[DH-5512] Fix refresh endpoint and uri validation (#413)

Dataherald · Feb 28, 2024 · ea3e009 · ea3e009
1 parent 8e45fb9
commit ea3e009
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 14 deletions.
diff --git a/dataherald/db_scanner/sqlalchemy.py b/dataherald/db_scanner/sqlalchemy.py
@@ -64,28 +64,23 @@ def refresh_tables(
         metadata: dict = None,
     ) -> list[TableDescription]:
         stored_tables = repository.find_by({"db_connection_id": str(db_connection_id)})
-        stored_tables_list = [table.table_name for table in stored_tables]
 
         rows = []
         for table_description in stored_tables:
             if table_description.table_name not in tables:
                 table_description.status = TableDescriptionStatus.DEPRECATED.value
                 rows.append(repository.save_table_info(table_description))
-            else:
-                rows.append(TableDescription(**table_description.dict()))
-
         for table in tables:
-            if table not in stored_tables_list:
-                rows.append(
-                    repository.save_table_info(
-                        TableDescription(
-                            db_connection_id=db_connection_id,
-                            table_name=table,
-                            status=TableDescriptionStatus.NOT_SCANNED.value,
-                            metadata=metadata,
-                        )
+            rows.append(
+                repository.save_table_info(
+                    TableDescription(
+                        db_connection_id=db_connection_id,
+                        table_name=table,
+                        status=TableDescriptionStatus.NOT_SCANNED.value,
+                        metadata=metadata,
                     )
                 )
+            )
         return rows
 
     @override

diff --git a/dataherald/sql_database/models/types.py b/dataherald/sql_database/models/types.py
@@ -89,7 +89,7 @@ class DatabaseConnection(BaseModel):
 
     @classmethod
     def validate_uri(cls, input_string):
-        pattern = r"([^:/]+)://([^/]+)/([^/]+)"
+        pattern = r"([^:/]+):/+([^/]+)/([^/]+)"
         match = re.match(pattern, input_string)
         if not match:
             raise InvalidURIFormatError(f"Invalid URI format: {input_string}")