fix(dataset): silent erros if dataset did not exist

docker-compose.dev.yml

@@ -23,6 +23,7 @@ services:
     container_name: crosspost_flask
     volumes:
       - .:/app
+      - model_cache:/models
     env_file:
       - .env
     ports:
@@ -36,6 +37,7 @@ services:
     build: .
     volumes:
       - .:/app
+      - model_cache:/models
     container_name: crosspost_worker
     env_file:
       - .env
@@ -46,3 +48,6 @@ services:
     depends_on:
       - postgres
       - redis
+
+volumes:
+  model_cache:

server/analysis/enrichment.py

@@ -3,7 +3,7 @@ import pandas as pd
 from server.analysis.nlp import NLP
 
 class DatasetEnrichment:
-    def __init__(self, df, topics):
+    def __init__(self, df: pd.DataFrame, topics: dict):
         self.df = self._explode_comments(df)
         self.topics = topics
         self.nlp = NLP(self.df, "title", "content", self.topics)

server/app.py

@@ -16,7 +16,7 @@ from flask_jwt_extended import (
 
 from server.analysis.stat_gen import StatGen
 from server.analysis.enrichment import DatasetEnrichment
-from server.exceptions import NotAuthorisedException, NotExistentDatasetException
+from server.exceptions import NotAuthorisedException, NonExistentDatasetException
 from server.db.database import PostgresConnector
 from server.core.auth import AuthManager
 from server.core.datasets import DatasetManager
@@ -162,7 +162,7 @@ def get_dataset(dataset_id):
         return jsonify(filtered_dataset), 200
     except NotAuthorisedException:
         return jsonify({"error": "User is not authorised to access this content"}), 403
-    except NotExistentDatasetException:
+    except NonExistentDatasetException:
         return jsonify({"error": "Dataset does not exist"}), 404
     except Exception:
         print(traceback.format_exc())

server/core/datasets.py

@@ -1,7 +1,7 @@
 import pandas as pd
 from server.db.database import PostgresConnector
 from psycopg2.extras import Json
-from server.exceptions import NotAuthorisedException
+from server.exceptions import NotAuthorisedException, NonExistentDatasetException
 
 class DatasetManager:
     def __init__(self, db: PostgresConnector):
@@ -23,21 +23,20 @@ class DatasetManager:
     def get_dataset_info(self, dataset_id: int) -> dict:
         query = "SELECT * FROM datasets WHERE id = %s"
         result = self.db.execute(query, (dataset_id,), fetch=True)
-        return result[0] if result else None
+
+        if not result:
+            raise NonExistentDatasetException(f"Dataset {dataset_id} does not exist")
+
+        return result[0]
     
     def save_dataset_info(self, user_id: int, dataset_name: str, topics: dict) -> int:
-            query = """
+        query = """
-                INSERT INTO datasets (user_id, name, topics)
+            INSERT INTO datasets (user_id, name, topics)
-                VALUES (%s, %s, %s)
+            VALUES (%s, %s, %s)
-                RETURNING id
+            RETURNING id
-            """
+        """
-            result = self.db.execute(query, (user_id, dataset_name, Json(topics)), fetch=True)
+        result = self.db.execute(query, (user_id, dataset_name, Json(topics)), fetch=True)
-            return result[0]["id"] if result else None
+        return result[0]["id"] if result else None
-
-    def get_dataset_content(self, dataset_id: int) -> pd.DataFrame:
-        query = "SELECT * FROM events WHERE dataset_id = %s"
-        result = self.db.execute(query, (dataset_id,), fetch=True)
-        return pd.DataFrame(result)
 
     def save_dataset_content(self, dataset_id: int, event_data: pd.DataFrame):
         if event_data.empty:

server/db/database.py

@@ -27,19 +27,21 @@ class PostgresConnector:
         self.connection.autocommit = False
 
     def execute(self, query, params=None, fetch=False) -> list:
-        with self.connection.cursor(cursor_factory=RealDictCursor) as cursor:
+        try:
-            cursor.execute(query, params)
+            with self.connection.cursor(cursor_factory=RealDictCursor) as cursor:
-            if fetch:
+                cursor.execute(query, params)
-                return cursor.fetchall()
+                result = cursor.fetchall() if fetch else None
             self.connection.commit()
+            return result
+        except Exception:
+            self.connection.rollback()
+            raise
 
     def execute_batch(self, query, values):
         with self.connection.cursor(cursor_factory=RealDictCursor) as cursor:
             execute_batch(cursor, query, values)
             self.connection.commit()
 
-
-    ## User Management Methods
     def close(self):
         if self.connection:
             self.connection.close()

server/exceptions.py

@@ -1,7 +1,7 @@
 class NotAuthorisedException(Exception):
     pass
 
-class NotExistentDatasetException(Exception):
+class NonExistentDatasetException(Exception):
     pass
 
 class DatabaseNotConfiguredException(Exception):