dora: A Rust repository from AutoCore

Dataflow Oriented Robotic Architecture ⚡

dora goal is to be a low latency, composable, and distributed data flow.

This project is in early development, and many features have yet to be implemented with breaking changes. Please don't take for granted the current design.

dora primary support is with Linux ( Ubuntu 20.04 and Ubuntu 22.04 ) as it is the primary OS for both Cloud and small computers. If you wish to use dora with another OS, please compile from source.

📖 Documentation

The documentation can be found here: https://dora-rs.github.io/dora/

Getting started

Install dora binaries from GitHub releases

For linux

wget https://github.com/dora-rs/dora/releases/download/<version>/dora-<version>-x86_64-Linux.zip
unzip dora-<version>-x86_64-Linux.zip
python3 -m pip install dora-rs==<version>
PATH=$PATH:$(pwd):$(pwd)/iceoryx
dora --help

This is x86_64 only for the moment.

Create a new dataflow

dora new abc_project --lang python
cd abc_project

This creates the following abc_project directory

.
├── dataflow.yml
├── node_1
│   └── node_1.py
├── op_1
│   └── op_1.py
└── op_2
    └── op_2.py

Start dora-coordinator in a separate terminal window

# New terminal window
dora up

Start your dataflow

# Other window
dora start dataflow.yml
# Output: c95d118b-cded-4531-a0e4-cd85b7c3916c

The output is the unique ID of the dataflow instance, which can be used to control it through the dora CLI.

You will see in your dora-coordinator window operators receiving ticks.

Received input tick, with data: b''
Received input tick, with data: b''
Received input tick, with data: b''
...

Stop your dataflow

dora stop c95d118b-cded-4531-a0e4-cd85b7c3916c

(Pass the ID returned by dora start here.)

You can then add or modify operators or nodes. For adding nodes easily, you can use the dora CLI again:

Run dora new --kind operator --lang Rust <name> to create a new Rust operator named <name>.
Run dora new --kind custom-node --lang Rust <name> to create a new custom node named <name>.

You need to add the created operators/nodes to your dataflow YAML file.

You can also download already implemented operators by putting links in the dataflow. This example will launch a webcam plot stream.

communication:
  zenoh:
    prefix: /abc_project

nodes:
  - id: op_1
    operator:
      python: https://raw.githubusercontent.com/dora-rs/dora-drives/main/operators/webcam.py
      inputs:
        tick: dora/timer/millis/100
      outputs:
        - image
  - id: op_2
    operator:
      python: https://raw.githubusercontent.com/dora-rs/dora-drives/main/physicals/plot.py
      inputs:
        image: op_1/image

Make sure to have a webcam and cv2 install via: pip install opencv-python

✨ Features

Composability as:

Low latency as:

written in ...Cough...blazingly fast ...Cough... Rust.
PubSub communication with iceoryx

Distributed as:

PubSub communication with zenoh
Distributed telemetry with opentelemetry

🏁 Further reading

Check out dora-drives for a template of an autonomous vehicle within a simulation.

⚖️ LICENSE