Allow HTML tags to be parsable

PaulioRandall · May 1, 2024 · 31e09e4 · 31e09e4
1 parent fc70bf6
commit 31e09e4
Show file tree

Hide file tree

Showing 5 changed files with 174 additions and 17 deletions.
diff --git a/package-lock.json b/package-lock.json
diff --git a/package.json b/package.json
@@ -4,7 +4,7 @@
 	"description": "Simple tool for adding parsable metadata as comments within Svelte components.",
 	"type": "module",
 	"license": "MIT",
-	"version": "0.3.0",
+	"version": "0.4.0",
 	"engines": {
 		"node": ">=18"
 	},

diff --git a/src/parser.js b/src/parser.js
@@ -6,10 +6,17 @@ const defaultGlobOptions = {
 	nodir: true,
 }
 
-export const newNodeRegExp = (prefix = 'p23') => {
+export const newJavaScriptNodeRegexp = (prefix = 'p23') => {
 	return RegExp(
 		`^[^S\r\n]*//${prefix}((?:.[$a-zA-Z_][$a-zA-Z_0-9]*)+):(.*)$`,
-		'mgi'
+		'igm'
+	)
+}
+
+export const newHtmlNodeRegexp = (prefix = 'p23') => {
+	return RegExp(
+		`^[^S\r\n]*<!--${prefix}((?:.[$a-zA-Z_][$a-zA-Z_0-9]*)+):(.*?)-->`,
+		'igms'
 	)
 }
 
@@ -64,9 +71,25 @@ const extractNodes = (data, options) => {
 	// Examples:
 	//p23.name: Abc
 	//p23.group.name: Abc
+	const jsRegexp = newJavaScriptNodeRegexp(options.prefix)
+	const jsNodes = extractNodesWithRegexp(data, jsRegexp)
 
+	// Examples:
+	//<!--p23.name: Abc-->
+	//<!--p23.group.name: Abc-->
+	//<!--p23.group.name:
+	//  Abc
+	//-->
+	const htmlRegexp = newHtmlNodeRegexp(options.prefix)
+	const htmlNodes = extractNodesWithRegexp(data, htmlRegexp)
+
+	const n = joinNodeTrees(jsNodes, htmlNodes)
+	console.log(n)
+	return n
+}
+
+const extractNodesWithRegexp = (data, regexp) => {
 	const nodes = {}
-	const regexp = newNodeRegExp(options.prefix)
 	let next = null
 
 	while ((next = regexp.exec(data)) !== null) {
@@ -80,6 +103,24 @@ const extractNodes = (data, options) => {
 	return nodes
 }
 
+const joinNodeTrees = (a, b) => {
+	if (typeof a !== 'object') {
+		return structuredClone(b)
+	}
+
+	const nodes = structuredClone(a)
+
+	for (const name in b) {
+		if (typeof b[name] === 'object') {
+			nodes[name] = joinNodeTrees(nodes[name], b[name])
+		} else {
+			nodes[name] = b[name]
+		}
+	}
+
+	return nodes
+}
+
 const parseNodeNames = (nodeNames) => {
 	return nodeNames.split('.').filter((s) => !!s)
 }

diff --git a/src/parser.test.js b/src/parser.test.js
@@ -1,5 +1,5 @@
 import upath from 'upath'
-import { parse, newNodeRegExp } from './parser.js'
+import { parse, newJavaScriptNodeRegexp, newHtmlNodeRegexp } from './parser.js'
 
 const testdataDir = './src/testdata'
 
@@ -24,15 +24,15 @@ const parseToUnix = (f) => {
 }
 
 describe('p23', () => {
-	describe('newNodeRegExp', () => {
+	describe('newJavaScriptNodeRegexp', () => {
 		test('Non-nested node', () => {
-			const act = newNodeRegExp().exec('//p23.name:value')
+			const act = newJavaScriptNodeRegexp().exec('//p23.name:value')
 			const exp = expect.arrayContaining(['//p23.name:value', '.name', 'value'])
 			expect(act).toEqual(exp)
 		})
 
 		test('Custom prefix', () => {
-			const act = newNodeRegExp('my_custom_prefix').exec(
+			const act = newJavaScriptNodeRegexp('my_custom_prefix').exec(
 				'//my_custom_prefix.name:value'
 			)
 			const exp = expect.arrayContaining([
@@ -44,7 +44,7 @@ describe('p23', () => {
 		})
 
 		test('Nested node', () => {
-			const act = newNodeRegExp().exec('//p23.group.name:value')
+			const act = newJavaScriptNodeRegexp().exec('//p23.group.name:value')
 			const exp = expect.arrayContaining([
 				'//p23.group.name:value',
 				'.group.name',
@@ -54,7 +54,7 @@ describe('p23', () => {
 		})
 
 		test('Messed up but valid node names', () => {
-			const act = newNodeRegExp().exec(
+			const act = newJavaScriptNodeRegexp().exec(
 				'//p23.$$$12313___.__dsfjk12$$6389__$$:value'
 			)
 			const exp = expect.arrayContaining([
@@ -66,28 +66,100 @@ describe('p23', () => {
 		})
 
 		test('Empty value still returns entry', () => {
-			const act = newNodeRegExp().exec('//p23.name:')
+			const act = newJavaScriptNodeRegexp().exec('//p23.name:')
 			const exp = expect.arrayContaining(['//p23.name:', '.name', ''])
 			expect(act).toEqual(exp)
 		})
 
 		test('Fails if number first in node name', () => {
-			const act = newNodeRegExp().exec('//p23.1name:value')
+			const act = newJavaScriptNodeRegexp().exec('//p23.1name:value')
 			expect(act).toEqual(null)
 		})
 
 		test('Fails if missing node name', () => {
-			const act = newNodeRegExp().exec('//p23:value')
+			const act = newJavaScriptNodeRegexp().exec('//p23:value')
 			expect(act).toEqual(null)
 		})
 
 		test('Fails if missing node name (version two)', () => {
-			const act = newNodeRegExp().exec('//p23.:value')
+			const act = newJavaScriptNodeRegexp().exec('//p23.:value')
 			expect(act).toEqual(null)
 		})
 
 		test('Fails if missing colon', () => {
-			const act = newNodeRegExp().exec('//p23.name')
+			const act = newJavaScriptNodeRegexp().exec('//p23.name')
+			expect(act).toEqual(null)
+		})
+	})
+
+	describe('newHtmlNodeRegexp', () => {
+		test('Non-nested node', () => {
+			const act = newHtmlNodeRegexp().exec('<!--p23.name:value-->')
+			const exp = expect.arrayContaining([
+				'<!--p23.name:value-->',
+				'.name',
+				'value',
+			])
+			expect(act).toEqual(exp)
+		})
+
+		test('Custom prefix', () => {
+			const act = newHtmlNodeRegexp('my_custom_prefix').exec(
+				'<!--my_custom_prefix.name:value-->'
+			)
+			const exp = expect.arrayContaining([
+				'<!--my_custom_prefix.name:value-->',
+				'.name',
+				'value',
+			])
+			expect(act).toEqual(exp)
+		})
+
+		test('Nested node', () => {
+			const act = newHtmlNodeRegexp().exec('<!--p23.group.name:value-->')
+			const exp = expect.arrayContaining([
+				'<!--p23.group.name:value-->',
+				'.group.name',
+				'value',
+			])
+			expect(act).toEqual(exp)
+		})
+
+		test('Messed up but valid node names', () => {
+			const act = newHtmlNodeRegexp().exec(
+				'<!--p23.$$$12313___.__dsfjk12$$6389__$$:value-->'
+			)
+			const exp = expect.arrayContaining([
+				'<!--p23.$$$12313___.__dsfjk12$$6389__$$:value-->',
+				'.$$$12313___.__dsfjk12$$6389__$$',
+				'value',
+			])
+			expect(act).toEqual(exp)
+		})
+
+		test('Empty value still returns entry', () => {
+			const act = newHtmlNodeRegexp().exec('<!--p23.name:-->')
+			const exp = expect.arrayContaining(['<!--p23.name:-->', '.name', ''])
+			expect(act).toEqual(exp)
+		})
+
+		test('Fails if number first in node name', () => {
+			const act = newHtmlNodeRegexp().exec('<!--p23.1name:value-->')
+			expect(act).toEqual(null)
+		})
+
+		test('Fails if missing node name', () => {
+			const act = newHtmlNodeRegexp().exec('<!--p23:value-->')
+			expect(act).toEqual(null)
+		})
+
+		test('Fails if missing node name (version two)', () => {
+			const act = newHtmlNodeRegexp().exec('<!--p23.:value-->')
+			expect(act).toEqual(null)
+		})
+
+		test('Fails if missing colon', () => {
+			const act = newHtmlNodeRegexp().exec('<!--p23.name-->')
 			expect(act).toEqual(null)
 		})
 	})
@@ -232,5 +304,32 @@ describe('p23', () => {
 				},
 			])
 		})
+
+		test('Parses with HTML docs', () => {
+			const file = createSvelteFilePath('htmlDocs')
+			const metadata = parse(file)
+
+			expect(metadata).toEqual([
+				{
+					...generateFileFields(file),
+					nodes: {
+						type: 'Music',
+						music: {
+							type: 'Band',
+							band: {
+								name: 'Rhapsody of Fire',
+								genre: 'Symphonic Power Metal',
+								albums: `
+	[
+		"Rhapsody of Fire",
+		"Triumph or Agony",
+	]
+`,
+							},
+						},
+					},
+				},
+			])
+		})
 	})
 })
diff --git a/src/testdata/files/htmlDocs.svelte b/src/testdata/files/htmlDocs.svelte
@@ -0,0 +1,17 @@
+<script>
+	//p23.type:Music
+	//p23.music.type:Band
+</script>
+
+<!--p23.music.band.genre:Symphonic Power Metal-->
+
+<div>
+	<!--p23.music.band.name:Rhapsody of Fire-->
+</div>
+
+<!--p23.music.band.albums:
+	[
+		"Rhapsody of Fire",
+		"Triumph or Agony",
+	]
+-->