Merge pull request #68 from ens-lgil/main

Update import scripts and optimise REST API

Author: ens-lgil

imports/applications/parsers/phecode.py

@@ -1,5 +1,5 @@
 import re
-from psycopg2.extras import NumericRange
+from psycopg.types.range import Range
 
 class PhecodeFullParser():
 
@@ -31,7 +31,7 @@ def parse_hazard_ratio(self):
         if len(values) == 3:
             ci_low = values[1].replace('(','')
             ci_upper = values[2].replace(')','')
-            self.score_application['hr_ci'] = NumericRange(lower=float(ci_low), upper=float(ci_upper), bounds='[]')
+            self.score_application['hr_ci'] = Range(lower=float(ci_low), upper=float(ci_upper), bounds='[]')
 
 
     def fix_phecode_id(self):

imports/omicspred/models/cohort.py

@@ -8,6 +8,7 @@ class CohortData(GenericData):
     def __init__(self,name_short:str,cohort_data:dict):
         GenericData.__init__(self)
         self.name_short = name_short
+        self.data['name_short'] = name_short
         for item in cohort_data.keys():
             if cohort_data[item]:
                  self.data[item] = cohort_data[item]

imports/omicspred/models/dataset.py

@@ -23,8 +23,8 @@ def __init__(self, publication:PublicationData, platform:PlatformData, tissue:Ti
         self.publication = publication
         self.platform = platform
         self.tissue = tissue
+        self.name = name
         if name:
-            self.name = name
             self.data['name'] = name
         if data_type in self.omics_data_types.keys():
             self.data['omics_type'] = self.omics_data_types[data_type]

imports/omicspred/models/performance.py

@@ -18,17 +18,23 @@ class PerformanceData(GenericData):
     ancestries = {
         'Ad Mixed American': 'AMR',
         'African': 'AFR',
+        'African American or Afro-Caribbean': 'AFR',
         'East Asian': 'EAS',
         'European': 'EUR',
         'European,Ad Mixed American,African,East Asian,South Asian': 'ALL',
+        'African American or Afro-Caribbean, East Asian, European, Hispanic or Latin American': 'ALL',
+        'Hispanic or Latin American': 'AMR',
         'South Asian': 'SAS'
     }
 
 
-    def __init__(self,score_name,metrics):
+    def __init__(self):
         GenericData.__init__(self)
-        self.score_name = score_name # Could be removed ?
-        self.metrics = metrics
+        self.metrics = []
+
+
+    def add_eval_type(self, eval_type:str):
+        self.data['eval_type'] = eval_type
 
 
     def add_score_model(self, score_model:Score):
@@ -48,7 +54,7 @@ def add_metrics(self,metrics_data):
         '''
         Method adding MetricData objects to the metrics array.
         '''
-        self.metrics = metrics_data
+        self.metrics.append(metrics_data)
         # for metric_type in metric_values.keys():
         #     metric_val = None
         #     pvalue_val = None
@@ -67,7 +73,8 @@ def add_metrics(self,metrics_data):
     def get_cohort_label(self, sample_model:Sample) -> str:
         cohorts = [x.name_short for x in sample_model.cohorts.all()]
         cohort_label = '_'.join(sorted(cohorts))
-        if cohort_label in ['MEC','MESA','UKB_Withheld']:
+        # if cohort_label in ['MEC','MESA','UKB_Withheld']:
+        if cohort_label in ['MEC','UKB_Withheld']:
             sample_anc = sample_model.ancestry_broad
             if sample_anc in self.ancestries.keys():
                 cohort_label = f'{cohort_label}_{self.ancestries[sample_anc]}'
@@ -77,7 +84,7 @@ def get_cohort_label(self, sample_model:Sample) -> str:
     def update_eval_type(self):
         eval_type = self.data['eval_type']
         eval_type_choices = Performance.eval_type.field.choices
-        eval_types =  {}
+        eval_types = {}
         for choice in eval_type_choices:
             eval_types[choice[1]] = choice[0]
         if eval_type in eval_types.keys():
@@ -92,7 +99,8 @@ def create_model(self):
         '''
         try:
             with transaction.atomic():
-                self.update_eval_type()
+                if self.data['eval_type'] and len(self.data['eval_type']) > 2:
+                    self.update_eval_type()
                 self.model = Performance(**self.data)
                 self.model.save()
 

imports/omicspred/models/sample.py

@@ -1,6 +1,7 @@
 import re
 from django.db import IntegrityError, transaction
 from imports.generic_model import GenericData
+from imports.omicspred.models.cohort import CohortData
 from omicspred.models import Sample
 
 class SampleData(GenericData):
@@ -54,7 +55,10 @@ def create_model(self):
                         if field == 'cohorts':
                             # Stored as list of CohortData -> Cohort Model
                             for cohort_data in val:
-                                cohort = cohort_data.create_model()
+                                if isinstance(cohort_data,CohortData):
+                                    cohort = cohort_data.create_model()
+                                else:
+                                    cohort = cohort_data
                                 cohorts.append(cohort)
                             continue
                         elif field in ['ancestry_broad','ancestry_country','ancestry_free']:

imports/omicspred/models/score.py

@@ -43,7 +43,9 @@ def create_model(self):
         try:
             with transaction.atomic():
                 self.model = Score()
-                self.model.set_score_ids(self.next_id_number(Score, 'num'))
+                id_number = Score.objects.latest('num').pk + 1
+                # self.model.set_score_ids(self.next_id_number(Score, 'num'))
+                self.model.set_score_ids(id_number)
                 for field, val in self.data.items():
                     setattr(self.model, field, val)
                 self.model.save()

imports/omicspred/spreadsheets/spreadsheet.py

@@ -371,10 +371,12 @@ def extract_data(self):
                 metrics.append(metric)
             # print(f"==> Metrics: {len(metrics)}")
             # Perforance
-            performance = PerformanceData(score_name, metrics)
+            # performance = PerformanceData(score_name, metrics)
+            performance = PerformanceData()
+            for metric in metrics:
+                performance.add_metrics(metric)
             for p_key, p_val in performance_data.items():
                 performance.add_data(p_key, p_val)
-            performance.add_metrics(metrics)
             # print(f"==> Performance Metrics: {performance.metrics}")
 
             # Take into account when there are more than 1 performance per score

requirements.txt

@@ -1,8 +1,8 @@
 #### Django library ####
 Django==5.2.8
 #### Python PostgreSQL library ####
-psycopg2-binary==2.9.11
-# psycopg-binary==3.2.10
+# psycopg2-binary==2.9.11
+psycopg[binary]==3.3.2
 # psycopg==3.2.10
 #### Django REST API libraries ####
 djangorestframework==3.16.1