Merge branch 'main' of github.com:raznem/parsera

Author: raznem

parsera/engine/simple_extractor.py

@@ -109,7 +109,7 @@ async def run(self, content: str, attributes: dict[str, str]) -> list[dict]:
 ]
 ```
 
-If no data is found return empty json:
+If no data is found return empty list:
 ```json
 []
 ```

parsera/engine/structured_extractor.py

@@ -8,6 +8,7 @@
 from pydantic import BaseModel, Field, create_model
 
 from parsera.engine.chunks_extractor import ChunksTabularExtractor
+from parsera.utils import has_any_non_none_values
 
 
 class AttributeData(BaseModel):
@@ -67,7 +68,10 @@ async def extract(
         ]
         structured_output = await self.structured_model.ainvoke(messages)
         output_dict = structured_output.model_dump(mode="json")
-        return output_dict["data"]
+        if has_any_non_none_values(output_dict["data"]):
+            return output_dict["data"]
+        else:
+            return []
 
     async def merge_all_data(
         self, all_data: list[list[dict]], attributes: dict[str, str]
@@ -107,7 +111,7 @@ def create_schema(self, attributes: dict[str, dict[str, Any]]) -> Type[BaseModel
         )
 
         class ListSchemaModel(BaseModel):
-            data: List[RecordModel]
+            data: List[RecordModel] = Field(default_factory=list)
 
         return ListSchemaModel
 

parsera/utils.py

@@ -7,3 +7,34 @@ def getinstance(*args, **kwargs):
         return instances[class_]
 
     return getinstance
+
+
+def has_any_non_none_values(input_data: dict | list) -> bool:
+    """
+    Recursively checks if an arbitrarily nested dictionary or list contains any non-None values.
+
+    Args:
+        input_data (dict or list): The dictionary or list to check. Can be nested.
+
+    Returns:
+        bool: True if the data structure (or any nested structure) contains at least one value
+              that is not None, False otherwise (i.e., all values are None or containers
+              that ultimately only contain None).
+    """
+    if isinstance(input_data, dict):
+        for value in input_data.values():
+            if value is not None:
+                if not isinstance(value, (dict, list)):
+                    return True
+                elif has_any_non_none_values(value):
+                    return True
+
+    elif isinstance(input_data, list):
+        for item in input_data:
+            if item is not None:
+                if not isinstance(item, (dict, list)):
+                    return True
+                elif has_any_non_none_values(item):
+                    return True
+
+    return False