Skip to content

nathanstitt/jsb-autopilot

 
 

Repository files navigation

JP's notes

Run this in VSCode, using the provided devcontainer.

yarn jest --watch

rollup -c to build a minified version, then strip off the "use strict" at the front and the export at the end. Copy remainder into a tank script.

yard add blah --dev when new packages needed.

Nathan‘s notes

  • install libraries with: yarn install
  • test code by running in browser: yarn start
    • stimulation will open in your default browser and restart after every code change
    • future improvement will be to store selected competitor in localStorage and re-use it after reload
  • Add your code to main.js, after saving it the browser should reload so you can view results immediatly
  • Build code for copy'n'paste to competition: yarn build
  • Code will be built ininto a dist/index.<hash>.js file, copy from it

Usage

Copy the bundle/bundle.min.js into your JS Battle tank script. Remove

"use strict";Object.defineProperty(exports,"__esModule",{value:!0});

from the front and remove

exports.Autopilot=Autopilot,exports.Constants=Constants;

from the end after you paste it in.

See bundle/bundle.js for a non-minified version. Then in your tank script (below where you pasted the bundle), say:

let autopilot = new Autopilot();

tank.loop(function(state, control) {
	autopilot.update(state, control);

  // your code here

The Autopilot object gives you some utility functions to help you build your tank:

  • After your radar has scanned a horizontal and vertical wall, the autopilot will have calculated the battlefield origin for you. You can access it with autopilot.origin. autopilot.isOriginKnown() will return true when the origin is available.
  • If you know the origin through other means (e.g. from your teammates), you can set it with setOrigin(x,y).
  • lookEverywhere sets your radar to spin at max speed.
  • lookAtEnemy(enemy) locks your radar onto an enemy. enemy is an object with x and y coordinates.
  • isWallCollisionImminent(inTicks=3) returns true if it looks like your tank will hit a wall in the number of ticks provided.
  • turnToAngle(angle) turns your bot to the given absolute angle. Returns the angle remaining to turn.
  • turnToPoint(x,y,baseOnZeroOrigin=false) will turn your tank to face the given coordinates. If the origin is known and you set baseOnZeroOrigin to true, you can give coordinates assuming the origin is (0,0). Returns the angle remaining to turn.
  • moveToPoint(x,y,baseOnZeroOrigin=false) will move your bot to the provided coordinates. If the origin is known and you set baseOnZeroOrigin to true, you can give coordinates assuming the origin is (0,0).
  • moveAlongAngle(angle) moves your tank in the given direction. You can use constants here, e.g. moveAlongAngle(Constants.EAST).
  • loopOnPath(positions, basedOnZeroOrigin, tolerance=50) cycles through an array of {x:, y:} positions, which can be based on zero origin if the origin is known. The tolerance says how close is close enough to say we've made it to one of the positions.
  • stopLoopOnPath() clears a loopOnPath call.
  • stop() stops the turn, throttle, and boost.
  • shootEnemy() shoots the enemy (enemy object from the state), using a predictive aim based on the enemy's speed and angle.
  • extrapolatedPosition(inTicks) gives your bot's linear extrapolated position in the given number of ticks. May not totally account for speed that is ramping up or slowing down.
  • extrapolatedOuterPosition(inTicks) is the same as the above except it uses a circle around the tank to give the position (to help with collision detection)
  • speed() gives the tank's speed.
  • Autopilot.extrapolatedPosition(startPosition, travelAngle, travelSpeed, inTicks) computes a general linear extrapolated position.

javascript-library-template GitHub license Tests codecov npm minizipped size

Click below to create a new GitHub repository using this template:


This JavaScript library template allows you to easily develop, collaborate on and publish a JavaScript library with all the modern tooling you'd expect from the current JavaScript ecosystem.

Why should you use this? One of the hidden challenges of authoring opensource JavaScript libraries is to provide a project that is easy to contribute to. You want people to join your project. Doing so requires a good amount of boilerplate: testing, code coverage, dependencies maintenance, release scripts, tooling requirements (Node.js, Yarn and which versions are we using again?), code editor configuration, formatting, linting... Well, this is the goal of this template: to provide sensible and modern defaults to all those subjects. So that once set up, you can focus on ⌨️ coding, 🙌 collaborating and 🚀 shipping.

The goals of the template are to:

  • Ease the contribution of the library by providing reproducible environments for developers and CI
  • Automate as much as possible, from testing to releasing and upgrading dependencies
  • Provide good defaults for users of Visual Studio Code

Features:

  • EditorConfig: easy contributions from any code editor.
  • ESLint: launched in the test script.
  • Prettier: launched in the test script, with markdown, JavaScript, CSS and JSON files support (including automatic package.json formatting).
  • Automatic VSCode formatting and linting: using VSCode extensions recommendations and workspace settings in .vscode/ folder.
  • Yarn version pinning: via Yarn policies, so anyone contributing or any system accessing your library will use the same Yarn version without having to think about it.
  • Node.js version pinning: via nvm, so anyone contributing or any system accessing your library will use the same Node.js version without having to think about it.
  • Jest: launched in the test script, also with the right VSCode settings providing a testing workflow inside VSCode using vscode-jest extension.
  • GitHub actions: automatic testing and releasing from GitHub: npm publish and GitHub releases are automatically created. Note that the package.json in your repository is never updated (the version is always 0.0.0-development), only the one in npm is updated. This is surprising at first but as long as you display the published version in your README (like this template does) then you're fine. Find more information about this in the semantic-release documentation.
  • semantic-release: allows for automatic releases based on semver.org and conventional commits specification. The defaults are taken from the Angular git commit guidelines.
  • Codecov: launched in the test script on CI, ensures code coverage does not decrease on pull requests (free for public and private repositories).
  • Renovate configurated with the JavaScript library preset: this will automatically update your dependencies by opening pull request for you to approve or not. So you never have to think about it (free for public and private repositories).

Setup

Using this template requires a bit of setup, but way less than if you had to start from 0. Here's what you need to do:

Required steps: (needed every time you want to use the template)

  1. Create a new repository on GitHub based on this template
  2. Setup renovate for your new repository. If you previously installed the Renovate application to your account then this is just a box to tick when creating the repository
  3. Clone the new repository
  4. Change the package name and description in package.json
  5. Setup Codecov for your new repository. If you previously installed the Codecov application to your account then this is just a box to tick when creating the repository
  6. Setup semantic releases: run yarn semantic-release-cli setup in a terminal (This will ask for your npm and GitHub credentials)
  7. Add the previously generated GH_TOKEN and NPM_TOKEN secrets to the GitHub secrets of the new repository
  8. Install dependencies: run yarn in your terminal
  9. Develop your library: change code in lib/
  10. Test your library: you can either see tests results inside VSCode directly or run yarn jest --watch
  11. Check formatting of your code: run yarn check-formatting in your terminal
  12. Create your first release: open a pull request on your project, wait for tests to pass, merge and 💥 your library will be automatically released to npm and a GitHub release will be created

Optional steps: (needed only if you're doing them for the first time)

  1. If you're not using VSCode, if your editor has no EditorConfig support, then setup EditorConfig EditorConfig support
  2. Make sure you have npm 2fa auth-only configured. Releases can't be automated if you have 2fa setup for both authentication and publish. See https://semantic-release.gitbook.io/semantic-release/usage/ci-configuration#authentication-for-plugins
  3. Install nvm
  4. Install yarn

Status and next steps

The template is still pretty new (March 2020) and was done to author JavaScript libraries using ECMAScript modules for Node.js >= 12. Gradually, or given requests, we could update it to support:

  • Using different CI environments than GitHub actions
  • Authoring browser libraries
  • Generating README.md table of contents automatically
  • Better default README.md content (Install, API, Examples, ...)
  • create-javascript-library command line that would get most of the setup done easily
  • .github Pull requests template, issues templates, CONTRIBUTING files
  • add or change scripts to allow for auto-formatting
  • provide documentation on how to protect branches on GitHub
  • provide scripts to easily open a pull request once a branch is created
  • provide a way to check for semantic commits in PR
  • contributors list
  • https://github.com/apps/semantic-pull-requests

If you'd like to participate, if you have bugs or new ideas, open an issue or a pull request.

Recipes

Using yarn link

To use yarn link efficiently, do this:

> cd my-library
> yarn link
> yarn build --watch
> cd ../my-other-library
> yarn link my-library

Reformating all code

yarn format

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • JavaScript 83.6%
  • HTML 15.9%
  • Dockerfile 0.5%