feat: added tagged (accessible) PDFs option (#11182)

puppeteer · Oct 17, 2023 · 0316863 · 0316863
1 parent 443ee99
commit 0316863
Show file tree

Hide file tree

Showing 6 changed files with 50 additions and 9 deletions.
diff --git a/docs/api/puppeteer.pdfoptions.md b/docs/api/puppeteer.pdfoptions.md
@@ -29,5 +29,6 @@ export interface PDFOptions
 | preferCSSPageSize   | <code>optional</code> | boolean                                   | Give any CSS <code>@page</code> size declared in the page priority over what is declared in the <code>width</code> or <code>height</code> or <code>format</code> option.                                                                                                                                                                                                        | <code>false</code>, which will scale the content to fit the paper size.  |
 | printBackground     | <code>optional</code> | boolean                                   | Set to <code>true</code> to print background graphics.                                                                                                                                                                                                                                                                                                                          | <code>false</code>                                                       |
 | scale               | <code>optional</code> | number                                    | Scales the rendering of the web page. Amount must be between <code>0.1</code> and <code>2</code>.                                                                                                                                                                                                                                                                               | <code>1</code>                                                           |
+| tagged              | <code>optional</code> | boolean                                   | Generate tagged (accessible) PDF.                                                                                                                                                                                                                                                                                                                                               | <code>false</code>                                                       |
 | timeout             | <code>optional</code> | number                                    | Timeout in milliseconds. Pass <code>0</code> to disable timeout.                                                                                                                                                                                                                                                                                                                | <code>30_000</code>                                                      |
 | width               | <code>optional</code> | string \| number                          | Sets the width of paper. You can pass in a number or a string with a unit.                                                                                                                                                                                                                                                                                                      |                                                                          |
diff --git a/packages/puppeteer-core/src/api/Page.ts b/packages/puppeteer-core/src/api/Page.ts
@@ -2677,7 +2677,7 @@ export abstract class Page extends EventEmitter<PageEvents> {
     options: PDFOptions = {},
     lengthUnit: 'in' | 'cm' = 'in'
   ): ParsedPDFOptions {
-    const defaults = {
+    const defaults: Omit<ParsedPDFOptions, 'width' | 'height' | 'margin'> = {
       scale: 1,
       displayHeaderFooter: false,
       headerTemplate: '',
@@ -2688,6 +2688,7 @@ export abstract class Page extends EventEmitter<PageEvents> {
       preferCSSPageSize: false,
       omitBackground: false,
       timeout: 30000,
+      tagged: false,
     };
 
     let width = 8.5;
@@ -2714,15 +2715,13 @@ export abstract class Page extends EventEmitter<PageEvents> {
         convertPrintParameterToInches(options.margin?.right, lengthUnit) || 0,
     };
 
-    const output = {
+    return {
       ...defaults,
       ...options,
       width,
       height,
       margin,
     };
-
-    return output;
   }
 
   /**

diff --git a/packages/puppeteer-core/src/cdp/Page.ts b/packages/puppeteer-core/src/cdp/Page.ts
@@ -1140,6 +1140,7 @@ export class CdpPage extends Page {
       preferCSSPageSize,
       omitBackground,
       timeout: ms,
+      tagged: generateTaggedPDF,
     } = this._getPDFOptions(options);
 
     if (omitBackground) {
@@ -1164,6 +1165,7 @@ export class CdpPage extends Page {
         marginRight: margin.right,
         pageRanges,
         preferCSSPageSize,
+        generateTaggedPDF,
       }
     );
 

diff --git a/packages/puppeteer-core/src/common/PDFOptions.ts b/packages/puppeteer-core/src/common/PDFOptions.ts
@@ -166,6 +166,12 @@ export interface PDFOptions {
    * @defaultValue `false`
    */
   omitBackground?: boolean;
+  /**
+   * Generate tagged (accessible) PDF.
+   * @defaultValue `false`
+   * @experimental
+   */
+  tagged?: boolean;
   /**
    * Timeout in milliseconds. Pass `0` to disable timeout.
    * @defaultValue `30_000`

diff --git a/test/TestExpectations.json b/test/TestExpectations.json
@@ -1007,6 +1007,12 @@
     "parameters": ["chrome", "webDriverBiDi"],
     "expectations": ["PASS"]
   },
+  {
+    "testIdPattern": "[page.spec] Page Page.pdf can print to PDF with accessible",
+    "platforms": ["darwin", "linux", "win32"],
+    "parameters": ["webDriverBiDi"],
+    "expectations": ["SKIP"]
+  },
   {
     "testIdPattern": "[page.spec] Page Page.pdf should respect timeout",
     "platforms": ["darwin", "linux", "win32"],
@@ -3197,6 +3203,12 @@
     "parameters": ["cdp", "firefox"],
     "expectations": ["FAIL"]
   },
+  {
+    "testIdPattern": "[page.spec] Page Page.pdf can print to PDF with accessible",
+    "platforms": ["darwin", "linux", "win32"],
+    "parameters": ["cdp", "firefox"],
+    "expectations": ["SKIP"]
+  },
   {
     "testIdPattern": "[page.spec] Page Page.removeExposedFunction should work",
     "platforms": ["darwin", "linux", "win32"],

diff --git a/test/src/page.spec.ts b/test/src/page.spec.ts
@@ -1989,8 +1989,30 @@ describe('Page', function () {
       const outputFile = __dirname + '/../assets/output.pdf';
       await page.goto(server.PREFIX + '/pdf.html');
       await page.pdf({path: outputFile});
-      expect(fs.readFileSync(outputFile).byteLength).toBeGreaterThan(0);
-      fs.unlinkSync(outputFile);
+      try {
+        expect(fs.readFileSync(outputFile).byteLength).toBeGreaterThan(0);
+      } finally {
+        fs.unlinkSync(outputFile);
+      }
+    });
+
+    it('can print to PDF with accessible', async () => {
+      const {page, server} = await getTestState();
+
+      const outputFile = __dirname + '/../assets/output.pdf';
+      const outputFileAccessible =
+        __dirname + '/../assets/output-accessible.pdf';
+      await page.goto(server.PREFIX + '/pdf.html');
+      await page.pdf({path: outputFile});
+      await page.pdf({path: outputFileAccessible, tagged: true});
+      try {
+        expect(
+          fs.readFileSync(outputFileAccessible).byteLength
+        ).toBeGreaterThan(fs.readFileSync(outputFile).byteLength);
+      } finally {
+        fs.unlinkSync(outputFileAccessible);
+        fs.unlinkSync(outputFile);
+      }
     });
 
     it('can print to PDF and stream the result', async () => {
@@ -2009,9 +2031,8 @@ describe('Page', function () {
 
       await page.goto(server.PREFIX + '/pdf.html');
 
-      let error!: Error;
-      await page.pdf({timeout: 1}).catch(_error => {
-        return (error = _error);
+      const error = await page.pdf({timeout: 1}).catch(err => {
+        return err;
       });
       expect(error).toBeInstanceOf(TimeoutError);
     });