Polishing

Author: Intrinsical-AI

rre-dataset-generator/src/search_engine/data_store.py

@@ -152,7 +152,7 @@ def default_serializer(obj):
             """Default function to handle non-serializable objects"""
             if isinstance(obj, QueryRatingContext):
                 return obj.to_dict()
-            elif isinstance(obj, Document): # from pydantic import BaseModel
+            elif isinstance(obj, Document): # more generall: from pydantic import BaseModel
                 return obj.model_dump()
             else:
                 # Convert to string as fallback
@@ -172,9 +172,6 @@ def load_tmp_file_content(self) -> None:
         """
         Loads the content from a file on disk deserializing queries-ratings and documents.
         """
-        self._documents.clear()
-        self._queries_by_id.clear()
-        self._query_text_to_query_id.clear()
 
         filepath: Path = self.ensure_path_and_folder_exists()
 
@@ -185,6 +182,10 @@ def load_tmp_file_content(self) -> None:
         with filepath.open("r", encoding="utf-8") as f:
             file_content = json.load(f)
 
+        self._documents.clear()
+        self._queries_by_id.clear()
+        self._query_text_to_query_id.clear()
+
         queries_data: Dict[str, Dict[str, Any]] = file_content.get("queries", {})
         documents_data: Dict[str, Dict[str, Any]] = file_content.get("documents", {})
 

rre-dataset-generator/src/writers/abstract_writer.py

@@ -1,8 +1,6 @@
 from abc import ABC, abstractmethod
 from typing import List, Tuple
 from pathlib import Path
-
-from src.model.query_rating_context import QueryRatingContext
 from src.search_engine.data_store import DataStore