Make input dir in DataProcessor required (#18910)

(cherry picked from commit 61a286d)
Lightning-AI · Dec 19, 2023 · 7eabb21 · 7eabb21
1 parent 285e784
commit 7eabb21
Showing 1 changed file with 3 additions and 4 deletions.
diff --git a/src/lightning/data/streaming/data_processor.py b/src/lightning/data/streaming/data_processor.py
@@ -731,7 +731,7 @@ def prepare_item(self, output_dir: str, item_metadata: T) -> None:  # type: igno
 class DataProcessor:
     def __init__(
         self,
-        input_dir: Optional[Union[str, Dir]] = None,
+        input_dir: Union[str, Dir],
         output_dir: Optional[Union[str, Dir]] = None,
         num_workers: Optional[int] = None,
         num_downloaders: Optional[int] = None,
@@ -771,9 +771,8 @@ def __init__(
         self.stop_queues: List[Queue] = []
         self.reorder_files = reorder_files
 
-        if self.input_dir:
-            # Ensure the input dir is the same across all nodes
-            self.input_dir = self._broadcast_object(self.input_dir)
+        # Ensure the input dir is the same across all nodes
+        self.input_dir = self._broadcast_object(self.input_dir)
 
         if self.output_dir:
             # Ensure the output dir is the same across all nodes