[FIX] duplicate analyses (#809)

* do not allow duplicate analyses

* fix update logic

* fix condition when both record and data have no id

store/neurostore/resources/base.py

@@ -69,6 +69,10 @@ class BaseView(MethodView):
     _view_fields = {}
     # _default_exclude = None
 
+    @classmethod
+    def check_duplicate(cls, data, record):
+        return False
+
     def get_affected_ids(self, ids):
         """
         Get all the ids that are affected by a change to a record..
@@ -223,16 +227,16 @@ def load_nested_records(cls, data, record=None):
     @classmethod
     def update_or_create(cls, data, id=None, user=None, record=None, flush=True):
         """
-        scenerios:
-        1. cloning a study
-          a. clone everything, a study is an object
-        2. cloning a studyset
-          a. studies are linked to a studyset, so create a new studyset with same links
-        3. cloning an annotation
-          a. annotations are linked to studysets, update when studyset updates
-        4. creating an analysis
+        Scenarios:
+        1. Cloning a study
+          a. Clone everything, a study is an object
+        2. Cloning a studyset
+          a. Studies are linked to a studyset, so create a new studyset with same links
+        3. Cloning an annotation
+          a. Annotations are linked to studysets, update when studyset updates
+        4. Creating an analysis
           a. I should have to own all (relevant) parent objects
-        5. creating an annotation
+        5. Creating an annotation
             a. I should not have to own the studyset to create an annotation
         """
 
@@ -291,6 +295,14 @@ def update_or_create(cls, data, id=None, user=None, record=None, flush=True):
 
             return record
 
+        data["user_id"] = current_user.external_id
+        if hasattr(record, "id"):
+            data["id"] = record.id
+        # check to see if duplicate
+        duplicate = cls.check_duplicate(data, record)
+        if duplicate:
+            return duplicate
+
         # Update all non-nested attributes
         for k, v in data.items():
             if k in cls._parent and v is not None:

store/neurostore/resources/data.py

@@ -931,6 +931,57 @@ def join_tables(self, q, args):
             )
         return super().join_tables(q, args)
 
+    @classmethod
+    def check_duplicate(cls, data, record):
+        study_id = data.get("study_id")
+
+        if hasattr(record, "id") and record.id and record.id == data.get("id"):
+            # not a duplicate, same record
+            return False
+
+        if hasattr(record, "study") and record.study:
+            study = record.study
+        else:
+            study = Study.query.filter_by(id=study_id).first()
+
+        if not study:
+            return False
+
+        name = data.get("name")
+        user_id = data.get("user_id")
+        coordinates = data.get("points")
+
+        for analysis in study.analyses:
+            if (
+                analysis.name == name
+                and analysis.user_id == user_id
+                and cls._compare_coordinates(analysis.points, coordinates)
+            ):
+                return analysis
+
+        return False
+
+    @staticmethod
+    def _compare_coordinates(existing_points, new_points):
+        # Create a dictionary to map point IDs to their coordinates
+        existing_points_dict = {
+            point.id: (point.x, point.y, point.z) for point in existing_points
+        }
+
+        # Create sets for comparison
+        existing_points_set = {(point.x, point.y, point.z) for point in existing_points}
+        new_points_set = set()
+
+        for point in new_points:
+            if "x" in point and "y" in point and "z" in point:
+                new_points_set.add((point["x"], point["y"], point["z"]))
+            elif "id" in point and point["id"] in existing_points_dict:
+                new_points_set.add(existing_points_dict[point["id"]])
+            else:
+                return False  # If the point doesn't have coordinates or a valid ID, return False
+
+        return existing_points_set == new_points_set
+
 
 @view_maker
 class ConditionsView(ObjectView, ListView):

store/neurostore/tests/api/test_analyses.py

@@ -151,3 +151,34 @@ def test_post_analysis_without_order(auth_client, ingest_neurosynth, session):
 
     # Check if the 'order' field is not None
     assert resp.json()["order"] is not None
+
+
+def test_create_duplicate_analysis(auth_client, ingest_neurosynth, session):
+    # Get an existing analysis from the database
+    analysis_db = Analysis.query.first()
+    analysis = AnalysisSchema().dump(analysis_db)
+    id_ = auth_client.username
+    user = User.query.filter_by(external_id=id_).first()
+    analysis_db.study.user = user
+    for a in analysis_db.study.analyses:
+        a.user = user
+        session.add(a)
+    session.add(analysis_db.study)
+    session.commit()
+
+    # Remove fields that are auto-generated
+    for k in ["user", "id", "created_at", "updated_at", "entities"]:
+        analysis.pop(k, None)
+
+    # Create the first analysis
+    resp = auth_client.post("/api/analyses/", data=analysis)
+    assert resp.status_code == 200
+
+    # Attempt to create a duplicate analysis
+    resp_duplicate = auth_client.post("/api/analyses/", data=analysis)
+    assert resp_duplicate.status_code == 200
+
+    # Check if the duplicate analysis is the same as the original
+    original_analysis = resp.json()
+    duplicate_analysis = resp_duplicate.json()
+    assert original_analysis["id"] == duplicate_analysis["id"]