Skip to content

SINTEF/entities-service

Repository files navigation

Entities Service

REST API service for serving entities

This is a FastAPI-based REST API service running on onto-ns.com. It's purpose is to serve entities from an underlying database.

Other than the REST API service, the repository also contains a CLI for validating and uploading entities to the service, as well as create and manipulate the service' configuration file. See the CLI documentation for more information.

The repository also contains a pre-commit hook validate-entities, which may be used externally to validate entities before committing them a repository. See the pre-commit hook documentation for more information.

Install the service

First, download and install the Python package from GitHub:

# Download (git clone)
git clone https://github.com/SINTEF/entities-service.git
cd entities-service

# Install (using pip)
python -m pip install -U pip
pip install -U -e .

Important: If using this service locally alongside DLite, it is important to note that issues may occur if NumPy v2 is used. There is no known issues with NumPy v1. This is a registered issue found for DLite v0.5.16.

Run the service

The service requires a MongoDB server to be running, and the service needs to be able to connect to it. The service also requires a valid X.509 certificate, in order to connect to the MongoDB server.

The MongoDB server could be MongoDB Atlas, a local MongoDB server, or a Docker container running MongoDB.

Using the local environment and MongoDB Atlas

First, create a MongoDB Atlas cluster, and a user with read-only access to the entities database.

Set the necessary environment variables:

ENTITIES_SERVICE_MONGO_URI=<your MongoDB Atlas URI>
ENTITIES_SERVICE_X509_CERTIFICATE_FILE=<your X.509 certificate file>
ENTITIES_SERVICE_MONGO_USER=<your MongoDB Atlas user with read-only access (default: 'guest')>
ENTITIES_SERVICE_MONGO_PASSWORD=<your MongoDB Atlas user's password with read-only access (default: 'guest')>

Run the service:

uvicorn entities_service.main:APP --host localhost --port 8000 --no-server-header --header "Server:EntitiesService"

Finally, go to localhost:8000/docs and try out retrieving an entity.

--log-level debug can be added to the uvicorn command to get more verbose logging. --reload can be added to the uvicorn command to enable auto-reloading of the service when any files are changed.

Note, the environment variables can be set in a .env file, see the section on using a file for environment variables.

Using Docker and a local MongoDB server

First, we need to create self-signed certificates for the service to use. This is done by running the following command:

mkdir docker_security
cd docker_security
../.github/docker_init/setup_mongo_security.sh

Note, this is only possible with openssl installed on your system. And the OS on the system being Linux/Unix-based.

For development, start a local MongoDB server, e.g., through another Docker image:

docker run --rm -d \
  --env "IN_DOCKER=true" \
  --env "HOST_USER=${USER}" \
  --env "MONGO_INITDB_ROOT_USERNAME=root" \
  --env "MONGO_INITDB_ROOT_PASSWORD=root" \
  --name "mongodb" \
  -p "27017:27017" \
  -v "${PWD}/.github/docker_init/create_x509_user.js:/docker-entrypoint-initdb.d/0_create_x509_user.js" \
  -v "${PWD}/docker_security:/mongo_tls" \
  mongo:8 \
  --tlsMode allowTLS --tlsCertificateKeyFile /mongo_tls/test-server1.pem --tlsCAFile /mongo_tls/

Then build and run the Entities Service Docker image:

docker build --pull -t entities-service --target development .
docker run --rm -d \
  --env "ENTITIES_SERVICE_MONGO_URI=mongodb://localhost:27017" \
  --env "ENTITIES_SERVICE_X509_CERTIFICATE_FILE=docker_security/test-client.pem" \
  --env "ENTITIES_SERVICE_CA_FILE=docker_security/test-ca.pem" \
  --name "entities-service" \
  -u "${id -ur}:${id -gr}" \
  -p "8000:80" \
  entities-service

Now, fill up the MongoDB with valid entities at the entities_service database in the entities collection.

Then go to localhost:8000/docs and try out retrieving an entity.


For production, use a public MongoDB, and follow the same instructions above for building and running the Entities Service Docker image, but exchange the --target value with production, put in the proper value for the ENTITIES_SERVICE_MONGO_URI and ENTITIES_SERVICE_X509_CERTIFICATE_FILE environment values, possibly add the ENTITIES_SERVICE_MONGO_USER, ENTITIES_SERVICE_MONGO_PASSWORD, and ENTITIES_SERVICE_CA_FILE environment variables as well, if needed.

Using Docker Compose

Run the following commands:

docker compose pull
docker compose --env-file=.env up --build

By default the development target will be built, to change this, set the ENTITIES_SERVICE_DOCKER_TARGET environment variable accordingly, e.g.:

ENTITIES_SERVICE_DOCKER_TARGET=production docker compose --env-file=.env up --build

Furthermore, the used localhost port can be changed via the PORT environment variable.

The --env-file argument is optional, but if used, it should point to a file containing the environment variables needed by the service. See the section on using a file for environment variables for more information.

Using a file for environment variables

The service supports a "dot-env" file, i.e., a .env file with a list of (secret) environment variables.

In order to use this, create a new file named .env. This file will never be committed if you choose to git commit any files, as it has been hardcoded into the .gitignore file.

Fill up the .env file with (secret) environment variables.

For using it locally, no changes are needed, as the service will automatically check for a .env file and load it in, using it to set the service app configuration.

For using it with Docker, use the --env-file .env argument when calling docker run or docker compose up.

CLI

The CLI is a command-line interface for interacting with the Entities Service. It can be used to validate and upload entities to the service, as well as create and manipulate the service' configuration file.

To see the available commands and options, run:

entities-service --help

A complete list of commands and options can be found in the CLI documentation.

pre-commit hook validate-entities

The validate-entities pre-commit hook runs the CLI command entities-service validate on all files that are about to be committed. This is to ensure that all entities are valid before committing them to the repository.

By default it runs with the --verbose flag, which will print out detailed differences if an entity already exists externally and differs in its content. Furthermore, it will run such that all supported file formats (currently JSON and YAML/YML) will be validated.

Important: Add the .[cli] entry to the additional_dependencies argument.

It is also advisable to focus in which directories or files the hook should run for, by adding the files argument.

Here is an example of how to add the validate-entities pre-commit hook to your .pre-commit-config.yaml file, given a repository that contains entities in the entities directory in the root of the repository:

repos:
# ...
- repo: https://github.com/SINTEF/entities-service
  rev: v0.7.1
  hooks:
  - id: validate-entities
    additional_dependencies: [".[cli]"]
    files: ^entities/.*\.(json|yaml|yml)$

This will run for all JSON, YAML, and YML files in the entities directory and its subdirectories.

Note, you can add the --no-external-calls argument if you wish to not make external calls to the Entities Service when validating entities. This is useful when running the pre-commit hook in an environment where the Entities Service is not available, or when you wish to only validate the entities locally.

# ...
    args: ['--no-external-calls']
# ...

Testing

The repository code is tested using pytest. For the service, it can be tested against a local MongoDB server and Entities Service instance or against a mock MongoDB server and Entities Service instance utilizing Starlette's TestClient.

To run the tests, first install the test dependencies:

pip install -U -e .[testing]

Then run the tests (for mock MongoDB (mongomock) and Entities Service):

pytest

To run the tests against a live backend, you can pull, build, and run the Docker Compose file:

docker compose pull
docker compose build

Before running the services the self-signed certificates need to be created. See the section on using Docker and a local MongoDB server for more information.

Then run (up) the Docker Compose file and subsequently the tests:

docker compose up -d
pytest --live-backend

Remember to set the ENTITIES_SERVICE_X509_CERTIFICATE_FILE and ENTITIES_SERVICE_CA_FILE environment variables to docker_security/test-server1.pem and docker_security/test-ca.pem, respectively. Note, these environment variables are already specified in the docker-compose.yml file, however, one should still check that they are set correctly.

Test uploading entities using the CLI

To test uploading entities using the CLI, one must note that validation of the entities happens twice: First by the CLI, and then by the service. The validation that is most tricky when testing locally is the namespace validation, as the service will validate the namespace against the ENTITIES_SERVICE_BASE_URL environment variable set when starting the service, which defaults to http://onto-ns.com/meta. However, if using this namespace in the CLI, the CLI will connect to the publicly running service at http://onto-ns.com/meta, which will not work when testing locally.

So to make all this work together, one should start the service with the ENTITIES_SERVICE_BASE_URL environment variable set to http://localhost:8000 (which is done through the locally available environment variable ENTITIES_SERVICE_HOST), and then use the CLI to upload entities to the service running at http://localhost:8000.

In practice, this will look like this:

# Set the relevant environment variables
export ENTITIES_SERVICE_BASE_URL=http://localhost:8000
export ENTITIES_SERVICE_HOST=${ENTITIES_SERVICE_BASE_URL}

# Start the service
docker compose up -d

# Upload entities using the CLI
entities-service upload my_entities.yaml --format=yaml

The my_entities.yaml file should contain one or more entities with uri values of the form http://localhost:8000/....

Extra pytest markers

There are some custom pytest markers:

  • skip_if_live_backend: skips the test if the --live-backend flag is set. Add this marker to tests that should not be run against a live backend. Either because they are not relevant for a live backend, or because they currently impossible to replicate within a live backend.

    A reason can be specified as an argument to the marker, e.g.:

    @pytest.mark.skip_if_live_backend(reason="Cannot force an HTTP error")
    def test_something():
        ...

    Availability: This marker is available for all tests.

  • skip_if_not_live_backend: skips the test if the --live-backend flag is not set. Add this marker to tests that should only be run against a live backend. Mainly due to the fact that the mock backend does not support the test.

    A reason can be specified as an argument to the marker, e.g.:

    @pytest.mark.skip_if_not_live_backend(reason="Indexing is not supported by mongomock")
    def test_something():
        ...

    Availability: This marker is available for all tests.

Extra pytest fixtures

There is one fixture that may be difficult to locate, this is the parameterized_entity fixture. It can be invoked to automatically parameterize a test, iterating over all the valid entities that exist in the valid_entities.yaml static test file. It will return one of these entities as a parsed dictionary for each iteration, i.e., within each test.

The fixture is available for all tests.

Licensing & copyright

All files in this repository are MIT licensed.
Copyright by Casper Welzel Andersen, SINTEF.

Acknowledgements

This project is made possible by funding from:

  • MEDIATE (2022-2025) that receives funding from the RCN, Norway, FNR, Luxembourg, and SMWK, Germany via the M-ERA.NET programme, project9557. M-ERA.NET 2 and M-ERA.NET 3 have received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreements No 685451 and No 958174.
  • MatCHMaker (2022-2026) that receives funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No 101091687.