docs(plugins/languages): full rewrite

This commit is contained in:
lowlighter
2022-01-15 12:06:05 -05:00
parent 8f1e017beb
commit 72c736c5b7
2 changed files with 175 additions and 192 deletions

View File

@@ -1,139 +1,139 @@
### 🈷️ Most used languages
<!--header-->
<!--/header-->
The *languages* plugin displays which programming languages you use the most across all your repositories.
<table>
<td align="center">
<details open><summary>Indepth analysis (clone and analyze repositories)</summary>
<img src="https://github.com/lowlighter/lowlighter/blob/master/metrics.plugin.languages.indepth.svg">
</details>
<details open><summary>Recently used (analyze recent activity events)</summary>
<img src="https://github.com/lowlighter/lowlighter/blob/master/metrics.plugin.languages.recent.svg">
</details>
<details><summary>Default algorithm</summary>
<img src="https://github.com/lowlighter/lowlighter/blob/master/metrics.plugin.languages.svg">
</details>
<details><summary>Default algorithm (with details)</summary>
<img src="https://github.com/lowlighter/lowlighter/blob/master/metrics.plugin.languages.details.svg">
</details>
<img width="900" height="1" alt="">
</td>
</table>
It is possible to use custom colors for languages instead of those provided by GitHub by using `plugin_languages_colors` option.
You can specify either an index with a color, or a language name (case insensitive) with a color.
Colors can be either in hexadecimal format or a [named color](https://developer.mozilla.org/en-US/docs/Web/CSS/color_value).
It is also possible to use a predefined set of colors from [colorsets.json](colorsets.json)
#### Using `indepth` statistics
Languages statistics are computed using the top languages provided by GitHub of each repository you contributed to.
If you work a lot with other people, these numbers may be less representative of your actual work.
The `plugin_languages_indepth` option lets you get more accurate metrics by cloning each repository you contributed to, running [linguist-js](https://github.com/Nixinova/Linguist) and then iterating over patches matching your username from `git log`. This method is slower than the first one.
> ⚠️ Although *metrics* does not send any code to external sources, you must understand that when using this option repositories are cloned locally temporarly on the GitHub Action runner. If you work with sensitive data or company code, it is advised to keep this option disabled. *Metrics* and its authors cannot be held responsible for any resulting code leaks, use at your own risk.
> Source code is available for auditing at [analyzers.mjs](/source/plugins/languages/analyzers.mjs)
> 🔣 On web instances, `indepth` is an extra feature and must be enabled globally in `settings.json`
#### `commits_authoring` option
Since Git lets you use any email and name for commits, metrics may not be able to detect whether you own a commit or not. By default, it'll check whether it matches your GitHub login.
For better results, it's advised to add either your surnames and eventually no-reply email addresses.
#### ➡️ Available options
## ➡️ Available options
<!--options-->
| Option | Type *(format)* **[default]** *{allowed values}* | Description |
| ------ | -------------------------------- | ----------- |
| `plugin_languages` | `boolean` **[no]** | Display most used languages metrics |
| `plugin_languages_ignored` | `array` *(comma-separated)* **[]** | Languages to ignore |
| `plugin_languages_skipped` | `array` *(comma-separated)* **[]** | Repositories to skip |
| `plugin_languages_limit` | `number` **[8]** *{0 ≤ 𝑥 ≤ 8}* | Maximum number of languages to display |
| `plugin_languages_sections` <sup>🧰</sup> | `array` *(comma-separated)* **[most-used]** *{"most-used", "recently-used"}* | Sections to display |
| `plugin_languages_colors` | `array` *(comma-separated)* **[github]** | Custom languages colors |
| `plugin_languages_aliases` | `string` **[]** | Custom languages names |
| `plugin_languages_details` | `array` *(comma-separated)* **[]** *{"bytes-size", "percentage", "lines"}* | Additional details |
| `plugin_languages_threshold` | `string` **[0%]** | Minimum threshold |
| `plugin_languages_indepth` <sup>🧰</sup> | `boolean` **[false]** | Indepth languages processing (see documentation before enabling) |
| `plugin_languages_analysis_timeout` <sup>🧰</sup> | `number` **[15]** *{1 ≤ 𝑥 ≤ 30}* | Languages analysis timeout |
| `plugin_languages_categories` | `array` *(comma-separated)* **[markup, programming]** *{"data", "markup", "programming", "prose"}* | Language categories to display |
| `plugin_languages_recent_categories` | `array` *(comma-separated)* **[markup, programming]** *{"data", "markup", "programming", "prose"}* | Language categories to display (for recently used section) |
| `plugin_languages_recent_load` | `number` **[300]** *{100 ≤ 𝑥 ≤ 1000}* | Number of events to load (for recently used section) |
| `plugin_languages_recent_days` | `number` **[14]** *{0 ≤ 𝑥 ≤ 365}* | Maximum event age (for recently used section) |
Legend for option icons:
* 🧰 Must be enabled in `settings.json` (for web instances)
<!--/options-->
*[→ Full specification](metadata.yml)*
## 🔎 `indepth` mode
#### Examples workflows
The default algorithm use the top languages provided of each repository you contributed to.
When working in collaborative projects with a lot of people, these numbers may be less representative of your actual work.
The `plugin_languages_indepth` option lets you use a more advanced algorithm for more accurates statistics.
Under the hood, it will clone your repositories, run [linguist-js](https://github.com/Nixinova/Linguist) (a JavaScript port of [GitHub linguist](https://github.com/github/linguist)) and iterate over patches matching your `commits_authoring` setting.
Since git lets you use any email and username for commits, *metrics* may not be able to detect a commit ownership if it isn't the same as your GitHub personal data. By default, it will use your GitHub username, but you can configure additional matching usernames and email addresses using `commits_authoring` option.
*Example: configuring `indepth` mode*
```yml
- uses: lowlighter/metrics@latest
with:
plugin_languages: yes
plugin_languages_indepth: yes
commits_authoring: firstname lastname, username, username@users.noreply.github.com
```
> 💡 This feature unlocks the `lines` option in `plugin_languages_details`
> ⚠️ This feature significantly increase workflow time
> ⚠️ Since this mode iterates over **each commit of each repository**, it is not suited for large code base, especially those with a large amount of commits and the ones containing binaries. While `plugin_languages_analysis_timeout` can be used to increase the default timeout for analysis, please be responsible and keep this feature disabled if it cannot work on your account to save GitHub resources and our planet 🌏
> ⚠️ Although *metrics* does not send any code to external sources, repositories are temporarily cloned on the GitHub Action runner. It is advised to keep this option disabled when working with sensitive data or company code. Use at your own risk, *metrics* and its authors **cannot** be held responsible for any resulting code leaks. Source code is available for auditing at [analyzers.mjs](/source/plugins/languages/analyzers.mjs).
> 🌐 Web instances must enable this feature in `settings.json`
## 📅 Recently used languages
This feature uses a similar algorithm as `indepth` mode, but uses patches from your events feed instead.
It will fetch a specified amount of recent push events and perform linguistic analysis on it.
> ⚠️ Note that *metrics* won't be able to use more events than GitHub API is able to provide
*Example: display recently used languages from 400 GitHub events from last 2 weeks*
```yml
- uses: lowlighter/metrics@latest
with:
plugin_languages: yes
plugin_languages_sections: recently-used
plugin_languages_recent_load: 400
plugin_languages_recent_days: 14
```
> 🌐 Web instances must enable this feature in `settings.json`
## 🥽 Controling which languages are displayed
Several options lets you customize which languages should be displayed.
It is possible to ignore completely languages or those lower than a given threshold, skip repositories, and filter by language categories.
*Example: hide HTML and CSS languages, skip lowlighter/metrics repository*
```yml
- uses: lowlighter/metrics@latest
with:
plugin_languages: yes
plugin_languages_ignored: html, css
plugin_languages_skipped: lowlighter/metrics
```
*Example: hide languages with less than 2% usage*
```yml
- uses: lowlighter/metrics@latest
with:
plugin_languages: yes
plugin_languages_threshold: 2%
```
> 💡 The threshold feature will automatically scale remaining languages so the total percentage is always 100%. However, other stats like bytes count and lines are not affected.
When using `indepth` mode, it is possible to hide languages per category.
Supported categories are `data`, `markup`, `programming` and `prose`.
*Example: hide data and prose languages from stats*
```yml
- uses: lowlighter/metrics@latest
with:
plugin_languages: yes
plugin_languages_categories: data, prose
plugin_languages_recent_categories: data, prose
```
## 🎨 Using custom colors
The plugin uses GitHub language colors, but it may be hard to distinguish them depending on which languages you use.
It is possible to use custom colors using `plugin_languages_colors` option.
The following syntaxes are supported:
- A predefined set from [colorsets.json](colorsets.json) *(support limited to 8 languages max)*
- `${language}:${color}` to change the color of a language *(case insensitive)*
- `${n}:${color}` to change the color of the the n-th language
Both hexadecimal and [named color](https://developer.mozilla.org/en-US/docs/Web/CSS/color_value) are supported.
*Example: using a predefined color set*
```yml
- uses: lowlighter/metrics@latest
with:
plugin_languages: yes
plugin_languages_colors: rainbow
plugin_languages_limit: 8
```
*Example: setting JavaScript to red, the first language to blue and the second one to `#ff00aa`*
```yml
- uses: lowlighter/metrics@latest
with:
plugin_languages: yes
plugin_languages_colors: javascript:red, 0:blue, 1:#ff00aa
```
## ✍️ Using custom languages name
This plugin is limited by [GitHub linguist](https://github.com/github/linguist) capabilities, meaning that some languages may be mislabeled in some cases.
To mitigate this, it is possible to use `plugin_languages_aliases` option and provide a list of overrides using the following syntax: `${language}:${alias}` *(case insensitive)*.
*Example: display JavaScript as JS and TypeScript as TS*
```yml
- uses: lowlighter/metrics@latest
with:
plugin_languages: yes
plugin_languages_aliases: javascript:JS typescript:TS
```
## Examples workflows
<!--examples-->
```yaml
name: Most used
uses: lowlighter/metrics@latest
with:
filename: metrics.plugin.languages.svg
token: ${{ secrets.METRICS_TOKEN }}
base: ''
plugin_languages: 'yes'
plugin_languages_ignored: >-
html, css, tex, less, dockerfile, makefile, qmake, lex, cmake, shell,
gnuplot
plugin_languages_limit: 4
```
```yaml
name: Most used (with details)
uses: lowlighter/metrics@latest
with:
filename: metrics.plugin.languages.details.svg
token: ${{ secrets.METRICS_TOKEN }}
base: ''
plugin_languages: 'yes'
plugin_languages_ignored: >-
html, css, tex, less, dockerfile, makefile, qmake, lex, cmake, shell,
gnuplot
plugin_languages_details: bytes-size, percentage
plugin_languages_limit: 4
```
```yaml
name: Recently used
uses: lowlighter/metrics@latest
with:
filename: metrics.plugin.languages.recent.svg
token: ${{ secrets.METRICS_TOKEN }}
base: ''
plugin_languages: 'yes'
plugin_languages_ignored: >-
html, css, tex, less, dockerfile, makefile, qmake, lex, cmake, shell,
gnuplot
plugin_languages_sections: recently-used
plugin_languages_details: bytes-size, percentage
plugin_languages_limit: 4
```
```yaml
name: Indepth analysis
uses: lowlighter/metrics@latest
with:
filename: metrics.plugin.languages.indepth.svg
token: ${{ secrets.METRICS_TOKEN }}
base: ''
plugin_languages: 'yes'
plugin_languages_ignored: >-
html, css, tex, less, dockerfile, makefile, qmake, lex, cmake, shell,
gnuplot
plugin_languages_indepth: 'yes'
plugin_languages_details: lines, bytes-size
plugin_languages_limit: 4
plugin_languages_analysis_timeout: 15
```
<!--/examples-->