Merge pull request #8637 from kahghi/add-gcs-storage-provider

feat:add GCSStorageProvider
This commit is contained in:
Timothy Jaeryang Baek
2025-01-21 18:14:22 -08:00
committed by GitHub
6 changed files with 1935 additions and 1112 deletions

View File

@@ -1,10 +1,12 @@
import io
import os
import boto3
import pytest
from botocore.exceptions import ClientError
from moto import mock_aws
from open_webui.storage import provider
from gcp_storage_emulator.server import create_server
from google.cloud import storage
def mock_upload_dir(monkeypatch, tmp_path):
@@ -19,6 +21,7 @@ def test_imports():
provider.StorageProvider
provider.LocalStorageProvider
provider.S3StorageProvider
provider.GCSStorageProvider
provider.Storage
@@ -27,6 +30,8 @@ def test_get_storage_provider():
assert isinstance(Storage, provider.LocalStorageProvider)
Storage = provider.get_storage_provider("s3")
assert isinstance(Storage, provider.S3StorageProvider)
Storage = provider.get_storage_provider("gcs")
assert isinstance(Storage, provider.GCSStorageProvider)
with pytest.raises(RuntimeError):
provider.get_storage_provider("invalid")
@@ -42,6 +47,7 @@ def test_class_instantiation():
Test()
provider.LocalStorageProvider()
provider.S3StorageProvider()
provider.GCSStorageProvider()
class TestLocalStorageProvider:
@@ -175,3 +181,91 @@ class TestS3StorageProvider:
self.Storage.delete_all_files()
assert not (upload_dir / self.filename).exists()
assert not (upload_dir / self.filename_extra).exists()
class TestGCSStorageProvider:
Storage = provider.GCSStorageProvider()
Storage.bucket_name = "my-bucket"
file_content = b"test content"
filename = "test.txt"
filename_extra = "test_exyta.txt"
file_bytesio_empty = io.BytesIO()
@pytest.fixture(scope="class")
def setup(self):
host, port = "localhost", 9023
server = create_server(host, port, in_memory=True)
server.start()
os.environ["STORAGE_EMULATOR_HOST"] = f"http://{host}:{port}"
gcs_client = storage.Client()
bucket = gcs_client.bucket(self.Storage.bucket_name)
bucket.create()
self.Storage.gcs_client, self.Storage.bucket = gcs_client, bucket
yield
bucket.delete(force=True)
server.stop()
def test_upload_file(self, monkeypatch, tmp_path, setup):
upload_dir = mock_upload_dir(monkeypatch, tmp_path)
# catch error if bucket does not exist
with pytest.raises(Exception):
self.Storage.bucket = monkeypatch(self.Storage, "bucket", None)
self.Storage.upload_file(io.BytesIO(self.file_content), self.filename)
contents, gcs_file_path = self.Storage.upload_file(
io.BytesIO(self.file_content), self.filename
)
object = self.Storage.bucket.get_blob(self.filename)
assert self.file_content == object.download_as_bytes()
# local checks
assert (upload_dir / self.filename).exists()
assert (upload_dir / self.filename).read_bytes() == self.file_content
assert contents == self.file_content
assert gcs_file_path == "gs://" + self.Storage.bucket_name + "/" + self.filename
# test error if file is empty
with pytest.raises(ValueError):
self.Storage.upload_file(self.file_bytesio_empty, self.filename)
def test_get_file(self, monkeypatch, tmp_path, setup):
upload_dir = mock_upload_dir(monkeypatch, tmp_path)
contents, gcs_file_path = self.Storage.upload_file(
io.BytesIO(self.file_content), self.filename
)
file_path = self.Storage.get_file(gcs_file_path)
assert file_path == str(upload_dir / self.filename)
assert (upload_dir / self.filename).exists()
def test_delete_file(self, monkeypatch, tmp_path, setup):
upload_dir = mock_upload_dir(monkeypatch, tmp_path)
contents, gcs_file_path = self.Storage.upload_file(
io.BytesIO(self.file_content), self.filename
)
# ensure that local directory has the uploaded file as well
assert (upload_dir / self.filename).exists()
assert self.Storage.bucket.get_blob(self.filename).name == self.filename
self.Storage.delete_file(gcs_file_path)
# check that deleting file from gcs will delete the local file as well
assert not (upload_dir / self.filename).exists()
assert self.Storage.bucket.get_blob(self.filename) == None
def test_delete_all_files(self, monkeypatch, tmp_path, setup):
upload_dir = mock_upload_dir(monkeypatch, tmp_path)
# create 2 files
self.Storage.upload_file(io.BytesIO(self.file_content), self.filename)
object = self.Storage.bucket.get_blob(self.filename)
assert (upload_dir / self.filename).exists()
assert (upload_dir / self.filename).read_bytes() == self.file_content
assert self.Storage.bucket.get_blob(self.filename).name == self.filename
assert self.file_content == object.download_as_bytes()
self.Storage.upload_file(io.BytesIO(self.file_content), self.filename_extra)
object = self.Storage.bucket.get_blob(self.filename_extra)
assert (upload_dir / self.filename_extra).exists()
assert (upload_dir / self.filename_extra).read_bytes() == self.file_content
assert self.Storage.bucket.get_blob(self.filename_extra).name == self.filename_extra
assert self.file_content == object.download_as_bytes()
self.Storage.delete_all_files()
assert not (upload_dir / self.filename).exists()
assert not (upload_dir / self.filename_extra).exists()
assert self.Storage.bucket.get_blob(self.filename) == None
assert self.Storage.bucket.get_blob(self.filename_extra) == None