`
    standard_terminologies : List[str], optional
        **EXAMPLE**: `["LOINC", "AnaBio"]`


    Returns
    -------
    pd.DataFrame
        DataFrame with a column for the source concepts codes and columns for the selected standard concepts codes"""
    check_columns(
        concept,
        required_columns=[
            "concept_id",
            "concept_code",
            "concept_name",
            "vocabulary_id",
        ],
        df_name="concept",
    )
    check_columns(
        concept_relationship,
        required_columns=[
            "concept_id_1",
            "concept_id_2",
            "relationship_id",
        ],
        df_name="concept_relationship",
    )

    # Get desired concepts
    filtered_concepts = _filter_concepts(concept, concepts_sets, standard_concept_regex)

    # Get only parent concepts
    concept_relationship = concept_relationship[
        concept_relationship.relationship_id.isin(["Maps to", "Mapped from"])
    ]
    concept_relationship = concept_relationship.drop(columns="relationship_id")

    # Get the complete standard concept id list
    original_concpet_id = filtered_concepts.merge(
        concept_relationship, left_on="concept_id", right_on="concept_id_2", how="inner"
    )[["concept_id_1", "concepts_set"]]
    original_concpet_id.drop_duplicates("concept_id_1", inplace=True)
    related_concept_id = concept_relationship.merge(
        original_concpet_id, on="concept_id_1", how="inner"
    )
    long_src_to_std = concept.merge(
        related_concept_id, left_on="concept_id", right_on="concept_id_2", how="inner"
    )
    long_src_to_std = long_src_to_std.drop(columns=["concept_id", "concept_id_2"])
    long_src_to_std.rename(
        columns={
            "concept_id_1": "source_concept_id",
        },
        inplace=True,
    )
    long_src_to_std = to("pandas", long_src_to_std)

    # Convert long src_std to a wide src_std
    related_terminologies_concept_id = []
    for terminology in standard_terminologies:
        # Filter each terminology concept id
        terminology_filter = long_src_to_std.vocabulary_id.str.contains(
            terminology, case=False, regex=False
        )
        related_terminology_concept_id = long_src_to_std[terminology_filter].copy()
        related_terminology_concept_id.rename(
            columns={
                "concept_code": "{}_concept_code".format(terminology),
                "concept_name": "{}_concept_name".format(terminology),
                "vocabulary_id": "{}_vocabulary_id".format(terminology),
            },
            inplace=True,
        )
        related_terminology_concept_id.drop_duplicates(
            ["source_concept_id", "{}_concept_code".format(terminology)], inplace=True
        )
        related_terminologies_concept_id.append(related_terminology_concept_id)

    # Merge all terminologies
    if len(related_terminologies_concept_id) >= 2:
        wide_src_to_std = reduce(
            lambda left, right: left.merge(
                right, on=["source_concept_id", "concepts_set"], how="outer"
            ),
            related_terminologies_concept_id,
        )
    elif len(related_terminologies_concept_id) == 1:
        wide_src_to_std = related_terminologies_concept_id[0]

    else:
        return long_src_to_std[["source_concept_id"]]

    if all(
        terminology in standard_terminologies
        for terminology in default_standard_concept_regex
    ):
        # Get LOINC NAME and code from ITM
        wide_src_to_std = _override_name_code_with_itm(wide_src_to_std)

    wide_src_to_std = _rename_duplicate_code_with_different_names(wide_src_to_std)
    wide_src_to_std.fillna("Non renseigné", inplace=True)

    return wide_src_to_std