From a7a5f49f923c5e3fc8e704e0b4a5398b9e4304e0 Mon Sep 17 00:00:00 2001
From: Kai Carhuallanqui <kai_carhuallanqui@yahoo.com>
Date: Mon, 12 Aug 2024 14:14:56 +0200
Subject: [PATCH] Implemented from_documents() method for IterableDataset

---
 src/pie_datasets/core/dataset.py | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

diff --git a/src/pie_datasets/core/dataset.py b/src/pie_datasets/core/dataset.py
index 0368ba51..6f58864b 100644
--- a/src/pie_datasets/core/dataset.py
+++ b/src/pie_datasets/core/dataset.py
@@ -493,8 +493,19 @@ def from_documents(
         documents: List[Document],
         document_converters: Optional[DocumentConvertersType] = None,
         **dataset_kwargs,
-    ) -> "Dataset":
-        raise NotImplementedError("from_documents is not implemented for IterableDataset")
+    ) -> "IterableDataset":
+        if len(documents) == 0:
+            raise ValueError("No documents to create dataset from")
+        document_type = type(documents[0])
+        data = [doc.asdict() for doc in documents]
+        hf_dataset = datasets.Dataset.from_list(mapping=data, **dataset_kwargs)
+        hf_iterable_dataset = hf_dataset.to_iterable_dataset()
+        dataset = cls.from_hf_dataset(
+            hf_iterable_dataset,
+            document_type=document_type,
+            document_converters=document_converters,
+        )
+        return dataset
 
     def __iter__(self):
         for example in iter(super().__iter__()):