fix: Read file-like values before upload in the impit transport

Author: vdusek

src/apify_client/_utils/encoding.py

@@ -22,19 +22,20 @@ def encode_key_value_store_record_value(value: Any, *, content_type: str | None
     Returns:
         A tuple of (encoded_value, content_type).
     """
+    # Read file-like values into memory; the underlying HTTP transport only accepts bytes-like bodies,
+    # so a file object would otherwise reach it unread and raise a raw `TypeError`.
+    if isinstance(value, io.IOBase):
+        value = value.read()
+
     if not content_type:
-        if isinstance(value, (bytes, bytearray, io.IOBase)):
+        if isinstance(value, (bytes, bytearray)):
             content_type = 'application/octet-stream'
         elif isinstance(value, str):
             content_type = 'text/plain; charset=utf-8'
         else:
             content_type = 'application/json; charset=utf-8'
 
-    if (
-        'application/json' in content_type
-        and not isinstance(value, (bytes, bytearray, io.IOBase))
-        and not isinstance(value, str)
-    ):
+    if 'application/json' in content_type and not isinstance(value, (bytes, bytearray, str)):
         # Don't use indentation to reduce size.
         value = json.dumps(
             value,

tests/unit/test_key_value_store.py

@@ -0,0 +1,60 @@
+from __future__ import annotations
+
+import gzip
+import io
+from typing import TYPE_CHECKING
+
+from werkzeug import Request, Response
+
+from apify_client import ApifyClient, ApifyClientAsync
+
+if TYPE_CHECKING:
+    from pytest_httpserver import HTTPServer
+
+_MOCKED_KVS_ID = 'test_kvs_id'
+_RECORD_PATH = f'/v2/key-value-stores/{_MOCKED_KVS_ID}/records/f'
+
+
+def _decode_body(request: Request) -> bytes:
+    raw = request.get_data()
+    return gzip.decompress(raw) if request.headers.get('Content-Encoding') == 'gzip' else raw
+
+
+def test_set_record_reads_file_like_value_sync(httpserver: HTTPServer) -> None:
+    """Regression test: a file-like value is read and its bytes are uploaded, not passed through unread."""
+    captured_requests: list[Request] = []
+
+    def capture_request(request: Request) -> Response:
+        captured_requests.append(request)
+        return Response(status=201)
+
+    httpserver.expect_request(_RECORD_PATH, method='PUT').respond_with_handler(capture_request)
+
+    api_url = httpserver.url_for('/').removesuffix('/')
+    client = ApifyClient(token='test_token', api_url=api_url)
+
+    client.key_value_store(_MOCKED_KVS_ID).set_record('f', io.BytesIO(b'buffer data'))
+
+    assert len(captured_requests) == 1
+    assert _decode_body(captured_requests[0]) == b'buffer data'
+    assert captured_requests[0].headers['content-type'] == 'application/octet-stream'
+
+
+async def test_set_record_reads_file_like_value_async(httpserver: HTTPServer) -> None:
+    """Regression test: a file-like value is read and its bytes are uploaded, not passed through unread."""
+    captured_requests: list[Request] = []
+
+    def capture_request(request: Request) -> Response:
+        captured_requests.append(request)
+        return Response(status=201)
+
+    httpserver.expect_request(_RECORD_PATH, method='PUT').respond_with_handler(capture_request)
+
+    api_url = httpserver.url_for('/').removesuffix('/')
+    client = ApifyClientAsync(token='test_token', api_url=api_url)
+
+    await client.key_value_store(_MOCKED_KVS_ID).set_record('f', io.BytesIO(b'buffer data'))
+
+    assert len(captured_requests) == 1
+    assert _decode_body(captured_requests[0]) == b'buffer data'
+    assert captured_requests[0].headers['content-type'] == 'application/octet-stream'

tests/unit/test_utils.py

@@ -249,13 +249,21 @@ def test_encode_key_value_store_record_value(
 
 
 def test_encode_key_value_store_record_value_bytesio() -> None:
-    """Test that BytesIO is encoded as octet-stream."""
+    """Test that BytesIO is read into bytes and encoded as octet-stream."""
     buffer = io.BytesIO(b'buffer data')
     value, content_type = encode_key_value_store_record_value(buffer)
-    assert value == buffer
+    assert value == b'buffer data'
     assert content_type == 'application/octet-stream'
 
 
+def test_encode_key_value_store_record_value_stringio() -> None:
+    """Test that StringIO is read into text and encoded as text/plain."""
+    buffer = io.StringIO('buffer data')
+    value, content_type = encode_key_value_store_record_value(buffer)
+    assert value == 'buffer data'
+    assert content_type == 'text/plain; charset=utf-8'
+
+
 def test_response_to_dict() -> None:
     """Test parsing response as dictionary."""
     mock_response = Mock()