fix(core): support string data in Blob.as_bytes_io

libs/core/langchain_core/documents/base.py

@@ -203,6 +203,8 @@ def as_bytes_io(self) -> Generator[BytesIO | BufferedReader, None, None]:
         """
         if isinstance(self.data, bytes):
             yield BytesIO(self.data)
+        elif isinstance(self.data, str):
+            yield BytesIO(self.data.encode(self.encoding))
         elif self.data is None and self.path:
             with Path(self.path).open("rb") as f:
                 yield f

libs/core/tests/unit_tests/documents/test_base.py

@@ -0,0 +1,21 @@
+"""Tests for `langchain_core.documents.base.Blob`."""
+
+from langchain_core.documents.base import Blob
+
+
+def test_as_bytes_io_with_string_data() -> None:
+    blob = Blob.from_data("hello")
+    with blob.as_bytes_io() as stream:
+        assert stream.read() == b"hello"
+
+
+def test_as_bytes_io_with_string_data_respects_encoding() -> None:
+    blob = Blob.from_data("café", encoding="utf-16")
+    with blob.as_bytes_io() as stream:
+        assert stream.read() == "café".encode("utf-16")
+
+
+def test_as_bytes_io_with_bytes_data() -> None:
+    blob = Blob.from_data(b"\x00\x01\x02")
+    with blob.as_bytes_io() as stream:
+        assert stream.read() == b"\x00\x01\x02"