Update DEFT2020.py

DEFT2020.py

@@ -171,6 +171,8 @@ class FrenchMedMCQA(datasets.GeneratorBasedBuilder):
         all_res = []
 
         if self.config.name == "task_1":
+            
+            ratio = 0.8316
                         
             with open(filepath) as fd:
                 doc = xmltodict.parse(fd.read())
@@ -189,6 +191,8 @@ class FrenchMedMCQA(datasets.GeneratorBasedBuilder):
                     })
 
         elif self.config.name == "task_2":
+            
+            ratio = 0.8059
                         
             with open(filepath) as fd:
                 doc = xmltodict.parse(fd.read())
@@ -209,5 +213,26 @@ class FrenchMedMCQA(datasets.GeneratorBasedBuilder):
                 
                 all_res.append(obj)
 
-        for r in all_res:
-            yield r["@id"], r
+        if split != "test":
+            
+            ids = [r["id"] for r in all_res]
+    
+            random.seed(4)
+            random.shuffle(ids)
+            random.shuffle(ids)
+            random.shuffle(ids)
+            
+            train, validation = np.split(ids, [int(len(ids)*ratio)])
+    
+            if split == "train":
+                allowed_ids = list(train)
+            elif split == "validation":
+                allowed_ids = list(validation)
+            
+            for r in all_res:
+                if r["id"] in allowed_ids:
+                    yield r["id"], r
+        else:
+            
+            for r in all_res:
+                yield r["id"], r