fix(Files): Fix pyodide fstat not returning actual file size but 0 (#2649)

Author: haakonvt

cognite/client/_api/files.py

@@ -3,6 +3,7 @@
 import asyncio
 import copy
 import math
+import os
 import warnings
 from collections import defaultdict
 from collections.abc import AsyncIterator, Sequence
@@ -14,6 +15,7 @@
 
 from cognite.client._api_client import APIClient
 from cognite.client._constants import (
+    _RUNNING_IN_BROWSER,
     DEFAULT_LIMIT_READ,
     FILE_DEFAULT_MULTIPART_SIZE,
     FILE_MAX_MULTIPART_COUNT,
@@ -503,7 +505,7 @@ async def upload_content(
         if not path.is_file():
             raise FileNotFoundError(path)
 
-        file_size = path.stat().st_size
+        file_size = self._get_file_size(path)
         part_size, num_parts = self.calculate_part_size_and_count(file_size)
         session = await self.multipart_upload_content_session(
             parts=num_parts, external_id=external_id, instance_id=instance_id
@@ -641,7 +643,7 @@ async def upload(
     async def _upload_file_from_path(
         self, file_metadata: FileMetadataWrite, path: Path, overwrite: bool
     ) -> FileMetadata:
-        file_size = path.stat().st_size
+        file_size = self._get_file_size(path)
         part_size, num_parts = self.calculate_part_size_and_count(file_size)
         session = await self.multipart_upload_session(
             parts=num_parts,
@@ -674,6 +676,17 @@ async def upload_part(part_no: int) -> None:
         async with session:
             await asyncio.gather(*(upload_part(i) for i in range(num_parts)))
 
+    @staticmethod
+    def _get_file_size(path: Path) -> int:
+        size = path.stat().st_size
+        # Pyodide's File System Access API sometimes lies: stat may report st_size=0 even when there's readable
+        # bytes on disk (that can be read normally...) Fall back to seek/tell, which reads the true size:
+        if _RUNNING_IN_BROWSER and size == 0:
+            with path.open("rb") as fh:
+                fh.seek(0, os.SEEK_END)
+                return fh.tell()
+        return size
+
     def calculate_part_size_and_count(self, file_size: int) -> tuple[int, int]:
         """Calculate part size and count for a multipart upload, for a given file size.
 

cognite/client/_sync_api/files.py

@@ -1,6 +1,6 @@
 """
 ===============================================================================
-1f615115f322d471c0ae15347e13f8c2
+d8fa16955b8c30a8af32a19dcb33c4a8
 This file is auto-generated from the Async API modules, - do not edit manually!
 ===============================================================================
 """

tests/tests_unit/test_api/test_files.py

@@ -6,6 +6,7 @@
 from collections.abc import Callable, Iterator
 from pathlib import Path
 from tempfile import TemporaryDirectory
+from types import SimpleNamespace
 from typing import TYPE_CHECKING, Any, NoReturn
 from unittest.mock import AsyncMock, patch
 
@@ -14,6 +15,7 @@
 from pytest_httpx import HTTPXMock
 
 from cognite.client import CogniteClient
+from cognite.client._api.files import FilesAPI
 from cognite.client._constants import FILE_MAX_MULTIPART_COUNT, FILE_MAX_MULTIPART_SIZE, FILE_MIN_MULTIPART_SIZE
 from cognite.client.config import global_config
 from cognite.client.data_classes import GeoLocation, GeoLocationFilter, Geometry, GeometryFilter, TimestampRange
@@ -1185,6 +1187,46 @@ def test_upload_semaphore_limits_concurrent_parts(
         assert peak == concurrency_limit
 
 
+@pytest.fixture
+def lying_stat(monkeypatch: pytest.MonkeyPatch) -> None:
+    """Monkeypatch Path.stat to report st_size=0, mimicking Pyodide's broken fstat."""
+    monkeypatch.setattr(Path, "stat", lambda _path, **_kw: SimpleNamespace(st_size=0))
+
+
+class TestGetFileSize:
+    def test_browser_falls_back_to_seek_when_stat_lies(
+        self, tmp_path: Path, monkeypatch: pytest.MonkeyPatch, lying_stat: None
+    ) -> None:
+        """Pyodide /drive/... regression: stat reports 0 but the file -has- bytes."""
+        p = tmp_path / "lied_about.bin"
+        p.write_bytes(b"hey\nyo\n")  # 7 bytes really on disk
+        # We have to pretend we are running in the browser with Pyodide:
+        monkeypatch.setattr("cognite.client._api.files._RUNNING_IN_BROWSER", True)
+
+        # Sanity-check that stat is now lying:
+        assert p.stat().st_size == 0
+        # The helper should still report correct size:
+        assert FilesAPI._get_file_size(p) == 7
+
+    def test_empty_file_in_browser_still_reports_zero(self, tmp_path: Path, monkeypatch: pytest.MonkeyPatch) -> None:
+        """A genuinely empty file must still report size=0 in browser mode."""
+        p = tmp_path / "empty.bin"
+        p.touch()
+        monkeypatch.setattr("cognite.client._api.files._RUNNING_IN_BROWSER", True)
+
+        assert FilesAPI._get_file_size(p) == 0
+
+    def test_returns_stat_size_for_non_empty_file(self, tmp_path: Path) -> None:
+        p = tmp_path / "x.bin"
+        p.write_bytes(b"hello world")
+        assert FilesAPI._get_file_size(p) == 11
+
+    def test_returns_zero_for_genuinely_empty_file(self, tmp_path: Path) -> None:
+        p = tmp_path / "empty.bin"
+        p.touch()
+        assert FilesAPI._get_file_size(p) == 0
+
+
 @pytest.fixture
 def mock_files_empty(
     httpx_mock: HTTPXMock, cognite_client: CogniteClient, async_client: AsyncCogniteClient