add loading into HF datasets library

Signed-off-by: Michele Dolfi <[email protected]>

examples/export_multimodal.py

@@ -94,6 +94,18 @@ def main():
             f"The example failed converting {failure_count} on {len(input_doc_paths)}."
         )
 
+    # This block demonstrates how the file can be opened with the HF datasets library
+    # from datasets import Dataset
+    # from PIL import Image
+    # multimodal_df = pd.read_parquet(output_filename)
+
+    # # Convert pandas DataFrame to Hugging Face Dataset and load bytes into image
+    # dataset = Dataset.from_pandas(multimodal_df)
+    # def transforms(examples):
+    #     examples["image"] = Image.frombytes('RGB', (examples["image.width"], examples["image.height"]), examples["image.bytes"], 'raw')
+    #     return examples
+    # dataset = dataset.map(transforms)
+
 
 if __name__ == "__main__":
     main()