feat: implement "append_from_file" (#1686)

chandra-siri · web-flow · commit 1333c956da18 · 2025-12-22T18:09:32.000+05:30
feat: implement `append_from_file`
diff --git a/google/cloud/storage/_experimental/asyncio/async_appendable_object_writer.py b/google/cloud/storage/_experimental/asyncio/async_appendable_object_writer.py
@@ -21,6 +21,7 @@
 if you want to use these Rapid Storage APIs.
 
 """
+from io import BufferedReader
 from typing import Optional, Union
 
 from google_crc32c import Checksum
@@ -339,6 +340,16 @@ async def append_from_stream(self, stream_obj):
         """
         raise NotImplementedError("append_from_stream is not implemented yet.")
 
-    async def append_from_file(self, file_path: str):
-        """Create a file object from `file_path` and call append_from_stream(file_obj)"""
-        raise NotImplementedError("append_from_file is not implemented yet.")
+    async def append_from_file(
+        self, file_obj: BufferedReader, block_size: int = _DEFAULT_FLUSH_INTERVAL_BYTES
+    ):
+        """
+        Appends data to an Appendable Object using file_handle which is opened
+        for reading in binary mode.
+
+        :type file_obj: file
+        :param file_obj: A file handle opened in binary mode for reading.
+
+        """
+        while block := file_obj.read(block_size):
+            await self.append(block)
diff --git a/tests/unit/asyncio/test_async_appendable_object_writer.py b/tests/unit/asyncio/test_async_appendable_object_writer.py
@@ -12,6 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from io import BytesIO
 import pytest
 from unittest import mock
 
@@ -23,6 +24,7 @@
 )
 from google.cloud.storage._experimental.asyncio.async_appendable_object_writer import (
     _MAX_CHUNK_SIZE_BYTES,
+    _DEFAULT_FLUSH_INTERVAL_BYTES,
 )
 from google.cloud import _storage_v2
 
@@ -287,9 +289,6 @@ async def test_unimplemented_methods_raise_error(mock_client):
     with pytest.raises(NotImplementedError):
         await writer.append_from_stream(mock.Mock())
 
-    with pytest.raises(NotImplementedError):
-        await writer.append_from_file("file.txt")
-
 
 @pytest.mark.asyncio
 @mock.patch(
@@ -536,9 +535,6 @@ async def test_append_flushes_when_buffer_is_full(
     mock_write_object_stream, mock_client
 ):
     """Test that append flushes the stream when the buffer size is reached."""
-    from google.cloud.storage._experimental.asyncio.async_appendable_object_writer import (
-        _DEFAULT_FLUSH_INTERVAL_BYTES,
-    )
 
     writer = AsyncAppendableObjectWriter(mock_client, BUCKET, OBJECT)
     writer._is_stream_open = True
@@ -559,9 +555,6 @@ async def test_append_flushes_when_buffer_is_full(
 )
 async def test_append_handles_large_data(mock_write_object_stream, mock_client):
     """Test that append handles data larger than the buffer size."""
-    from google.cloud.storage._experimental.asyncio.async_appendable_object_writer import (
-        _DEFAULT_FLUSH_INTERVAL_BYTES,
-    )
 
     writer = AsyncAppendableObjectWriter(mock_client, BUCKET, OBJECT)
     writer._is_stream_open = True
@@ -602,3 +595,32 @@ async def test_append_data_two_times(mock_write_object_stream, mock_client):
     total_data_length = len(data1) + len(data2)
     assert writer.offset == total_data_length
     assert writer.simple_flush.await_count == 0
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize(
+    "file_size, block_size",
+    [
+        (10, 4 * 1024),
+        (0, _DEFAULT_FLUSH_INTERVAL_BYTES),
+        (20 * 1024 * 1024, _DEFAULT_FLUSH_INTERVAL_BYTES),
+        (16 * 1024 * 1024, _DEFAULT_FLUSH_INTERVAL_BYTES),
+    ],
+)
+async def test_append_from_file(file_size, block_size, mock_client):
+    # arrange
+    fp = BytesIO(b"a" * file_size)
+    writer = AsyncAppendableObjectWriter(mock_client, BUCKET, OBJECT)
+    writer._is_stream_open = True
+    writer.append = mock.AsyncMock()
+
+    # act
+    await writer.append_from_file(fp, block_size=block_size)
+
+    # assert
+    exepected_calls = (
+        file_size // block_size
+        if file_size % block_size == 0
+        else file_size // block_size + 1
+    )
+    assert writer.append.await_count == exepected_calls
