Feature/130 cohort ontology

tracex_project/extraction/logic/modules/module_cohort_tagger.py

@@ -28,13 +28,14 @@ def execute_and_save(
         super().execute_and_save(
             df,
             patient_journey=patient_journey,
-            patient_journey_sentences=patient_journey_sentences
+            patient_journey_sentences=patient_journey_sentences,
         )
 
         cohort_tags = self.__extract_cohort_tags(patient_journey)
         cohort_dict = self.__prepare_cohort_dict(cohort_tags)
+        cohort_dict_normalized = self.normalize_coniditons_snomed(cohort_dict)
 
-        return cohort_dict
+        return cohort_dict_normalized
 
     @staticmethod
     def __extract_cohort_tags(patient_journey):
@@ -62,3 +63,23 @@ def __prepare_cohort_dict(cohort_data):
             return None
 
         return cohort_dict
+
+    @staticmethod
+    def normalize_coniditons_snomed(cohort_dict):
+        """Normalizes conditions to a SNOMED code."""
+        condition = cohort_dict.get("condition")
+        preexisting_condition = cohort_dict.get("preexisting_condition")
+
+        if condition is not None:
+            (
+                cohort_dict["condition"],
+                cohort_dict["condition_snomed_code"],
+            ) = u.get_snomed_ct_info(condition)
+
+        if preexisting_condition is not None:
+            (
+                cohort_dict["preexisting_condition"],
+                cohort_dict["preexisting_condition_snomed_code"],
+            ) = u.get_snomed_ct_info(preexisting_condition)
+
+        return cohort_dict

tracex_project/extraction/logic/orchestrator.py

@@ -246,7 +246,9 @@ def set_default_values(self):
                 "gender": None,
                 "origin": None,
                 "condition": None,
+                "condition_snomed_code": None,
                 "preexisting_condition": None,
+                "preexisting_condition_snomed_code": None,
             }
             self.set_cohort(cohort_default_values)
 

tracex_project/extraction/migrations/0024_cohort_condition_snomed_code_and_more.py

@@ -0,0 +1,22 @@
+# Generated by Django 4.2.11 on 2024-05-16 11:03
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("extraction", "0023_alter_cohort_age_alter_cohort_condition_and_more"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="cohort",
+            name="condition_snomed_code",
+            field=models.IntegerField(blank=True, null=True),
+        ),
+        migrations.AddField(
+            model_name="cohort",
+            name="preexisting_condition_snomed_code",
+            field=models.IntegerField(blank=True, null=True),
+        ),
+    ]

tracex_project/extraction/models.py

@@ -42,7 +42,9 @@ class Cohort(models.Model):
     gender = models.CharField(max_length=25, null=True, blank=True)
     origin = models.CharField(max_length=50, null=True, blank=True)
     condition = models.CharField(max_length=50, null=True, blank=True)
+    condition_snomed_code = models.IntegerField(null=True, blank=True)
     preexisting_condition = models.CharField(max_length=100, null=True, blank=True)
+    preexisting_condition_snomed_code = models.IntegerField(null=True, blank=True)
     manager = models.Manager()
 
     def __str__(self):

tracex_project/tracex/logic/constants.py

@@ -44,3 +44,16 @@
     ("attribute_location", "Location"),
 ]
 THRESHOLD_FOR_MATCH = 0.5
+SNOMED_CT_API_URL = (
+    "https://browser.ihtsdotools.org/snowstorm/snomed-ct/browser/MAIN/descriptions"
+)
+SNOMED_CT_PARAMS = params = {
+    "limit": 50,
+    "conceptActive": "true",
+    "lang": "english",
+    "skipTo": 0,
+    "returnLimit": 100,
+}
+SNOMED_CT_HEADERS = {
+    "User-Agent": "browser",
+}

tracex_project/tracex/logic/utils.py

@@ -1,26 +1,36 @@
 """Module providing various utility functions for the project."""
 import os
-from io import StringIO
-from pathlib import Path
+import json
 
 import base64
 import tempfile
 import functools
 import warnings
+
+from io import StringIO
+from pathlib import Path
+
 import pandas as pd
 import pm4py
 import numpy as np
+import requests
 
 from django.conf import settings
 from django.db.models import Q
 from openai import OpenAI
+
 from tracex.logic.logger import log_tokens_used
 from tracex.logic.constants import (
     MAX_TOKENS,
     TEMPERATURE_SUMMARIZING,
     MODEL,
     oaik,
 )
+from tracex.logic.constants import (
+    SNOMED_CT_API_URL,
+    SNOMED_CT_PARAMS,
+    SNOMED_CT_HEADERS,
+)
 
 from extraction.models import Trace
 
@@ -101,6 +111,24 @@ def calculate_linear_probability(logprob):
     return linear_prob
 
 
+def get_snomed_ct_info(term):
+    """Get the first matched name and code of a SNOMED CT term."""
+    SNOMED_CT_PARAMS["term"] = term
+    response = requests.get(
+        SNOMED_CT_API_URL, params=SNOMED_CT_PARAMS, headers=SNOMED_CT_HEADERS
+    )
+    data = json.loads(response.text)
+
+    term = None
+    code = None
+
+    if data["items"]:
+        term = data["items"][0]["term"]
+        code = data["items"][0]["concept"]["conceptId"]
+
+    return term, code
+
+
 class Conversion:
     """Class for all kinds of conversions"""