Overview
This section contains application examples demonstrating how to deploy and interact with software running on Clustergate-2.
These examples are still under active development. More use cases and integrations will be added over time.
Please note that, by default, all output produced by a DPhi Pod—whether written to standard output (stdout) or standard error (stderr)—is captured and stored by the platform. This data is included in the downlink package and therefore counts against the user's downlink budget.
Users are strongly encouraged to write any data they wish to preserve, reuse across executions, or downlink intentionally to the mounted /data volume. Logging large amounts of information directly to the console should be avoided, as it can significantly increase the amount of generated downlink data and unnecessarily consume the available data budget.
All example source code is available under the examples/ directory of the repository.
Getting Started
Clone the public documentation repository:
git clone git@github.com:DPhi-Space/public-documentation.git
cd public-documentation
Index of Examples
- User API example A Python client that creates a namespace, creates or updates an operation, runs an execution, and lists results.
Application Examples
- Telemetry Client API A Python client that connects to the onboard telemetry API, fetches health, telemetry, and statistics data, and stores it locally in JSON format. It can run both natively or inside Docker and integrates directly with the onboard database and telemetry REST service.
- Fisheye Telemetry Analysis This example is designed to fetch fisheye images from the onboard telemetry API, perform GPU-accelerated analysis using CUDA, and generate detailed insights including deep features, color statistics, edge detection, texture, and histogram analysis.
Getting the most out of the examples
- Review the operation YAML before running so you understand task order and file paths.
- Upload any required files to the namespace file area first.
- Start in
cg2.embefore running incg2.fm.