feat: ensure notebook endpoints do their job (#388)

This work brings the notebook related endpoints to a working state
to serve as replacement for the renku-notebooks external service.

Short summary:
- API tests have been added
- Code has been fixed to answer / handle exception
- Automated test cluster creation has been added using k3d

---------

Co-authored-by: Tasko Olevski <tasko.olevski@sdsc.ethz.ch>

.devcontainer/.poetry_cache/.keep

@@ -1 +0,0 @@
-

.devcontainer/devcontainer.json

@@ -11,7 +11,6 @@
 			"command": "poetry self add poetry-polylith-plugin"
 		},
 		"ghcr.io/devcontainers/features/docker-in-docker:2": {},
-		"ghcr.io/mpriscella/features/kind:1": {},
 		"ghcr.io/devcontainers-contrib/features/gh-release:1": {
 			"repo": "authzed/zed",
 			"binaryNames": "zed"
@@ -28,7 +27,8 @@
 		"ghcr.io/EliiseS/devcontainer-features/bash-profile:1": {
 			"command": "alias k=kubectl"
 		},
-		"ghcr.io/devcontainers-contrib/features/rclone:1": {}
+		"ghcr.io/devcontainers-contrib/features/rclone:1": {},
+		"./k3d": {}
 	},
 	"overrideFeatureInstallOrder": [
 		"ghcr.io/devcontainers-contrib/features/poetry",

.devcontainer/k3d/devcontainer-feature.json

@@ -0,0 +1,17 @@
+{
+    "id": "k3d",
+    "version": "1.0.0",
+    "name": "k3s based kubernetes cluster in docker",
+    "postCreateCommand": "k3d --version",
+    "installsAfter": [
+        "ghcr.io/devcontainers-contrib/features/bash-command"
+    ],
+    "options": {
+    "k3d_version": {
+      "type": "string",
+      "description": "k3d version to install",
+      "proposals": ["latest", "5.7.4"],
+      "default": "latest"
+    }
+  }
+}

.devcontainer/k3d/install.sh

@@ -0,0 +1,14 @@
+if [ "${K3D_VERSION}" != "none" ]; then
+    echo "Downloading k3d..."
+    if [ "${K3D_VERSION}" = "latest" ]; then
+        # Install and check the hash
+        curl -sSL https://raw.githubusercontent.com/k3d-io/k3d/main/install.sh | bash
+    else
+        find_version_from_git_tags K3D_VERSION https://github.com/kubernetes/K3D
+        if [ "${K3D_VERSION::1}" != "v" ]; then
+            K3D_VERSION="v${K3D_VERSION}"
+        fi
+        # Install and check the hash
+        curl -sSL https://raw.githubusercontent.com/k3d-io/k3d/main/install.sh | TAG="${K3D_VERSION}" bash
+    fi
+fi

.github/workflows/test_publish.yml

@@ -95,6 +95,11 @@ jobs:
       - uses: actions/checkout@v4
         with:
           fetch-depth: 0
+      - uses: actions/cache/restore@v3
+        name: Restore cache
+        with:
+          path: ${{ env.CACHE_PATH }}
+          key: ${{ env.CACHE_KEY }}
       - name: Set Git config
         shell: bash
         run: |

.pre-commit-config.yaml

@@ -18,8 +18,10 @@ repos:
       - id: check-toml
       - id: debug-statements
       - id: end-of-file-fixer
+        exclude: 'components/renku_data_services/message_queue/(avro_models|schemas)'
       - id: mixed-line-ending
       - id: trailing-whitespace
+        exclude: 'components/renku_data_services/message_queue/(avro_models|schemas)'
   - repo: https://github.com/asottile/yesqa
     rev: v1.5.0
     hooks:

Makefile

@@ -1,7 +1,7 @@
-.PHONY: schemas tests test_setup main_tests schemathesis_tests collect_coverage style_checks pre_commit_checks run download_avro check_avro avro_models update_avro kind_cluster install_amaltheas all
+.PHONY: schemas tests test_setup main_tests schemathesis_tests collect_coverage style_checks pre_commit_checks run download_avro check_avro avro_models update_avro k3d_cluster install_amaltheas all
 
 AMALTHEA_JS_VERSION ?= 0.12.2
-AMALTHEA_SESSIONS_VERSION ?= 0.0.9-new-operator-chart
+AMALTHEA_SESSIONS_VERSION ?= 0.0.10-new-operator-chart
 codegen_params = --input-file-type openapi --output-model-type pydantic_v2.BaseModel --use-double-quotes --target-python-version 3.12 --collapse-root-models --field-constraints --strict-nullable --set-default-enum-member --openapi-scopes schemas paths parameters --set-default-enum-member --use-one-literal-as-default --use-default
 
 define test_apispec_up_to_date
@@ -151,21 +151,15 @@ help:  ## Display this help.
 
 ##@ Helm/k8s
 
-kind_cluster:  ## Creates a kind cluster for testing
-	kind delete cluster
-	docker network rm -f kind
-	docker network create -d=bridge -o com.docker.network.bridge.enable_ip_masquerade=true -o com.docker.network.driver.mtu=1500 --ipv6=false kind
-	kind create cluster --config kind_config.yaml
-	kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/main/deploy/static/provider/kind/deploy.yaml
-	echo "Waiting for ingress controller to initialize"
-	sleep 15
-	kubectl wait --namespace ingress-nginx --for=condition=ready pod --selector=app.kubernetes.io/component=controller --timeout=90s
+k3d_cluster:  ## Creates a k3d cluster for testing
+	k3d cluster delete
+	k3d cluster create --agents 1 --k3s-arg --disable=metrics-server@server:0
 
 install_amaltheas:  ## Installs both version of amalthea in the. NOTE: It uses the currently active k8s context.
 	helm repo add renku https://swissdatasciencecenter.github.io/helm-charts
 	helm repo update
 	helm upgrade --install amalthea-js renku/amalthea --version $(AMALTHEA_JS_VERSION)
-	helm upgrade --install amalthea-sessions amalthea-sessions-0.0.9-new-operator-chart.tgz --version $(AMALTHEA_SESSIONS_VERSION)
+	helm upgrade --install amalthea-se renku/amalthea-sessions --version ${AMALTHEA_SESSIONS_VERSION}
 
 # TODO: Add the version variables from the top of the file here when the charts are fully published
 amalthea_schema:  ## Updates generates pydantic classes from CRDs

components/renku_data_services/base_api/auth.py

@@ -180,3 +180,30 @@ async def decorated_function(*args: _P.args, **kwargs: _P.kwargs) -> _T:
         return response
 
     return decorated_function
+
+
+def internal_gitlab_authenticate(
+    authenticator: Authenticator,
+) -> Callable[
+    [Callable[Concatenate[Request, APIUser, APIUser, _P], Coroutine[Any, Any, _T]]],
+    Callable[Concatenate[Request, APIUser, _P], Coroutine[Any, Any, _T]],
+]:
+    """Decorator for a Sanic handler that that adds a user for the internal gitlab user."""
+
+    def decorator(
+        f: Callable[Concatenate[Request, APIUser, APIUser, _P], Coroutine[Any, Any, _T]],
+    ) -> Callable[Concatenate[Request, APIUser, _P], Coroutine[Any, Any, _T]]:
+        @wraps(f)
+        async def decorated_function(
+            request: Request,
+            user: APIUser,
+            *args: _P.args,
+            **kwargs: _P.kwargs,
+        ) -> _T:
+            access_token = str(request.headers.get("Gitlab-Access-Token"))
+            internal_gitlab_user = await authenticator.authenticate(access_token, request)
+            return await f(request, user, internal_gitlab_user, *args, **kwargs)
+
+        return decorated_function
+
+    return decorator

components/renku_data_services/notebooks/api.spec.yaml

@@ -454,11 +454,8 @@ components:
             message:
               type: string
               example: "Something went wrong - please try again later"
-          required:
-            - "code"
-            - "message"
-      required:
-        - "error"
+          required: ["code", "message"]
+      required: ["error"]
     Generated:
       properties:
         enabled:
@@ -881,7 +878,7 @@ components:
           type: integer
           description: The size of disk storage for the session, in gigabytes
         resource_class_id:
-          default: 
+          default:
           nullable: true
           type: integer
         cloudstorage:

components/renku_data_services/notebooks/api/amalthea_patches/general.py

@@ -30,7 +30,7 @@ def session_tolerations(server: "UserServer") -> list[dict[str, Any]]:
                     "op": "add",
                     "path": "/statefulset/spec/template/spec/tolerations",
                     "value": default_tolerations
-                    + [i.json_match_expression() for i in server.server_options.tolerations],
+                    + [toleration.json_match_expression() for toleration in server.server_options.tolerations],
                 }
             ],
         }

components/renku_data_services/notebooks/api/amalthea_patches/init_containers.py

@@ -312,9 +312,9 @@ def certificates_container(config: _NotebooksConfig) -> tuple[client.V1Container
         projected=client.V1ProjectedVolumeSource(
             default_mode=440,
             sources=[
-                {"secret": {"name": i.get("secret")}}
-                for i in config.sessions.ca_certs.secrets
-                if isinstance(i, dict) and i.get("secret") is not None
+                {"secret": {"name": secret.get("secret")}}
+                for secret in config.sessions.ca_certs.secrets
+                if isinstance(secret, dict) and secret.get("secret") is not None
             ],
         ),
     )

components/renku_data_services/notebooks/api/classes/k8s_client.py

@@ -79,11 +79,11 @@ async def get_pod_logs(self, name: str, max_log_lines: Optional[int] = None) ->
         """Get the logs of all containers in the session."""
         pod = cast(Pod, await Pod.get(name=name, namespace=self.namespace))
         logs: dict[str, str] = {}
-        containers = [i.name for i in pod.spec.containers] + [i.name for i in pod.spec.initContainers]
+        containers = [container.name for container in pod.spec.containers + pod.spec.get("initContainers", [])]
         for container in containers:
             try:
                 # NOTE: calling pod.logs without a container name set crashes the library
-                clogs: list[str] = [i async for i in pod.logs(container=container, tail_lines=max_log_lines)]
+                clogs: list[str] = [clog async for clog in pod.logs(container=container, tail_lines=max_log_lines)]
             except NotFoundError:
                 raise errors.MissingResourceError(message=f"The session pod {name} does not exist.")
             except ServerError as err:
@@ -243,8 +243,10 @@ async def patch_statefulset_tokens(self, name: str, renku_tokens: RenkuTokens) -
         except NotFoundError:
             return None
 
-        containers: list[V1Container] = [V1Container(**i) for i in sts.spec.template.spec.containers]
-        init_containers: list[V1Container] = [V1Container(**i) for i in sts.spec.template.spec.init_containers]
+        containers: list[V1Container] = [V1Container(**container) for container in sts.spec.template.spec.containers]
+        init_containers: list[V1Container] = [
+            V1Container(**container) for container in sts.spec.template.spec.init_containers
+        ]
 
         git_proxy_container_index, git_proxy_container = next(
             ((i, c) for i, c in enumerate(containers) if c.name == "git-proxy"),
@@ -368,7 +370,7 @@ async def list_servers(self, safe_username: str) -> list[_SessionType]:
             )
             raise JSCacheError(f"The JSCache produced an unexpected status code: {res.status_code}")
 
-        return [self.server_type.model_validate(i) for i in res.json()]
+        return [self.server_type.model_validate(server) for server in res.json()]
 
     async def get_server(self, name: str) -> _SessionType | None:
         """Get a specific jupyter server."""
@@ -441,7 +443,11 @@ async def get_server_logs(
     ) -> dict[str, str]:
         """Get the logs from the server."""
         # NOTE: this get_server ensures the user has access to the server without it you could read someone elses logs
-        _ = await self.get_server(server_name, safe_username)
+        server = await self.get_server(server_name, safe_username)
+        if not server:
+            raise MissingResourceError(
+                f"Cannot find server {server_name} for user " f"{safe_username} to retrieve logs."
+            )
         pod_name = f"{server_name}-0"
         return await self.renku_ns_client.get_pod_logs(pod_name, max_log_lines)
 
@@ -481,9 +487,10 @@ async def delete_server(self, server_name: str, safe_username: str) -> None:
         """Delete the server."""
         server = await self.get_server(server_name, safe_username)
         if not server:
-            return None
-        await self.renku_ns_client.delete_server(server_name)
-        return None
+            raise MissingResourceError(
+                f"Cannot find server {server_name} for user " f"{safe_username} in order to delete it."
+            )
+        return await self.renku_ns_client.delete_server(server_name)
 
     async def patch_tokens(self, server_name: str, renku_tokens: RenkuTokens, gitlab_token: GitlabToken) -> None:
         """Patch the Renku and Gitlab access tokens used in a session."""

components/renku_data_services/notebooks/api/classes/server.py

@@ -7,6 +7,7 @@
 from typing import Any
 from urllib.parse import urljoin, urlparse
 
+from gitlab.v4.objects.projects import Project
 from sanic.log import logger
 
 from renku_data_services.base_models import AnonymousAPIUser, AuthenticatedAPIUser
@@ -22,7 +23,6 @@
 from renku_data_services.notebooks.api.classes.cloud_storage import ICloudStorageRequest
 from renku_data_services.notebooks.api.classes.k8s_client import JupyterServerV1Alpha1Kr8s, K8sClient
 from renku_data_services.notebooks.api.classes.repository import GitProvider, Repository
-from renku_data_services.notebooks.api.classes.user import NotebooksGitlabClient
 from renku_data_services.notebooks.api.schemas.secrets import K8sUserSecrets
 from renku_data_services.notebooks.api.schemas.server_options import ServerOptions
 from renku_data_services.notebooks.config import _NotebooksConfig
@@ -155,7 +155,8 @@ async def start(self) -> JupyterServerV1Alpha1 | None:
                     f"or Docker resources are missing: {', '.join(errors)}"
                 )
             )
-        manifest = JupyterServerV1Alpha1.model_validate(await self._get_session_manifest())
+        session_manifest = await self._get_session_manifest()
+        manifest = JupyterServerV1Alpha1.model_validate(session_manifest)
         return await self._k8s_client.create_server(manifest, self.safe_username)
 
     @staticmethod
@@ -321,7 +322,9 @@ def get_labels(self) -> dict[str, str | None]:
             f"{prefix}commit-sha": None,
             f"{prefix}gitlabProjectId": None,
             f"{prefix}safe-username": self.safe_username,
-            f"{prefix}quota": self.server_options.priority_class,
+            f"{prefix}quota": self.server_options.priority_class
+            if self.server_options.priority_class is not None
+            else "",
             f"{prefix}userId": self._user.id,
         }
         return labels
@@ -378,23 +381,23 @@ def __init__(
         workspace_mount_path: Path,
         work_dir: Path,
         config: _NotebooksConfig,
-        gitlab_client: NotebooksGitlabClient,
+        gitlab_project: Project | None,
         internal_gitlab_user: APIUser,
         using_default_image: bool = False,
         is_image_private: bool = False,
+        **_: dict,
     ):
-        self.gitlab_client = gitlab_client
+        self.gitlab_project = gitlab_project
         self.internal_gitlab_user = internal_gitlab_user
-        gitlab_project_name = f"{namespace}/{project}"
-        gitlab_project = self.gitlab_client.get_renku_project(gitlab_project_name)
+        self.gitlab_project_name = f"{namespace}/{project}"
         single_repository = (
             Repository(
-                url=gitlab_project.http_url_to_repo,
-                dirname=gitlab_project.path,
+                url=self.gitlab_project.http_url_to_repo,
+                dirname=self.gitlab_project.path,
                 branch=branch,
                 commit_sha=commit_sha,
             )
-            if gitlab_project is not None
+            if self.gitlab_project is not None
             else None
         )
 
@@ -422,8 +425,6 @@ def __init__(
         self.commit_sha = commit_sha
         self.notebook = notebook
         self.git_host = urlparse(config.git.url).netloc
-        self.gitlab_project_name = gitlab_project_name
-        self.gitlab_project = gitlab_project
         self.single_repository = single_repository
 
     def _get_start_errors(self) -> list[str]:
@@ -509,6 +510,7 @@ def __init__(
         internal_gitlab_user: APIUser,
         using_default_image: bool = False,
         is_image_private: bool = False,
+        **_: dict,
     ):
         super().__init__(
             user=user,
@@ -531,11 +533,27 @@ def __init__(
         self.project_id = project_id
         self.launcher_id = launcher_id
 
+    def get_labels(self) -> dict[str, str | None]:
+        """Get the labels of the jupyter server."""
+        prefix = self._get_renku_annotation_prefix()
+        labels = super().get_labels()
+
+        # for validation purpose
+        for item in ["commit-sha", "gitlabProjectId"]:
+            labels[f"{prefix}{item}"] = ""
+
+        return labels
+
     def get_annotations(self) -> dict[str, str | None]:
         """Get the annotations of the session."""
         prefix = self._get_renku_annotation_prefix()
         annotations = super().get_annotations()
         annotations[f"{prefix}renkuVersion"] = "2.0"
         annotations[f"{prefix}projectId"] = self.project_id
         annotations[f"{prefix}launcherId"] = self.launcher_id
+
+        # for validation purpose
+        for item in ["commit-sha", "branch", "git-host", "namespace", "projectName", "gitlabProjectId", "repository"]:
+            annotations[f"{prefix}{item}"] = ""
+
         return annotations

components/renku_data_services/notebooks/api/classes/server_manifest.py

@@ -97,7 +97,7 @@ def server_name(self) -> str:
     @property
     def hibernation(self) -> Optional[dict[str, Any]]:
         """Return hibernation annotation."""
-        hibernation = self.manifest.metadata.annotations.get("hibernation")
+        hibernation = self.manifest.metadata.annotations.get("renku.io/hibernation")
         return json.loads(hibernation) if hibernation else None
 
     @property

components/renku_data_services/notebooks/api/schemas/server_options.py

@@ -64,12 +64,12 @@ def __post_init__(self) -> None:
             self.storage = 1
         elif self.storage is None and not self.gigabytes:
             self.storage = 1_000_000_000
-        if not all([isinstance(i, NodeAffinity) for i in self.node_affinities]):
+        if not all([isinstance(affinity, NodeAffinity) for affinity in self.node_affinities]):
             raise ProgrammingError(
                 message="Cannot create a ServerOptions dataclass with node "
                 "affinities that are not of type NodeAffinity"
             )
-        if not all([isinstance(i, Toleration) for i in self.tolerations]):
+        if not all([isinstance(toleration, Toleration) for toleration in self.tolerations]):
             raise ProgrammingError(
                 message="Cannot create a ServerOptions dataclass with tolerations that are not of type Toleration"
             )

components/renku_data_services/notebooks/apispec.py

@@ -1,6 +1,6 @@
 # generated by datamodel-codegen:
 #   filename:  api.spec.yaml
-#   timestamp: 2024-09-23T08:31:51+00:00
+#   timestamp: 2024-09-24T09:26:37+00:00
 
 from __future__ import annotations