genesis/api_clients/bioportal_api.py

# genesis/api_clients/bioportal_api.py
import requests
import os
from urllib.parse import quote

BIOPORTAL_API_KEY = os.getenv("BIOPORTAL_API_KEY")  # Store this in Hugging Face secrets
BIOPORTAL_BASE_URL = "https://data.bioontology.org"

if not BIOPORTAL_API_KEY:
    raise ValueError("BIOPORTAL_API_KEY not found in environment variables.")

HEADERS = {
    "Authorization": f"apikey token={BIOPORTAL_API_KEY}"
}


def search_concept(term: str, ontology: str = None, max_results: int = 10):
    """
    Search for a biomedical concept across BioPortal ontologies.
    
    Args:
        term (str): The search term (e.g., "BRCA1", "glioblastoma").
        ontology (str): Optional ontology acronym (e.g., "NCIT", "SNOMEDCT").
        max_results (int): Maximum number of results.
    """
    params = {
        "q": term,
        "pagesize": max_results
    }
    if ontology:
        params["ontology"] = ontology

    url = f"{BIOPORTAL_BASE_URL}/search"
    response = requests.get(url, headers=HEADERS, params=params)
    response.raise_for_status()
    return response.json().get("collection", [])


def get_concept_details(ontology_acronym: str, concept_id: str):
    """
    Fetch detailed information about a concept given ontology acronym and concept ID.
    """
    encoded_id = quote(concept_id, safe="")
    url = f"{BIOPORTAL_BASE_URL}/ontologies/{ontology_acronym}/classes/{encoded_id}"
    response = requests.get(url, headers=HEADERS)
    response.raise_for_status()
    return response.json()


def get_ontologies():
    """
    List all available ontologies in BioPortal.
    """
    url = f"{BIOPORTAL_BASE_URL}/ontologies"
    response = requests.get(url, headers=HEADERS)
    response.raise_for_status()
    return [
        {"acronym": o.get("acronym"), "name": o.get("name"), "description": o.get("description")}
        for o in response.json()
    ]


def map_term_across_ontologies(term: str, ontologies: list = None):
    """
    Map a term across multiple ontologies to unify synonyms and cross-references.
    """
    mappings = {}
    target_ontologies = ontologies or [o["acronym"] for o in get_ontologies()]
    
    for ont in target_ontologies:
        results = search_concept(term, ontology=ont, max_results=1)
        if results:
            mappings[ont] = {
                "label": results[0].get("prefLabel"),
                "uri": results[0].get("@id"),
                "synonyms": results[0].get("synonym", []),
            }
    return mappings