Examples¶

We provide a couple of examples of application services and , Jupyter Notebooks and configuration files for getting started quickly:

Configuration Files¶

Upon startup of the platform, we automatically copy all files within the /store folder of the GitHub repo to the data store.

These files include general configuration files, like scheduler specifications, access control lists and Jupyter Notebooks.

Jupyter Notebooks¶

We provide simple Jupyter Notebooks inside /store/notebooks which are evaluated by the [notebook executor](architecture.md#Notebook Executor).

The evaluation of these notebooks is usually triggered by the scheduler whenever new sample data has been added to the sore by the [sample recorder](architecture.md#Sample Recorder) or at regular intervals (e.g. daily).

Application Services¶

Application services are the primary approach to implement business logic in the platform. They allow you to integrate third-party tools.

Two examples for such services are shipped as standalone Docker images in images/examples.

Service Deployment¶

To deploy your application services, follow these steps:

Package Your Application as a Docker Container If your application uses parts of the SEGuRo Platform toolchain (e.g., the broker or store), it is recommended to use the seguro-platform base image. Below is an example Dockerfile:
```
# Dockerfile
FROM registry.localhost/seguro:latest

COPY my_service.py /platform/seguro
WORKDIR /platform/seguro
```
Login to the Platform Docker Registry Use the admin credentials configured in the environment file to log in to the platform’s Docker registry. By default, the credentials are admin / s3gur0herne (see Configuration and Access for more details).
```
docker login registry.localhost
```
Push Your Image to the Docker Registry Push your Docker image to the registry using the following command:
```
docker push registry.localhost/my-image:my-tag
```
Upload a Job Configuration File To start your service on the platform using the scheduler, upload a YAML job specification describing the scheduling process to the object store at seguro/config/jobs/. This can be done via the store UI frontend. Below is a minimal example configuration. For more details, refer to the job.yaml.example.
```
# my-job.yaml
# Example configuration for a job

container:
  image: registry.localhost/my-image:my-tag

  environment:
    MY_VAR: "foo"

  command: ["python my_service.py"]
```
Note

Since no explicit trigger is defined, the job will start immediately after uploading the file and after each restart of the scheduler service.

Job Worker¶

The example job worker is a script that processes store data triggered by the creation of specific store objects. When triggered, it retrieves a data frame from the created object, doubles its values, and saves the processed frame to a new location in the store (measurements_processed).

For accesing the store, it uses the store helper class.

Dockerfile:

# SPDX-FileCopyrightText: 2023-2024 Steffen Vogel, OPAL-RT Germany GmbH
# SPDX-License-Identifier: Apache-2.0

FROM python:3.11.8-bookworm

RUN mkdir /src
WORKDIR /src

COPY . /src
RUN pip install --no-cache-dir .

CMD ["job-worker"]

Job Specification:

---
# SPDX-FileCopyrightText: 2023-2024 Steffen Vogel, OPAL-RT Germany GmbH
# SPDX-License-Identifier: Apache-2.0

build: true

triggers:
  new-data:
    type: created
    prefix: XXX

container:
  build:
    context: ./images/examples/job-worker

Streaming Worker¶

The example streaming worker is a script that processes all incoming samples on a given topic (data/measurements/mp1) at the broker. When triggered, it scales the values based on their index within the sample, and publishes the resulting scaled samples to another topic (data/measurements/processed_by_streaming_worker/mp1).

For accessing the broker, it uses the broker helper class

Dockerfile:

# SPDX-FileCopyrightText: 2023-2024 Steffen Vogel, OPAL-RT Germany GmbH
# SPDX-License-Identifier: Apache-2.0

FROM python:3.11.8-bookworm

RUN mkdir /src
WORKDIR /src

COPY . /src
RUN pip install --no-cache-dir .

CMD ["streaming-worker"]

Job Specification: