fixes with openml.config.[server|apikey] leakage

Author: geetu040

openml/_config.py

@@ -50,12 +50,12 @@
     },
     "local": {
         APIVersion.V1: {
-            "server": "http://localhost:8000/api/v1/xml/",
+            "server": "http://localhost:8080/api/v1/xml/",
             "apikey": "normaluser",
         },
         APIVersion.V2: {
-            "server": "http://localhost:8002/api/v1/xml/",
-            "apikey": "normaluser",
+            "server": "http://localhost:8082/",
+            "apikey": "AD000000000000000000000000000000",
         },
     },
 }
@@ -137,15 +137,10 @@ def __init__(self) -> None:
         self.console_handler: logging.StreamHandler | None = None
         self.file_handler: logging.handlers.RotatingFileHandler | None = None
 
-        server_test_v1_apikey = self.get_servers("test")[APIVersion.V1]["apikey"]
-        server_test_v1_server = self.get_servers("test")[APIVersion.V1]["server"]
-
         self.OPENML_CACHE_DIR_ENV_VAR = "OPENML_CACHE_DIR"
         self.OPENML_SKIP_PARQUET_ENV_VAR = "OPENML_SKIP_PARQUET"
-        self._TEST_SERVER_NORMAL_USER_KEY = server_test_v1_apikey
-        self._HEADERS: dict[str, str] = {"user-agent": f"openml-python/{__version__}"}
         self.OPENML_TEST_SERVER_ADMIN_KEY_ENV_VAR = "OPENML_TEST_SERVER_ADMIN_KEY"
-        self.TEST_SERVER_URL = cast("str", server_test_v1_server).split("/api/v1/xml")[0]
+        self._HEADERS: dict[str, str] = {"user-agent": f"openml-python/{__version__}"}
 
         self._config: OpenMLConfig = OpenMLConfig()
         # for legacy test `test_non_writable_home`
@@ -178,7 +173,6 @@ def __setattr__(self, name: str, value: Any) -> None:
             "_examples",
             "OPENML_CACHE_DIR_ENV_VAR",
             "OPENML_SKIP_PARQUET_ENV_VAR",
-            "_TEST_SERVER_NORMAL_USER_KEY",
             "_HEADERS",
         }:
             return object.__setattr__(self, name, value)

openml/cli.py

@@ -8,10 +8,12 @@
 from collections.abc import Callable
 from dataclasses import fields
 from pathlib import Path
+from typing import cast
 from urllib.parse import urlparse
 
 import openml
 from openml.__version__ import __version__
+from openml.enums import APIVersion
 
 
 def is_hex(string_: str) -> bool:
@@ -110,9 +112,9 @@ def check_server(server: str) -> str:
 
     def replace_shorthand(server: str) -> str:
         if server == "test":
-            return f"{openml.config.TEST_SERVER_URL}/api/v1/xml"
+            return cast("str", openml.config.get_servers("test")[APIVersion.V1]["server"])
         if server == "production_server":
-            return "https://www.openml.org/api/v1/xml"
+            return cast("str", openml.config.get_servers("production")[APIVersion.V1]["server"])
         return server
 
     configure_field(

openml/testing.py

@@ -49,9 +49,7 @@ class TestBase(unittest.TestCase):
         "user": [],
     }
     flow_name_tracker: ClassVar[list[str]] = []
-    test_server = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/"
     admin_key = os.environ.get(openml.config.OPENML_TEST_SERVER_ADMIN_KEY_ENV_VAR)
-    user_key = openml.config._TEST_SERVER_NORMAL_USER_KEY
 
     # creating logger for tracking files uploaded to test server
     logger = logging.getLogger("unit_tests_published_entities")
@@ -106,8 +104,6 @@ def setUp(self, n_levels: int = 1, tmpdir_suffix: str = "") -> None:
         os.chdir(self.workdir)
 
         self.cached = True
-        openml.config.apikey = TestBase.user_key
-        self.production_server = "https://www.openml.org/api/v1/xml"
         openml.config.set_root_cache_directory(str(self.workdir))
 
         # Increase the number of retries to avoid spurious server failures
@@ -128,8 +124,7 @@ def use_production_server(self) -> None:
 
         Please use this sparingly - it is better to use the test server.
         """
-        openml.config.server = self.production_server
-        openml.config.apikey = ""
+        openml.config.set_servers("production")
 
     def tearDown(self) -> None:
         """Tear down the test"""

tests/conftest.py

@@ -99,8 +99,7 @@ def delete_remote_files(tracker, flow_names) -> None:
     :param tracker: Dict
     :return: None
     """
-    openml.config.server = TestBase.test_server
-    openml.config.apikey = TestBase.user_key
+    openml.config.set_servers("test")
 
     # reordering to delete sub flows at the end of flows
     # sub-flows have shorter names, hence, sorting by descending order of flow name length
@@ -252,8 +251,23 @@ def test_files_directory() -> Path:
 
 
 @pytest.fixture(scope="session")
-def test_api_key() -> str:
-    return TestBase.user_key
+def test_server_v1() -> str:
+    return openml.config.get_servers("test")[APIVersion.V1]["server"]
+
+
+@pytest.fixture(scope="session")
+def test_apikey_v1() -> str:
+    return openml.config.get_servers("test")[APIVersion.V1]["apikey"]
+
+
+@pytest.fixture(scope="session")
+def test_server_v2() -> str:
+    return openml.config.get_servers("test")[APIVersion.V2]["server"]
+
+
+@pytest.fixture(scope="session")
+def test_apikey_v2() -> str:
+    return openml.config.get_servers("test")[APIVersion.V2]["apikey"]
 
 
 @pytest.fixture(autouse=True, scope="function")
@@ -275,12 +289,10 @@ def as_robot() -> Iterator[None]:
 @pytest.fixture(autouse=True)
 def with_server(request):
     if "production_server" in request.keywords:
-        openml.config.server = "https://www.openml.org/api/v1/xml/"
-        openml.config.apikey = None
+        openml.config.set_servers("production")
         yield
         return
-    openml.config.server = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/"
-    openml.config.apikey = TestBase.user_key
+    openml.config.set_servers("test")
     yield
 
 

tests/test_api/test_http.py

@@ -26,28 +26,28 @@ def sample_path() -> str:
 
 
 @pytest.fixture
-def sample_url(sample_path) -> str:
-    return urljoin(openml.config.server, sample_path)
+def sample_url_v1(sample_path, test_server_v1) -> str:
+    return urljoin(test_server_v1, sample_path)
 
 
 @pytest.fixture
-def sample_download_url() -> str:
-    server = openml.config.server.split("api/")[0]
+def sample_download_url_v1(test_server_v1) -> str:
+    server = test_server_v1.split("api/")[0]
     endpoint = "data/v1/download/1/anneal.arff"
     url = server + endpoint
     return url
 
 
-def test_cache(cache, sample_url):
+def test_cache(cache, sample_url_v1):
     params = {"param1": "value1", "param2": "value2"}
 
-    parsed_url = urlparse(sample_url)
+    parsed_url = urlparse(sample_url_v1)
     netloc_parts = parsed_url.netloc.split(".")[::-1]
     path_parts = parsed_url.path.strip("/").split("/")
     params_key = "&".join([f"{k}={v}" for k, v in params.items()])
 
 
-    key = cache.get_key(sample_url, params)
+    key = cache.get_key(sample_url_v1, params)
 
     expected_key = os.path.join(
         *netloc_parts,
@@ -58,10 +58,10 @@ def test_cache(cache, sample_url):
     assert key == expected_key
 
     # mock response
-    req = Request("GET", sample_url).prepare()
+    req = Request("GET", sample_url_v1).prepare()
     response = Response()
     response.status_code = 200
-    response.url = sample_url
+    response.url = sample_url_v1
     response.reason = "OK"
     response._content = b"<xml>test</xml>"
     response.headers = {"Content-Type": "text/xml"}
@@ -73,7 +73,7 @@ def test_cache(cache, sample_url):
     cached = cache.load(key)
 
     assert cached.status_code == 200
-    assert cached.url == sample_url
+    assert cached.url == sample_url_v1
     assert cached.content == b"<xml>test</xml>"
     assert cached.headers["Content-Type"] == "text/xml"
 
@@ -87,13 +87,13 @@ def test_get(http_client):
 
 
 @pytest.mark.uses_test_server()
-def test_get_with_cache_creates_cache(http_client, cache, sample_url, sample_path):
+def test_get_with_cache_creates_cache(http_client, cache, sample_url_v1, sample_path):
     response = http_client.get(sample_path, enable_cache=True)
 
     assert response.status_code == 200
     assert cache.path.exists()
 
-    cache_key = cache.get_key(sample_url, {})
+    cache_key = cache.get_key(sample_url_v1, {})
     cache_path = cache._key_to_path(cache_key)
 
     assert (cache_path / "meta.json").exists()
@@ -102,8 +102,8 @@ def test_get_with_cache_creates_cache(http_client, cache, sample_url, sample_pat
 
 
 @pytest.mark.uses_test_server()
-def test_get_uses_cached_response(http_client, cache, sample_url, sample_path):
-    key = cache.get_key(sample_url, {})
+def test_get_uses_cached_response(http_client, cache, sample_url_v1, sample_path):
+    key = cache.get_key(sample_url_v1, {})
     meta_path = cache._key_to_path(key) / "meta.json"
 
     r1 = http_client.get(sample_path, enable_cache=True)
@@ -118,8 +118,8 @@ def test_get_uses_cached_response(http_client, cache, sample_url, sample_path):
 
 
 @pytest.mark.uses_test_server()
-def test_get_refresh_cache(http_client, cache, sample_url, sample_path):
-    key = cache.get_key(sample_url, {})
+def test_get_refresh_cache(http_client, cache, sample_url_v1, sample_path):
+    key = cache.get_key(sample_url_v1, {})
     meta_path = cache._key_to_path(key) / "meta.json"
 
     r1 = http_client.get(sample_path, enable_cache=True)
@@ -148,9 +148,9 @@ def test_get_without_api_key_raises(http_client):
 
 
 @pytest.mark.uses_test_server()
-def test_download_creates_file(http_client, sample_download_url):
+def test_download_creates_file(http_client, sample_download_url_v1):
     path = http_client.download(
-        url=sample_download_url,
+        url=sample_download_url_v1,
         file_name="downloaded.bin",
     )
 
@@ -160,15 +160,15 @@ def test_download_creates_file(http_client, sample_download_url):
 
 
 @pytest.mark.uses_test_server()
-def test_download_is_cached_on_disk(http_client, sample_download_url):
+def test_download_is_cached_on_disk(http_client, sample_download_url_v1):
     path1 = http_client.download(
-        url=sample_download_url,
+        url=sample_download_url_v1,
         file_name="cached.bin",
     )
     mtime1 = path1.stat().st_mtime
 
     path2 = http_client.download(
-        url=sample_download_url,
+        url=sample_download_url_v1,
         file_name="cached.bin",
     )
     mtime2 = path2.stat().st_mtime
@@ -178,13 +178,13 @@ def test_download_is_cached_on_disk(http_client, sample_download_url):
 
 
 @pytest.mark.uses_test_server()
-def test_download_respects_custom_handler(http_client, sample_download_url):
+def test_download_respects_custom_handler(http_client, sample_download_url_v1):
     def handler(response, path: Path, encoding: str):
         path.write_text("HANDLED", encoding=encoding)
         return path
 
     path = http_client.download(
-        url=sample_download_url,
+        url=sample_download_url_v1,
         file_name="handler.bin",
         handler=handler,
     )
@@ -193,7 +193,7 @@ def handler(response, path: Path, encoding: str):
     assert path.read_text() == "HANDLED"
 
 
-def test_post(http_client):
+def test_post(http_client, test_server_v1, test_apikey_v1):
     resource_name = "resource"
     resource_files = {"description": "Resource Description File"}
 
@@ -205,15 +205,15 @@ def test_post(http_client):
 
         mock_request.assert_called_once_with(
             method="POST",
-            url=urljoin(openml.config.server, resource_name),
+            url=urljoin(test_server_v1, resource_name),
             params={},
-            data={"api_key": openml.config.apikey},
+            data={"api_key": test_apikey_v1},
             headers=openml.config._HEADERS,
             files=resource_files,
         )
 
 
-def test_delete(http_client):
+def test_delete(http_client, test_server_v1, test_apikey_v1):
     resource_name = "resource"
     resource_id = 123
 
@@ -226,12 +226,12 @@ def test_delete(http_client):
         mock_request.assert_called_once_with(
             method="DELETE",
             url=(
-                openml.config.server
+                test_server_v1
                 + resource_name
                 + "/"
                 + str(resource_id)
             ),
-            params={"api_key": openml.config.apikey},
+            params={"api_key": test_apikey_v1},
             data={},
             headers=openml.config._HEADERS,
             files=None,