Update the list indenting to match GOV.UK style

Since we now handle the list markdown rather than turndown we have the
opportunity to update the indenting to be more consistent with GOV.UK
style. This sets the lists to have just one space after a marker and 3
spaces before it.

I wrestled for a bit trying to work out what should be the correct level
of indentation. Some places advise 2 characters, others 3 and others 4.
If you are working in monotype fonts it's most readable to have 2 for
ul's and 3 for ol's. As we expect users not to be typing in monotype
fonts this isn't too big a concern and 3 seems to offer a nice balance
of readability and without special cases for each list type.

src/to-govspeak.js

@@ -1,7 +1,8 @@
 import TurndownService from 'turndown'
 
 const service = new TurndownService({
-  bulletListMarker: '-'
+  bulletListMarker: '-',
+  listIndent: '   ' // 3 spaces
 })
 
 // As a user may have pasted markdown we rather crudley
@@ -125,37 +126,38 @@ service.addRule('invalidNestedLists', {
       return previousNodeName === 'li'
     }
   },
-  replacement: (content, node) => {
+  replacement: (content, node, options) => {
     content = content
       .replace(/^\n+/, '') // remove leading newlines
       .replace(/\n+$/, '') // replace trailing newlines
-      .replace(/\n/gm, '\n    ') // indent
+      .replace(/\n/gm, `\n${options.listIndent}`) // indent all nested content in the list
 
-    return '    ' + content + '\n'
+    // indent this list to match sibling
+    return options.listIndent + content + '\n'
   }
 })
 
 // This is ported from https://github.com/domchristie/turndown/blob/80297cebeae4b35c8d299b1741b383c74eddc7c1/src/commonmark-rules.js#L61-L80
-// It is modified to handle items from the invalidNestedLists as if these
-// are ordered lists they will be output with incorrect numbering. It also
-// removed handling of start attribute of <ol> elements as this has no effect
-// on govspeak.
+// It is modified in the following ways:
+// - Only determines ol ordering based on li elements
+// - Removes handling of ol start attribute as this doesn't affect govspeak output
+// - Makes spacing consistent with gov.uk markdown guidance
 service.addRule('listItems', {
   filter: 'li',
   replacement: function (content, node, options) {
     content = content
       .replace(/^\n+/, '') // remove leading newlines
       .replace(/\n+$/, '\n') // replace trailing newlines with just a single one
-      .replace(/\n/gm, '\n    ') // indent
+      .replace(/\n/gm, `\n${options.listIndent}`) // indent all nested content in the list
 
-    let prefix = options.bulletListMarker + '   '
+    let prefix = options.bulletListMarker + ' '
     const parent = node.parentNode
     if (parent.nodeName.toLowerCase() === 'ol') {
       const listItems = Array.prototype.filter.call(
         parent.children, (element) => element.nodeName.toLowerCase() === 'li'
       )
       const index = Array.prototype.indexOf.call(listItems, node)
-      prefix = (index + 1).toString() + '.  '
+      prefix = (index + 1).toString() + '. '
     }
     return (
       prefix + content + (node.nextSibling && !/\n$/.test(content) ? '\n' : '')
@@ -178,7 +180,7 @@ function removeBrParagraphs (govspeak) {
 function extractHeadingsFromLists (govspeak) {
   // This finds instances of headings within ordered lists and replaces them
   // with the headings only. This only applies to H2 and H3.
-  const headingsInListsRegExp = new RegExp(/\d\.\s{2}(#{2,3})/, 'g')
+  const headingsInListsRegExp = new RegExp(/\d\.\s(\#{2,3})/, 'g')
   return govspeak.replace(headingsInListsRegExp, '$1')
 }
 

test/to-govspeak.test.js

@@ -27,7 +27,7 @@ it('converts lists to a dash bullet style', () => {
       <li>Item 2</li>
     </ul>
   `
-  expect(toGovspeak(html)).toEqual('-   Item 1\n-   Item 2')
+  expect(toGovspeak(html)).toEqual('- Item 1\n- Item 2')
 })
 
 it('maintains H2 and H3 headers', () => {
@@ -110,7 +110,7 @@ it('extracts headers from lists', () => {
       <li><h3>Item 2</h3></li>
     </ol>
   `
-  expect(toGovspeak(html)).toEqual('## Item 1\n    \n### Item 2')
+  expect(toGovspeak(html)).toEqual('## Item 1\n   \n### Item 2')
 })
 
 it('strips paragraph elements within a list item', () => {
@@ -120,7 +120,7 @@ it('strips paragraph elements within a list item', () => {
       <li><p>Item 2</p></li>
     </ul>
   `
-  expect(toGovspeak(html)).toEqual('-   Item 1\n-   Item 2')
+  expect(toGovspeak(html)).toEqual('- Item 1\n- Item 2')
 })
 
 it('removes nested links when link markdown text is wrapped in an element', () => {
@@ -151,10 +151,10 @@ it('fixes an invalid nested unordered list that Google Docs produces', () => {
     </ul>
   `
   expect(toGovspeak(html)).toEqual(
-    '-   Parent\n' +
-    '    -   Child\n' +
-    '        -   Grand child\n' +
-    '-   Parent sibling'
+    '- Parent\n' +
+    '   - Child\n' +
+    '      - Grand child\n' +
+    '- Parent sibling'
   )
 })
 
@@ -175,12 +175,12 @@ it('fixes an invalid nested ordered list that Google Docs produces', () => {
     </ol>
   `
   expect(toGovspeak(html)).toEqual(
-    '1.  Parent\n' +
-    '    1.  Child 1\n' +
-    '        1.  Grand child 1\n' +
-    '        2.  Grand child 2\n' +
-    '    2.  Child 2\n' +
-    '    3.  Child 3\n' +
-    '2.  Parent sibling'
+    '1. Parent\n' +
+    '   1. Child 1\n' +
+    '      1. Grand child 1\n' +
+    '      2. Grand child 2\n' +
+    '   2. Child 2\n' +
+    '   3. Child 3\n' +
+    '2. Parent sibling'
   )
 })