Drones for Marine Science and Agricultural

d4ms

[]()

A Cal Poly research project on using drones for marine and agricultural research. We are using this gitHub repo for software development and this Google Drive for other project documents and training data storage. We are using dvc for training data version control in order to avoid problems that git has with large files.

Setup

When working on multiple projects on the same machine it is a good idea to use a virtual Python environment. I like to use Anaconda but Python comes equipped with the virtualenv package if you don't want to install 3rd party software. Check out this tutorial if you would like some guidance. This document assumes you are using a bash shell, so your mileage may vary on other systems.

Run the following code in your virtual environment to install all dependencies:

pip install -r requirements.txt

Documentation

The full documentation can be found under /docs and in a more readable format at readTheDocs. We are using the Google documentation style.

Packages

This project is split into 4 modules, each containing separable code units for the different project groups.

drone

Drone flight and behavior to be utilized mainly by the Drones group. Use the #drones Slack channel to ask questions about this module.

mission

Mission planning and control to be used mainly by the Multi-Agent fleet group. Use the #agents Slack channel to ask questions about this module.

network

Ad hoc network protocol. This module is maintained by Joel Valdovinos Miranda.

model

Deep learning model for spotting interesting objects. Use the #machine-learning Slack channel to ask questions about this module.

Style Checking

In order to maintain project readability, we will use the following style checkers and linters. Linting is a commonly used term in software development for programs that automatically fix style issues whenever you save them.

PyCodeStyle for Python

To install PyCodeStyle (formerly known as PEP8), run the following command:

pip install pycodestyle

In your editor of choice, add the python path and enable linting. For VSCode, open the command palette (Ctrl+Shift+P) and select the Python: Enable Linting command.

Alternatively, for VSCode, add the following to .vscode/settings.json, replacing <path> with your python path:

{
  "python.pythonPath": "<path>",
  "python.linting.Enabled": true,
  "python.linting.lintOnSave": true
}

Format on Save

In your editor of choice, enable format on save so that linting is done automatically. For VSCode, add the following to .vscode/settings.json:

{
  "editor.defaultFormatter": null,
  "editor.formatOnSave": true,
  "[javascript]": {
    "editor.defaultFormatter": "esbenp.prettier-vscode"
  }
}

Prettier and ESLint for JavaScript

For modules involving a web interface, using Node.js is a powerful and ( relatively) easy approach. This setup is inspired by this tutorial. Please follow the link for further guidance.

To install ESLint and Prettier, run the following command in the frontend directory:

npm install --save-dev eslint prettier eslint-config-prettier eslint-plugin-prettier husky lint-staged

The following Prettier options are modified from their default value:

TODO

• Import all data to Google Drive
• Refactor code in the drone module
• Update each README.
• Agree on project style standards
• Move style guides to CODE_OF_CONDUCT file?
• add ReadTheDocs page
• add Sphinx (with napoleon for Google style comments?)
• add dvc