The technical document curation service for you.

Easy switching between languages.

• Get Extension from Chrome Web Store - v1.0.1

Check the list of sitefile.

We've just started the project. Pls let us know which document you need!

• Comment on the issue
• Send pull request to this repo

The sitefile of the original document - typically written in English - would be like below. This example is a sitefile for Bootstrap.

{
  "id": "bootstrap",
  "title": "Bootstrap · The world's most popular mobile-first and responsive front-end framework.",
  "url": "http://getbootstrap.com",
  "language": "en",
  "contribute": "https://github.com/twbs/bootstrap",
  "version": "3.3.*"
}

• id: The ID in TechDocs. It must be unique.
• title: The title of the document.
• url: The URL of the document. No trailing slash needed.
• language: The language the document written in. en fr ja cn kr...
• contribute: The URL for contributing to the document. Typically a GitHub repo.
• version: The version of the product.

Translated documents have two more attributes in sitefile.

{
  "id": "bootstrap-ja",
  "title": "Bootstrap · The world's most popular mobile-first and responsive front-end framework.",
  "url": "http://www.wivern.com/bootstrap",
  "language": "ja",
  "contribute": "https://twitter.com/smokyjp",
  "version": "3.2.*",
  "origin": "bootstrap",
  "rules": [
    "getting-started/ > getting-started.html",
    "css/ > css.html",
    "components/ > components.html"
  ]
}

• origin: The ID of the original document.
• rules: The list of pairs of URL. "[original] > [translated]"

We can use wildcards in rules section.

"rules": [
  "getting-started/ > getting-started.html",
  "css/ > css.html",
  "components/ > components.html"
]

The part of sitefile above would be rewritten like below.

"rules": [
  "*/ > *.html"
]

A wildcard * will be replaced into RegExp (.+?) internally. The code below is a pseudo code for explanation.

var url1, url2;
var original = 'css/';
var translated = 'css.html';

url1 = original.replace(new RegExp('^(.+?)/$'), '$1.html');
// url1 == translated

url2 = translated.replace(new RegExp('^(.+?).html$'), '$1/');
// url2 == original

As above, wildcards are bidirectionally used - original to translate and vice versa.

Sometimes, URLs need more complicated matching. For example, the case below needs camelCase <--> snake_case conversion.

The (part of) sitefile would be like below.

"rules": [
  "ng/* > ng_* | camel2snake"
]

The pipe | can connect single of multiple filters.

• camel2snake: convert camelCase to snake_case
• snake2camel: convert snake_case to camelCase
• dot2snake: convert dot.case to snake_case
• snake2dot: convert snake_case to dot.case
• replace: convert letters to others

Wildcards and filters are useful, but it's possible to break the reversibility of URL conversion. Please note that we have to match original <--> translated.

Basically we should keep reversibility. If it couldn't, we have reverse match. The next two are equivalent.

"rules": [
  "ng/* > ng_* | camel2snake",
  "ng/* < ng_* | snake2camel"
]