Skip to content

Latest commit

 

History

History
656 lines (509 loc) · 30.3 KB

CONTRIBUTING.md

File metadata and controls

656 lines (509 loc) · 30.3 KB
Raw

📊 Metrics

💪 Interested in contributing?

Nice! Read the few sections below to understand how this project is structured and how to implement new features.

👨‍💻 Extending metrics

This section explains how to create new templates and plugins. It also contains references about used packages, global project structure and list which contributions on are accepted.

🤝 Accepted contributions

Thanks for your interest in metrics and wanting to help it growing!

Review below which contributions are accepted:

Section Addition Editions Notes
🧩 Plugins ✔️ ✔️
  • New plugins are welcomed provided they're not redundant with existing plugins
  • New features for existing plugins are allowed but must be optional
🖼️ Templates ❌ 〽️
  • New templates are not allowed (use 📕 Community templates instead)
  • Templates editions are allowed for new features additions (but must remain consistent with current visuals)
🧪 Tests ✔️ ✔️
  • Everything that make metrics more stable is welcomed!
🧱 Core ❌ 〽️
  • Core editions impacts all rendering process and should be avoided unless necessary
🗃️ Repository ❌ ❌
  • Workflows, license, readmes, etc. usually don't need to be edited

Legend

  • ✔️: Contributions welcomed!
  • 〽️: Contributions welcomed, but must be discussed first
  • ❌: Only maintainers can manage these files

Before working on something, ensure that it isn't already in progress and that it will not duplicate any open pull requests (including drafts). If you're unsure, always open an issue first to get insights and gather feedback.

Even if your changes don't get merged in lowlighter/metrics, please don't be too sad. Metrics is designed to be highly customizable, so you can always decide to generate metrics on your forked repository 🙂!

📐 Project architecture

Following diagram explain how metrics code is structured:

Architecture

🗂️ Project structure

This section explain how metrics is structured.

  • source/app/metrics/ contains core metrics files
  • source/app/action/ contains GitHub action files
    • index.mjs contains GitHub action entry point
    • action.yml contains GitHub action descriptor
  • source/app/web/ contains web instance files
    • index.mjs contains web instance entry point
    • instance.mjs contains web instance source code
    • settings.example.json contains web instance settings example
    • statics/ contains web instance static files
      • app.js contains web instance client source code
      • app.placeholder.js contains web instance placeholder mocked data
  • source/app/mocks/ contains mocked data files
    • api/ contains mocked api data
      • axios/ contains external REST APIs mocked data
      • github/ contains mocked GitHub api data
    • index.mjs contains mockers
  • source/plugins/ contains source code of plugins
    • README.md contains plugin documentation
    • metadata.yml contains plugin metadata
    • index.mjs contains plugin source code
    • queries/ contains plugin GraphQL queries
  • source/templates/ contains templates files
    • README.md contains template documentation
    • metadata.yml contains template metadata
    • image.svg contains template image used to render metrics
    • style.css contains style used to render metrics
    • fonts.css contains additional fonts used to render metrics
    • template.mjs contains template source code
  • tests/ contains tests
    • metrics.test.js contains metrics testers
  • Dockerfile contains docker instructions used to build metrics image
  • package.json contains dependencies and command line aliases
📦 Packages

Below is a list of used packages.

🖼️ Templates

Templates require you to be comfortable with HTML, CSS and JavaScript (EJS flavored).

Metrics does not really accept new default templates in order to avoid bloating main repository (and it also makes maintaining easier upon new plugins releases).

Redesigns of existing templates are also typically avoided to keep visual consistency across all versions, but fear not! Users will still be able to use your custom templates thanks to community templates!

If you make something awesome, don't hesitate to share it!

💬 Creating a new template from scratch

Find a cool name for your template and run:

npm run quickstart -- template <template_name>

It will create a new folder in source/templates with the following files:

  • A README.md to describe your template and document it
  • An image.svg with base structure for rendering
  • A metadata.yml which list templates attributes and supported formats
  • A partials/ folder where you'll be able to implement parts of your template
    • A partials/_.json with a JSON array listing these parts in the order you want them displayed (unless overridden by user with config_order option)

If needed, you can also create the following optional files:

  • A fonts.css containing base64 encoded custom fonts
  • A styles.css with custom CSS that'll style your template
  • A template.mjs with additional data processing and formatting at template-level
    • When your template is used through setup_community_templates on official releases, this is disabled by default unless user trusts it by appending +trust at the end of source
    • You can specify the default template.mjs fallback by filling extends key in your metadata.yml (defaults to "classic" template)

If inexistent, these will fallback to classic template files.

Templates are auto-loaded based on their folder existence, so there's no need to register them somewhere.

💬 Creating image.svg and partials

The base structure for rendering looks like below:

<svg xmlns="http://www.w3.org/2000/svg" width="480" height="99999" class="<%= !animated ? 'no-animations' : '' %>">

  <defs><style><%= fonts %></style></defs>
  <style><%= style %></style>

  <foreignObject x="0" y="0" width="100%" height="100%">
    <div xmlns="http://www.w3.org/1999/xhtml" xmlns:xlink="http://www.w3.org/1999/xlink">
      <% for (const partial of [...partials]) { %>
        <%- await include(`partials/${partial}.ejs`) %>
      <% } %>

      <div id="metrics-end"></div>
    </div>
  </foreignObject>

</svg>
  • fonts and style variables will both be populated with the same content as your fonts.css and styles.css files
    • (or those of classic template files if inexistent)
  • partials variable will be populated with partials/_.json content
    • Main loop will iterate over this array to include all defined partials
  • #metrics-end is a special HTML tag which must remain at the bottom of SVG template

As you can see, we exploit the fact that SVG images are able to render HTML and CSS content so designing partials is the same as creating static web pages.

EJS framework is also used to programmatically create content through the help of templating tags (<% %>).

💬 Filling metadata.yml

metadata.yml is an optional file which describes what account types are allowed, which formats are supported, etc.

Here's an example:

name: "🖼️ Template name"
extends: classic  # Fallback to "classic" template "template.mjs" if not trusted
index: ~          # Leave as it (this is used to order plugins on metrics README.md)
supports:
  - user          # Support users account
  - organization  # Support organizations account
  - repository    # Support repositories metrics
formats:
  - svg           # Support SVG output
  - png           # Support PNG output
  - jpeg          # Support JPEG output
  - json          # Support JSON output
  - markdown      # Support markdown output
  - markdown-pdf  # Support PDF output

Core plugin will automatically check whether template supports given account or repository and output format and will throw an error in case they aren't compatible.

💬 Adding custom fonts

⚠️ This significantly increases rendered metrics filesize and thus not recommended

When using this feature, you should aim to restrict used charset to avoid including useless data.

Here's a quick step-by-step tutorial to create base64 encoded fonts:

    1. Find a font on fonts.google.com
    • Select regular, bold, italic and bold+italic fonts
    • Open embed tab and extract href
    1. Open extracted href in a browser and append &text= parameter with list of used characters
    • e.g. &text=0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz
    1. Download each font file from urls present in generated stylesheet
    1. Convert them into base64 with woff format on transfonter.org
    1. Download archive and extract it
    1. Copy content of generated stylesheet to fonts.css
    1. Update your template style.css to use the new font
🧩 Plugins

Plugins lets add new features with additional content to rendered metrics and are coded with JavaScript modules.

New plugins are welcomed, but maintainers have no obligation to maintain them. It means you (as author) may be notified about open issues regarding related plugin.

💬 Creating a new plugin from scratch

Find a cool name for your plugin and run:

npm run quickstart -- plugin <plugin_name>

It will create a new folder in source/plugins with the following files:

  • A README.md to describe your plugin and document it
  • An index.mjs with minimal plugin code
  • A metadata.yml which list plugin attributes and inputs
  • A tests.yml for unit tests

Here are some guidelines to follow about plugins:

  • They should never be dependent on output produced by other plugins (though allowed to re-use core and base data)
    • It allows parallelization of plugins execution
    • It avoids creating inter-dependencies which makes it confusing for both developers and users
  • Use of new external dependencies should be avoided
    • Adding new libraries to use only ~5% of its possibilities is just a waste
      • For APIs, most of the time a few HTTP calls instead of installing a full SDK wrapper is more than sufficient
      • imports probably already contains a library or a function that can help you achieving what you want
      • It also add more unstability since metrics may break upon external API breaking changes
  • Use of raw commands (spawning sub-process) should be avoided unless needed
    • If mandatory:
      • Use which detect whether command is available
      • For Windows, wrap command with WSL
      • It should at least work on Linux (action runs on ubuntu) but also on other OS to ease development and maintaining
  • Errors should be handled gracefully with error messages
  • Plugins arguments should NEVER be directly edited from inside a plugin
    • These are used by all plugins, including core and base so it would create unattended side effects
  • They should include customization options (limit entries, detailed output, etc.) for end user
    • Be sure to follow other options "lexical fields" to keep consistency and a certain abstraction level. For example use my_plugin_user and not my_plugin_username or my_plugin_login
    • Options should only affect content, not styling

You'll also need to an unused emoji to use as your plugin icon.

Plugins are auto-loaded based on their folder existence, so there's no need to register them somewhere.

💬 Implementing index.mjs and gathering new data from external APIs

Default exported function in index.mjs will receive the following inputs:

  • login, set to GitHub login
  • q, with query parameters (formatted with dots (.) instead of underscores (_) and without plugin_ prefix)
  • imports, with libraries and utilitaries
  • data and computed, with all data gathered from core and base
  • graphql and rest, with authenticated octokit clients (for GitHub APIs)
  • queries, with autoloaded GraphQL queries and replacers
  • account, set to account type ("user" or "organization")

Second input contains configuration settings from settings.json (which is mostly used by web instances) and also all user inputs of type token.

As said previously, plugins arguments should NEVER be directly edited from it, since these are used by all plugins, including core and base so it would create unattended side effects.

As for data gathering related to:

For GraphQL queries, use queries which will auto-load all queries from queries directory and will lets you create custom queries on the fly.

For example:

//Calling this
  await graphql(queries.myplugin.myquery({login:"github-user", account:"user"}))

//With this in source/plugins/myplugin/queries/myquery.graphql
  query MyQuery {
    $account(login: "$login") {
      name
    }
  }

//Will have the same result as calling this
  await graphql(`
    query MyQuery {
      user(login: "github-user") {
        name
      }
    }
  `)
💬 Filling metadata.yml

metadata.yml is a mandatory file which describes what inputs are allowed, which entities are supported, etc.

Here's an example:

name: "🧩 Plugin name (with emoji icon)"
cost: Estimates how many GitHub requests is used during plugin execution ("N/A" for Third-Party services)
categorie: github # Plugin categorie ("github", "social" or "other")
index: ~          # Leave as it (this is used to order plugins on metrics README.md)
supports:
  - user          # Support users account
  - organization  # Support organizations account
  - repository    # Support repositories metrics
inputs:

  # A comment detailing input purposes
  # An input must have at least a "description" and a "default" (used to generated GitHub Action `action.yml`)
  plugin_input:
    description: Short description (few words)
    type: boolean
    default: no

Because of GitHub Actions limitations, only strings and numbers are actually supported by action.yml inputs. Metrics apply additional post-processing to handle inputs.

Supported input types are boolean, string, number, array and json.

  • Allowed values for string and array types may be restricted using values attribute
    • Special default values .user.login, .user.twitter and .user.website will respectively be replaced by user's login, Twitter username and website (not available when token is set to NOT_NEEDED by user)
  • Lower and upper limits for number may be set using min and max attribute
  • Array format attribute define how string should be splitted (comma-separated or space-separated)

You can additionally specify an example which will also be used in web instance input placeholder.

Inputs will be available through imports.metadata.plugins.*.inputs with correct typing and default values (plugin_ prefix will be dropped, and all underscored (_) will be changed to dots (.) instead):

//Load inputs
  let {limit, "limit.field":limit_field} = imports.metadata.plugins.myplugin.inputs({data, account, q})

Additionally, if account type isn't supported, this method will automatically prevent your plugin from running by throwing an error.

💬 Creating a new partial

In templates you want to support, create a new .ejs file in partials folder and paste the following for a quick start:

<% if (plugins./* your plugin name */) { %>
  <section>
    <div class="row">
      <% if (plugins./* your plugin name */.error) { %>
        <section>
          <div class="field error">
            <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 16 16" width="16" height="16"><path fill-rule="evenodd" d="M2.343 13.657A8 8 0 1113.657 2.343 8 8 0 012.343 13.657zM6.03 4.97a.75.75 0 00-1.06 1.06L6.94 8 4.97 9.97a.75.75 0 101.06 1.06L8 9.06l1.97 1.97a.75.75 0 101.06-1.06L9.06 8l1.97-1.97a.75.75 0 10-1.06-1.06L8 6.94 6.03 4.97z"></path></svg>
            <%= plugins./* your plugin name */.error.message %>
          </div>
        </section>
      <% } else { %>
          <section>
            <%# Do stuff in there -%>
          </section>
      <% } %>
    </div>
  </section>
<% } %>
  • First conditional statement ensures that partial is displayed only when plugin is enabled
  • Nested conditional statement check plugin output
    • If it failed, an error message instead will be displayed instead
    • If it succeeded, second section in render.

Additional CSS rules may be added to style.css of edited template, but ensure it does not break other plugins rendering.

💬 Redacting plugin's README.md

README.md is a mandatory file which contains basic documentation about your plugin.

It should contain:

  • A brief description of plugin's purpose
  • A table including demos (this is automatically extracted upon metrics README.md auto-generation)
    • Multiple demo are allowed, but if they're too big, please embed them in <detail></details> tags
    • Note that the <img width="900" height="1" alt=""> must not be deleted, it is used to fill table height
  • Additional documentation on how to setup, token creation, etc.
  • A workflow example, with all available options
### 🧩 Your plugin name

<table>
  <td align="center">
    <img src="">
    <img width="900" height="1" alt="">
  </td>
</table>

#### ℹ️ Examples workflows

[➡️ Available options for this plugin](metadata.yml)

'''yaml
- uses: lowlighter/metrics@latest
  with:
    # ... other options
    plugin_custom: yes
'''
💬 Fast prototyping and testing

The easiest way to test and prototype your plugin is to setup a web instance. See documentation for more informations about that.

Open a browser and try to generate metrics with new your plugin enabled to see if it works as expected:

http://localhost:3000/your-github-login?base=0&your-plugin-name=1

Once ready, define test cases in your plugin directory tests.yml.

These tests will be run with:

  • Metrics action
  • Metrics web instance
  • Metrics web instance placeholder (rendered by browser)

Most APIs (including GitHub) usually have a rate-limit to ensure quality of service. This is why APIs output must be mocked and added in source/app/mocks/api/ in order for tests to be able to be performed anytime.

Files from these directories are auto-loaded, so you just need to create new functions (see other mocked data for examples).

Finally, edit source/app/web/statics/app.placeholder.js to add mocked result (but this time from metrics server) so users will be able to render placeholder preview in web instance.

💬 Submitting a pull request

Ensure that:

  • metadata.yml is correctly filled
  • tests.yml has defined test cases
  • mocks/api has mocked data for external APIs
  • app.placeholder.js has been updated with mocked plugin output
  • README.md of plugin explain how plugin works
    • <table> tag MUST remain present (along with <img width="900" height="1" alt="">) as these are extracted for global README.md
  • npm run linter outputs no errors
  • npm test is successful

Use config.output option to render a PNG version of your plugin:

http://localhost:3000/your-github-login?base=0&your-plugin-name=1&config.output=png

And finally open a new pull request and ensure that all builds succeed.

Additionally, you can test your changes with GitHub action by using your fork in your workflow:

- uses: github-login/metrics@feat-my-plugin
  with:
    # options...

Global README.md, plugins/README.md, templates/README.md, action.yml and settings.example.json are automatically rebuild by GitHub action, do not edit them manually.

🎬 Behind the scenes

This section explore some topics which explain globally how metrics was designed and how it works.

💬 Creating SVGs images on-the-fly

Metrics actually exploit the possibility of integrating HTML and CSS into SVGs, so basically creating these images is as simple as designing static web pages. It can even handle animations and transparency.

Metrics are html

SVGs are templated through EJS framework to make the whole rendering process easier thanks to variables, conditional and loop statements. Only drawback is that it tends to make syntax coloration a bit confused because templates are often misinterpreted as HTML tags markers (<%= "EJS templating syntax" %>).

Images (and custom fonts) are encoded into base64 to prevent cross-origin requests, while also removing any external dependencies, although it tends to increase files sizes.

Since SVG renders differently depending on OS and browsers (system fonts, CSS support, ...), it's pretty hard to compute dynamically height. Previously, it was computed with ugly formulas, but as it wasn't scaling really well (especially since the introduction of variable content length plugins). It was often resulting in large empty blank spaces or really badly cropped image.

To solve this, metrics now spawns a puppeteer instance and directly render SVG in a browser environment (with all animations disabled). An hidden "marker" element is placed at the end of the image, and is used to resize image through its Y-offset.

Metrics marker

Additional bonus of using pupeeter is that it can take screenshots, making it easy to convert SVGs to PNG output.

Finally, SVGs image can be optimized through svgo, which helps to remove unused attributes and blank space, while also reducing a bit the file size.

💬 Gathering external data from GitHub APIs and Third-Party services

Metrics mostly use GitHub APIs since it is its primary target. Most of the time, data are retrieved through GraphQL to save APIs requests, but it sometimes fallback on REST for other features. Octokit SDKs are used to make it easier.

As for other external services (Twitter, Spotify, PageSpeed, ...), metrics use their respective APIs, usually making https requests through axios and by following their documentation. It would be overkill to install entire SDKs for these since plugins rarely uses more than 2/3 calls.

In last resort, pupeeter is seldom used to scrap websites, though its use tends to make things slow and unstable (as it'll break upon HTML structural changes).

💬 Web instance and GitHub action similarities

Historically, metrics used to be only a web service without any customization possible. The single input was a GitHub username, and was composed of what is now base content (along with languages and followup plugin, which is why they can be computed without any additional queries). That's why base content is handled a bit differently from plugins.

As it gathered more and more plugins over time, generating a single user's metrics was becoming costly both in terms of resources but also in APIs requests. It was thus decided to switch to GitHub Action. At first, it was just a way to explore possibilities of this GitHub feature, but now it's basically the full-experience of metrics (unless you use your own self-hosted instance).

Both web instance and Action actually use the same entrypoint so they basically have the same features. Action just format inputs into a query-like object (similarly to when url params are parsed by web instance), from which metrics compute the rendered image. It also makes testing easier, as test cases can be reused since only inputs differs.

💬 Testing and mocking

Testing is done through jest framework.

While the best would be to work with real data during testing, to avoid consuming too much APIs requests for testing (and to be more planet friendly), they're mocked using JavaScript Proxies and Faker.js. Basically function calls are "trapped" and send randomly generated data from Faker.js if we're in a development environment.

Written by lowlighter