From 3e328561fd883f74862c4dce064e90e34b7cbdfc Mon Sep 17 00:00:00 2001
From: taoqf <tao_qiufeng@126.com>
Date: Tue, 27 Oct 2020 13:47:08 +0800
Subject: [PATCH] add option blockTextElements #78

---
 README.md         |  9 +++++---
 src/nodes/html.ts | 53 ++++++++++++++++++++++++++++++-----------------
 test/html.js      |  2 +-
 test/pre.js       | 41 +++++++++++++++++++++++++++++++++++-
 4 files changed, 81 insertions(+), 24 deletions(-)

diff --git a/README.md b/README.md
index 2774977..22bddb0 100644
--- a/README.md
+++ b/README.md
@@ -74,10 +74,13 @@ Parse given data, and return root of the generated DOM.
   ```js
   {
     lowerCaseTagName: false,  // convert tag name to lower case (hurt performance heavily)
-    script: true,            // retrieve content in <script> (hurt performance slightly)
-    style: true,             // retrieve content in <style> (hurt performance slightly)
-    pre: true,               // retrieve content in <pre> (hurt performance slightly)
     comment: false            // retrieve comments (hurt performance slightly)
+    blockTextElements: {
+      script: true,	// keep text content when parsing
+      noscript: true,	// keep text content when parsing
+      style: true,		// keep text content when parsing
+      pre: true			// keep text content when parsing
+    }
   }
   ```
 
diff --git a/src/nodes/html.ts b/src/nodes/html.ts
index 1bbafd1..5221f69 100644
--- a/src/nodes/html.ts
+++ b/src/nodes/html.ts
@@ -648,23 +648,13 @@ const kElementsClosedByClosing = {
 	th: { tr: true, table: true, TR: true, TABLE: true },
 	TH: { tr: true, table: true, TR: true, TABLE: true }
 };
-const kBlockTextElements = {
-	script: true,
-	SCRIPT: true,
-	noscript: true,
-	NOSCRIPT: true,
-	style: true,
-	STYLE: true,
-	pre: true,
-	PRE: true
-};
 
 export interface Options {
 	lowerCaseTagName: boolean;
-	script: boolean;
-	style: boolean;
-	pre: boolean;
 	comment: boolean;
+	blockTextElements: {
+		[tag: string]: boolean;
+	};
 }
 
 const frameflag = 'documentfragmentcontainer';
@@ -675,10 +665,35 @@ const frameflag = 'documentfragmentcontainer';
  * @param  {string} data      html
  * @return {HTMLElement}      root element
  */
-export function parse(data: string, options?: Options): HTMLElement & { valid: boolean };
-export function parse(data: string, options?: Options & { noFix: false }): HTMLElement & { valid: boolean };
-export function parse(data: string, options?: Options & { noFix: true }): (HTMLElement | TextNode) & { valid: boolean };
-export function parse(data: string, options = { pre: true, style: true, script: true, lowerCaseTagName: false, comment: false } as Options & { noFix?: boolean }) {
+export function parse(data: string, options?: Partial<Options>): HTMLElement & { valid: boolean };
+export function parse(data: string, options?: Partial<Options> & { noFix: false }): HTMLElement & { valid: boolean };
+export function parse(data: string, options?: Partial<Options> & { noFix: true }): (HTMLElement | TextNode) & { valid: boolean };
+export function parse(data: string, options = { lowerCaseTagName: false, comment: false } as Partial<Options & { noFix: boolean }>) {
+	const elements = options.blockTextElements || {
+		script: true,
+		noscript: true,
+		style: true,
+		pre: true
+	};
+	const element_names = Object.keys(elements);
+	const kBlockTextElements = element_names.map((it) => {
+		return new RegExp(it, 'i');
+	});
+	const kIgnoreElements = element_names.filter((it) => {
+		return elements[it];
+	}).map((it) => {
+		return new RegExp(it, 'i');
+	});
+	function element_should_be_ignore(tag: string) {
+		return kIgnoreElements.some((it) => {
+			return it.test(tag);
+		});
+	}
+	function is_block_text_element(tag: string) {
+		return kBlockTextElements.some((it) => {
+			return it.test(tag);
+		});
+	}
 	const root = new HTMLElement(null, {});
 	let currentParent = root;
 	const stack = [root];
@@ -728,7 +743,7 @@ export function parse(data: string, options = { pre: true, style: true, script:
 			// https://github.com/taoqf/node-html-parser/issues/38
 			currentParent = currentParent.appendChild(new HTMLElement(match[2], attrs, match[3]));
 			stack.push(currentParent);
-			if (kBlockTextElements[match[2]]) {
+			if (is_block_text_element(match[2])) {
 				// a little test to find next </script> or </style> ...
 				const closeMarkup = `</${match[2]}>`;
 				const index = (() => {
@@ -737,7 +752,7 @@ export function parse(data: string, options = { pre: true, style: true, script:
 					}
 					return data.indexOf(closeMarkup, kMarkupPattern.lastIndex);
 				})();
-				if (options[match[2]]) {
+				if (element_should_be_ignore(match[2])) {
 					let text: string;
 					if (index === -1) {
 						// there is no matching ending for the text element.
diff --git a/test/html.js b/test/html.js
index 3a04886..8231b00 100644
--- a/test/html.js
+++ b/test/html.js
@@ -539,7 +539,7 @@ describe('HTML Parser', function () {
 			it('set content pre', function () {
 				const root = parseHTML(`<html><head></head><body></body></html>`);
 				const body = root.querySelector("body");
-				body.set_content(`<pre>this    is some    preformatted    text</pre>`, { pre: true });
+				body.set_content(`<pre>this    is some    preformatted    text</pre>`);
 				root.toString().should.eql('<html><head></head><body><pre>this    is some    preformatted    text</pre></body></html>')
 			});
 		});
diff --git a/test/pre.js b/test/pre.js
index d0a28d7..8ea275d 100644
--- a/test/pre.js
+++ b/test/pre.js
@@ -9,7 +9,46 @@ describe('pre tag', function () {
       </div>
     </div>
     `;
-		const root = parse(html);
+		const root = parse(html, {
+			blockTextElements: {
+				script: true,
+				noscript: true,
+				style: true,
+				pre: true
+			}
+		});
 		root.toString().should.eql(html);
 	});
+	it('should ignore pre tag', function () {
+		const html = `
+    <div class="language-python highlighter-rouge">
+      <div class="highlight"> <pre class="highlight"><code><span class="k">print</span><span class="p">(</span><span class="s">'hello'</span><span class="p">)</span><br><span class="n">i</span> <span class="o">=</span> <span class="n">i</span> <span class="o">+</span> <span class="mi">1</span><br></code></pre>
+      </div>
+    </div>
+    `;
+		const root = parse(html, {
+			blockTextElements: {
+				pre: false
+			}
+		});
+		root.toString().should.eql(`
+    <div class="language-python highlighter-rouge">
+      <div class="highlight"> <pre class="highlight"></pre>
+      </div>
+    </div>
+    `);
+	});
+	it('do not treat pre as text block element', function () {
+		const html = `<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="k">print</span><span class="p">(</span><span class="s">'hello'</span><span class="p">)</span><br><span class="n">i</span><span class="o">=</span><span class="n">i</span><span class="o">+</span><span class="mi">1</span><br></code></pre>
+      </div>
+    </div>
+    `;
+		const root = parse(html, {
+			blockTextElements: {}
+		});
+		const div = root.firstChild.firstChild;
+		const pre = div.firstChild;
+		const code = pre.firstChild;
+		code.childNodes.length.should.eql(11);
+	});
 });