Merge pull request #160 from statisticsnorway/restrict-numpy

mallport · web-flow · commit 8fb00dadf4ca · 2024-06-19T19:07:45.000+02:00
Remove Spark workaround. Loosen pyarrow constraints
diff --git a/.vscode/settings.json b/.vscode/settings.json
@@ -1,11 +1,14 @@
 {
+    "python.testing.pytestArgs": [
+        "tests"
+    ],
+    "python.testing.unittestEnabled": false,
+    "python.testing.pytestEnabled": true,
     "python.testing.unittestArgs": [
         "-v",
         "-s",
         "./tests",
         "-p",
         "test*.py"
-    ],
-    "python.testing.pytestEnabled": false,
-    "python.testing.unittestEnabled": true
+    ]
 }
diff --git a/poetry.lock b/poetry.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "dapla-toolbelt"
-version = "2.0.17"
+version = "2.0.18"
 description = "Dapla Toolbelt"
 authors = ["Dapla Developers <dapla-platform-developers@ssb.no>"]
 license = "MIT"
@@ -17,7 +17,7 @@ Changelog = "https://github.com/statisticsnorway/dapla-toolbelt/releases"
 [tool.poetry.dependencies]
 python = ">=3.10,<4.0"
 requests = ">=2.27.1"
-pyarrow = ">=14.0.2, <15"                                   # This tight constraint is dependent on fixing https://issues.apache.org/jira/browse/ARROW-7867
+pyarrow = ">=14.0.2"
 pandas = { version = ">=1.4.2", extras = ["excel", "xml"] }
 gcsfs = ">=2022.7.1"
 ipython = ">=8.10.0"
@@ -78,6 +78,10 @@ warn_unreachable = true
 pretty = true
 show_column_numbers = true
 show_error_context = true
+disallow_untyped_calls = false
+# the above is added due to Google libs being untyped
+# note that this *does* enforce typing functions defined in dapla-toolbelt,
+# but it allows calling untyped functions in a typed context.
 
 [[tool.mypy.overrides]]
 # Allow missing type hints in third-party libraries without type information.
@@ -87,6 +91,7 @@ module = [
     "fsspec.*",
     "responses.*",
     "tomli.*",
+    "google.*",
     "google.cloud.*",
 ]
 ignore_missing_imports = true
diff --git a/src/dapla/auth.py b/src/dapla/auth.py
@@ -94,7 +94,7 @@ def fetch_local_user_from_jupyter() -> dict[str, Any]:
         hub = HubAuth()
         response = requests.get(
             os.environ["LOCAL_USER_PATH"],
-            headers={"Authorization": "token %s" % hub.api_token},
+            headers={"Authorization": f"token {hub.api_token}"},
             cert=(hub.certfile, hub.keyfile),
             verify=hub.client_ca,
             allow_redirects=False,
@@ -229,6 +229,6 @@ def _print_warning(self) -> None:
             display(
                 HTML(
                     'Your session has timed out. Please <a href="/hub/login">log in</a> to continue.'
-                )  # type: ignore [no-untyped-call]
+                )
             )
         )
diff --git a/src/dapla/collector.py b/src/dapla/collector.py
@@ -27,7 +27,7 @@ def start(self, specification: dict[str, Any]) -> Response:
         collector_response = requests.put(
             self.collector_url,
             headers={
-                "Authorization": "Bearer %s" % keycloak_token,
+                "Authorization": f"Bearer {keycloak_token}",
                 "Content-type": "application/json",
             },
             data=json.dumps(specification),
@@ -42,7 +42,7 @@ def running_tasks(self) -> Response:
         """Get all running collector tasks."""
         keycloak_token = AuthClient.fetch_personal_token()
         collector_response = requests.get(
-            self.collector_url, headers={"Authorization": "Bearer %s" % keycloak_token}
+            self.collector_url, headers={"Authorization": f"Bearer {keycloak_token}"}
         )
         collector_response.raise_for_status()
         return collector_response
@@ -59,7 +59,7 @@ def stop(self, task_id: int) -> Response:
         keycloak_token = AuthClient.fetch_personal_token()
         collector_response = requests.delete(
             f"{self.collector_url}/{task_id}",
-            headers={"Authorization": "Bearer %s" % keycloak_token},
+            headers={"Authorization": f"Bearer {keycloak_token}"},
         )
         if collector_response.status_code == 400:
             print(
diff --git a/src/dapla/converter.py b/src/dapla/converter.py
@@ -27,7 +27,7 @@ def start(self, job_config: dict[str, Any]) -> Response:
         converter_response = requests.post(
             f"{self.converter_url}/jobs",
             headers={
-                "Authorization": "Bearer %s" % keycloak_token,
+                "Authorization": f"Bearer {keycloak_token}",
                 "Content-type": "application/json",
             },
             data=json.dumps(job_config),
@@ -50,7 +50,7 @@ def start_simulation(self, job_config: dict[str, Any]) -> Response:
         converter_response = requests.post(
             f"{self.converter_url}/jobs/simulation",
             headers={
-                "Authorization": "Bearer %s" % keycloak_token,
+                "Authorization": f"Bearer {keycloak_token}",
                 "Content-type": "application/json",
             },
             data=json.dumps(job_config),
@@ -72,7 +72,7 @@ def get_job_summary(self, job_id: str) -> Response:
         job_summary = requests.get(
             f"{self.converter_url}/jobs/{job_id}/execution-summary",
             headers={
-                "Authorization": "Bearer %s" % keycloak_token,
+                "Authorization": f"Bearer {keycloak_token}",
                 "Content-type": "application/json",
             },
         )
@@ -93,7 +93,7 @@ def stop_job(self, job_id: str) -> Response:
         job_status = requests.post(
             f"{self.converter_url}/jobs/{job_id}/stop",
             headers={
-                "Authorization": "Bearer %s" % keycloak_token,
+                "Authorization": f"Bearer {keycloak_token}",
                 "Content-type": "application/json",
             },
         )
@@ -122,7 +122,7 @@ def get_pseudo_report(self, job_id: str) -> Response:
         pseudo_report = requests.get(
             f"{self.converter_url}/jobs/{job_id}/reports/pseudo",
             headers={
-                "Authorization": "Bearer %s" % keycloak_token,
+                "Authorization": f"Bearer {keycloak_token}",
                 "Content-type": "application/json",
             },
         )
@@ -145,7 +145,7 @@ def get_pseudo_schema(self, job_id: str) -> Response:
         pseudo_report = requests.get(
             f"{self.converter_url}/jobs/{job_id}/reports/pseudo-schema-hierarchy",
             headers={
-                "Authorization": "Bearer %s" % keycloak_token,
+                "Authorization": f"Bearer {keycloak_token}",
                 "Content-type": "application/json",
             },
         )
diff --git a/src/dapla/doctor.py b/src/dapla/doctor.py
@@ -68,11 +68,11 @@ def gcs_credentials_valid() -> bool:
         print("Checking your Google Cloud Storage credentials...")
 
         # Fetch the google token
-        google_token, _ = AuthClient.fetch_google_credentials().token
+        google_token = AuthClient.fetch_google_credentials().token
 
         try:
             requests.get(
-                "https://oauth2.googleapis.com/tokeninfo?access_token=%s" % google_token
+                f"https://oauth2.googleapis.com/tokeninfo?access_token={google_token}"
             )
         except HttpError as ex:
             if str(ex) == "Invalid Credentials, 401":
diff --git a/src/dapla/gcs.py b/src/dapla/gcs.py
@@ -14,50 +14,8 @@ def __init__(
     ) -> None:
         """Initialize GCSFileSystem."""
         super().__init__(token=token, **kwargs)
-        # Temporary bug fix for https://issues.apache.org/jira/browse/ARROW-7867
-        # Spark writes an empty file to GCS (to mimic a folder structure) before writing partitioned data
-        # Resolve this by ignoring the "empty" file when reading partitioned parquet files
-        try:
-            # Constant is moved to core module in Pyarrow 10.0.0
-            from pyarrow.parquet.core import (  # type: ignore [attr-defined]
-                EXCLUDED_PARQUET_PATHS,
-            )
-        except ImportError:
-            # Fallback for Pyarrow versions <10.0.0
-            from pyarrow.parquet.core import (  # type: ignore [attr-defined]
-                EXCLUDED_PARQUET_PATHS,
-            )
-        from pyarrow.parquet import ParquetManifest
-
-        EXCLUDED_PARQUET_PATHS.add("")
-        ParquetManifest._should_silently_exclude = (  # type: ignore [attr-defined]
-            GCSFileSystem._should_silently_exclude
-        )
 
     def isdir(self, path: str) -> bool:
         """Check if path is a directory."""
         info = super(gcsfs.GCSFileSystem, self).info(path)
         return t.cast(bool, info["type"] == "directory")
-
-    @staticmethod
-    # This code is from from pyarrow.parquet.core
-    def _should_silently_exclude(file_name: str) -> bool:
-        try:
-            # Constant is moved to core module in Pyarrow 10.0.0
-            from pyarrow.parquet.core import (  # type: ignore [attr-defined]
-                EXCLUDED_PARQUET_PATHS,
-            )
-        except ImportError:
-            # Fallback for Pyarrow versions <10.0.0
-            from pyarrow.parquet.core import (  # type: ignore [attr-defined]
-                EXCLUDED_PARQUET_PATHS,
-            )
-
-        return (
-            file_name.endswith(".crc")
-            or file_name.endswith("_$folder$")  # Checksums
-            or file_name.startswith(".")  # HDFS directories in S3
-            or file_name.startswith("_")  # Hidden files starting with .
-            or ".tmp" in file_name  # Hidden files starting with _
-            or file_name in EXCLUDED_PARQUET_PATHS  # Temp files
-        )
diff --git a/src/dapla/guardian.py b/src/dapla/guardian.py
@@ -42,7 +42,7 @@ def call_api(
         api_response = requests.get(
             api_endpoint_url,
             headers={
-                "Authorization": "Bearer %s" % maskinporten_token,
+                "Authorization": f"Bearer {maskinporten_token}",
                 "Accept": "application/json",
             },
         )
@@ -74,7 +74,7 @@ def get_guardian_token(
         guardian_response = requests.post(
             guardian_endpoint,
             headers={
-                "Authorization": "Bearer %s" % keycloak_token,
+                "Authorization": f"Bearer {keycloak_token}",
                 "Content-type": "application/json",
             },
             json=body,
diff --git a/src/dapla/jupyterhub.py b/src/dapla/jupyterhub.py
@@ -23,7 +23,7 @@ def generate_api_token(
         + os.environ["JUPYTERHUB_USER"]
         + "/tokens",
         json=body,
-        headers={"Authorization": "token %s" % hub.api_token},
+        headers={"Authorization": f"token {hub.api_token}"},
         allow_redirects=False,
     )
     hub_response.raise_for_status()
diff --git a/tests/test_auth.py b/tests/test_auth.py
@@ -170,12 +170,11 @@ def test_fetch_google_credentials_expired(
     client = AuthClient()
     credentials = client.fetch_google_credentials()
 
-    assert credentials.expired
-
     fetch_google_token_from_oidc_exchange_mock.return_value = (
         "google_token",
         datetime.now() + timedelta(hours=1),
     )
+
     credentials.refresh(None)
     assert not credentials.expired
 
diff --git a/tests/test_collector.py b/tests/test_collector.py
@@ -17,12 +17,9 @@ def test_initiate_201_response(auth_client_mock: Mock) -> None:
     auth_client_mock.fetch_personal_token.return_value = fake_token
     specification: dict[str, str] = {}
     worker_id = "abcd"
-    collector_response = (
-        """{
-        "workerId": "%s"
-    }"""
-        % worker_id
-    )
+    collector_response = f"""{{
+        "workerId": "{worker_id}"
+    }}"""
     responses.add(
         responses.PUT, collector_test_url, json=collector_response, status=201
     )
diff --git a/tests/test_gcs.py b/tests/test_gcs.py
@@ -28,11 +28,11 @@ def test_gcs_deadlock(mock_fetch_google_token: Mock, mock_is_ready: Mock) -> Non
 
     mock_is_ready.return_value = True  # Mock client ready to not use ADC
     mock_fetch_google_token.side_effect = [
-        ("FakeToken1", utcnow()),  # type: ignore[no-untyped-call]
-        ("FakeToken2", utcnow()),  # type: ignore[no-untyped-call]
-        ("FakeToken3", utcnow()),  # type: ignore[no-untyped-call]
-        ("FakeToken4", utcnow()),  # type: ignore[no-untyped-call]
-        ("FakeToken5Valid", utcnow() + timedelta(seconds=30)),  # type: ignore[no-untyped-call]
+        ("FakeToken1", utcnow()),
+        ("FakeToken2", utcnow()),
+        ("FakeToken3", utcnow()),
+        ("FakeToken4", utcnow()),
+        ("FakeToken5Valid", utcnow() + timedelta(seconds=30)),
     ]
 
     gcs_path = "gs://ssb-dapla-pseudo-data-produkt-test/integration_tests_data/personer.parquet"

Original file line number	Diff line number	Diff line change
`@@ -23,7 +23,7 @@ def generate_api_token(`
`23`	`23`	`+ os.environ["JUPYTERHUB_USER"]`
`24`	`24`	`+ "/tokens",`
`25`	`25`	`json=body,`
`26`		`- headers={"Authorization": "token %s" % hub.api_token},`
	`26`	`+ headers={"Authorization": f"token {hub.api_token}"},`
`27`	`27`	`allow_redirects=False,`
`28`	`28`	`)`
`29`	`29`	`hub_response.raise_for_status()`