PDF a11y: Use style info in the text-runs break heuristic

The text-runs break heuristic now considers the text style.
With this change, a text run will break if it encounters a difference
in the following text style properties:
* font-family
* font-weight
* rendering mode
* font-size
* fill color
* stroke color
* italic flag
* bold flag

Before this change, it used to break only if a font-size difference was
observed.

I added an "ultimate" dump test that tests for all of those properties.

Bug: 985604
Change-Id: Ice4b4a2dcdd8cc9642ca0e9d069a415c5b2700fd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1771987
Commit-Queue: Benjamin Beaudry <benjamin.beaudry@microsoft.com>
Reviewed-by: Lei Zhang <thestig@chromium.org>
Reviewed-by: Ian Prest <iapres@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#710522}

chrome/browser/pdf/pdf_extension_test.cc

@@ -2617,3 +2617,8 @@ IN_PROC_BROWSER_TEST_P(PDFExtensionAccessibilityTreeDumpTest,
                        LinksImagesAndText) {
   RunPDFTest(FILE_PATH_LITERAL("text-image-link.pdf"));
 }
+
+IN_PROC_BROWSER_TEST_P(PDFExtensionAccessibilityTreeDumpTest,
+                       TextRunStyleHeuristic) {
+  RunPDFTest(FILE_PATH_LITERAL("text-run-style-heuristic.pdf"));
+}

chrome/test/data/pdf/accessibility/text-format-expected-blink.txt

@@ -0,0 +1,15 @@
+embeddedObject fontFamily='Roboto' color=-16777216 fontSize=13.00 fontWeight=400.00
+++document fontFamily='Roboto' restriction=readOnly
+++++region fontFamily='Roboto' name='Page 1' restriction=readOnly isPageBreakingObject=true
+++++++paragraph fontFamily='Roboto' restriction=readOnly
+++++++++staticText fontFamily='Roboto' name='Case 1<newline>' restriction=readOnly
+++++++++++inlineTextBox fontFamily='Helvetica-Bold-Italic' name='Case 1<newline>' restriction=readOnly color=-65536 fontSize=38.00 fontWeight=600.00
+++++++paragraph fontFamily='Roboto' restriction=readOnly
+++++++++staticText fontFamily='Roboto' name='Case 2<newline>' restriction=readOnly
+++++++++++inlineTextBox fontFamily='Helvetica-Bold-Italic' name='Case 2<newline>' restriction=readOnly color=-16711936 fontSize=38.00 fontWeight=600.00
+++++++paragraph fontFamily='Roboto' restriction=readOnly
+++++++++staticText fontFamily='Roboto' name='Case 3<newline>' restriction=readOnly
+++++++++++inlineTextBox fontFamily='Helvetica-Bold-Italic' name='Case 3<newline>' restriction=readOnly color=-65536 fontSize=38.00 fontWeight=600.00
+++++++paragraph fontFamily='Roboto' restriction=readOnly
+++++++++staticText fontFamily='Roboto' name='Case 4' restriction=readOnly
+++++++++++inlineTextBox fontFamily='Helvetica-Bold-Italic' name='Case 4' restriction=readOnly fontSize=38.00 fontWeight=600.00

chrome/test/data/pdf/accessibility/text-run-style-heuristic-expected-auralinux.txt

@@ -0,0 +1,5 @@
+[embedded component]
+++[document frame]
+++++[landmark] name='Page 1'
+++++++[paragraph]
+++++++++[text] name='Hello world! One word equals to one text run.'

chrome/test/data/pdf/accessibility/text-run-style-heuristic-expected-blink.txt

@@ -0,0 +1,14 @@
+embeddedObject
+++document restriction=readOnly
+++++region name='Page 1' restriction=readOnly isPageBreakingObject=true
+++++++paragraph restriction=readOnly
+++++++++staticText name='Hello world! One word equals to one text run.' restriction=readOnly
+++++++++++inlineTextBox name='Hello ' restriction=readOnly
+++++++++++inlineTextBox name='world! ' restriction=readOnly
+++++++++++inlineTextBox name='One ' restriction=readOnly
+++++++++++inlineTextBox name='word ' restriction=readOnly
+++++++++++inlineTextBox name='equals ' restriction=readOnly
+++++++++++inlineTextBox name='to ' restriction=readOnly
+++++++++++inlineTextBox name='one ' restriction=readOnly
+++++++++++inlineTextBox name='text ' restriction=readOnly
+++++++++++inlineTextBox name='run.' restriction=readOnly

chrome/test/data/pdf/accessibility/text-run-style-heuristic-expected-mac.txt

@@ -0,0 +1,5 @@
+AXGroup AXDescription='Page 1'
+++AXGroup
+++++AXGroup AXDescription='Page 1'
+++++++AXGroup
+++++++++AXStaticText AXValue='Hello world! One word equals to one text run.'

chrome/test/data/pdf/accessibility/text-run-style-heuristic-expected-uia-win.txt

@@ -0,0 +1,5 @@
+group
+++document
+++++region Name='Page 1'
+++++++group
+++++++++description Name='Hello world! One word equals to one text run.'

chrome/test/data/pdf/accessibility/text-run-style-heuristic-expected-win.txt

@@ -0,0 +1,5 @@
+ROLE_SYSTEM_GROUPING FOCUSABLE
+++ROLE_SYSTEM_DOCUMENT READONLY FOCUSABLE
+++++IA2_ROLE_LANDMARK name='Page 1' READONLY
+++++++IA2_ROLE_PARAGRAPH READONLY
+++++++++ROLE_SYSTEM_STATICTEXT name='Hello world! One word equals to one text run.' READONLY

chrome/test/data/pdf/accessibility/text-run-style-heuristic.in

@@ -0,0 +1,143 @@
+{{header}}
+{{object 1 0}} <<
+  /Type /Catalog
+  /Pages 2 0 R
+>>
+endobj
+{{object 2 0}} <<
+  /Type /Pages
+  /MediaBox [ 0 0 200 200 ]
+  /Count 1
+  /Kids [ 3 0 R ]
+>>
+endobj
+{{object 3 0}} <<
+  /Type /Page
+  /Parent 2 0 R
+  /Resources <<
+    /Font <<
+      /F1 4 0 R %Basic font. Times-Roman, no style.
+      /F2 5 0 R %Basic font. Helvetica, no style.
+      /F3 6 0 R %F2 + font-weigth (/stemV and /Weight)
+      /F4 8 0 R %F3 + italic flag
+    >>
+  >>
+  /Contents 10 0 R
+>>
+endobj
+{{object 4 0}} <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Times-Roman
+>>
+endobj
+{{object 5 0}} <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Helvetica
+>>
+endobj
+{{object 6 0}} <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Helvetica-Bold
+  /FontDescriptor 7 0 R
+>>
+endobj
+{{object 7 0}} <<
+  /Type /FontDescriptor
+  /FontName /Helvetica-Bold
+  /FontWeight 600
+  /StemV 80
+  /ItalicAngle 0
+  /Ascent 776
+  /Flags 0
+  /FontBBox [-250 -236 2827 1000]
+  /CapHeight 763
+  /Descent -223
+>>
+endobj
+{{object 8 0}} <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Helvetica-Bold-Italic
+  /FontDescriptor 9 0 R
+>>
+endobj
+{{object 9 0}} <<
+  /Type /FontDescriptor
+  /FontName /Helvetica-Bold-Italic
+  /FontWeight 600
+  /StemV 80
+  /ItalicAngle 45
+  /Ascent 776
+  /Flags 64
+  /FontBBox [-250 -236 2827 1000]
+  /CapHeight 763
+  /Descent -223
+>>
+endobj
+{{object 10 0}} <<
+  {{streamlen}}
+>>
+stream
+BT
+0 0 Td
+/F1 9 Tf
+1 0 0 1 2 30 Tm
+(Hello) Tj
+
+0 0 Td
+/F2 9 Tf
+1 0 0 1 22 30 Tm
+( world!) Tj
+
+0 0 Td
+/F2 8 Tf
+1 0 0 1 48 30 Tm
+( One) Tj
+
+0 0 Td
+/F3 8 Tf
+1 0 0 1 65 30 Tm
+( word) Tj
+
+0 0 Td
+/F4 8 Tf
+1 0 0 1 85 30 Tm
+( equals) Tj
+
+0 0 Td
+/F4 8 Tf
+1 0 0 1 121 30 Tm
+3 Tr
+( to) Tj
+
+0 0 Td
+/F4 8 Tf
+1 0 0 1 137 30 Tm
+1 Tr
+( one) Tj
+
+0 0 Td
+/F4 8 Tf
+1 0 0 1 154 30 Tm
+1 0 0 rg
+0 1 0 RG
+2 Tr
+( text) Tj
+
+0 0 Td
+/F4 8 Tf
+1 0 0 1 170 30 Tm
+0 0 1 rg
+1 0 0 RG
+1 Tr
+( run.) Tj
+ET
+endstream
+endobj
+{{xref}}
+{{trailer}}
+{{startxref}}
+%%EOF

chrome/test/data/pdf/accessibility/text-run-style-heuristic.pdf

@@ -0,0 +1,160 @@
+%PDF-1.7
+%<25><><EFBFBD><EFBFBD>
+1 0 obj <<
+  /Type /Catalog
+  /Pages 2 0 R
+>>
+endobj
+2 0 obj <<
+  /Type /Pages
+  /MediaBox [ 0 0 200 200 ]
+  /Count 1
+  /Kids [ 3 0 R ]
+>>
+endobj
+3 0 obj <<
+  /Type /Page
+  /Parent 2 0 R
+  /Resources <<
+    /Font <<
+      /F1 4 0 R %Basic font. Times-Roman, no style.
+      /F2 5 0 R %Basic font. Helvetica, no style.
+      /F3 6 0 R %F2 + font-weigth (/stemV and /Weight)
+      /F4 8 0 R %F3 + italic flag
+    >>
+  >>
+  /Contents 10 0 R
+>>
+endobj
+4 0 obj <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Times-Roman
+>>
+endobj
+5 0 obj <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Helvetica
+>>
+endobj
+6 0 obj <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Helvetica-Bold
+  /FontDescriptor 7 0 R
+>>
+endobj
+7 0 obj <<
+  /Type /FontDescriptor
+  /FontName /Helvetica-Bold
+  /FontWeight 600
+  /StemV 80
+  /ItalicAngle 0
+  /Ascent 776
+  /Flags 0
+  /FontBBox [-250 -236 2827 1000]
+  /CapHeight 763
+  /Descent -223
+>>
+endobj
+8 0 obj <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Helvetica-Bold-Italic
+  /FontDescriptor 9 0 R
+>>
+endobj
+9 0 obj <<
+  /Type /FontDescriptor
+  /FontName /Helvetica-Bold-Italic
+  /FontWeight 600
+  /StemV 80
+  /ItalicAngle 45
+  /Ascent 776
+  /Flags 64
+  /FontBBox [-250 -236 2827 1000]
+  /CapHeight 763
+  /Descent -223
+>>
+endobj
+10 0 obj <<
+  /Length 469
+>>
+stream
+BT
+0 0 Td
+/F1 9 Tf
+1 0 0 1 2 30 Tm
+(Hello) Tj
+
+0 0 Td
+/F2 9 Tf
+1 0 0 1 22 30 Tm
+( world!) Tj
+
+0 0 Td
+/F2 8 Tf
+1 0 0 1 48 30 Tm
+( One) Tj
+
+0 0 Td
+/F3 8 Tf
+1 0 0 1 65 30 Tm
+( word) Tj
+
+0 0 Td
+/F4 8 Tf
+1 0 0 1 85 30 Tm
+( equals) Tj
+
+0 0 Td
+/F4 8 Tf
+1 0 0 1 121 30 Tm
+3 Tr
+( to) Tj
+
+0 0 Td
+/F4 8 Tf
+1 0 0 1 137 30 Tm
+1 Tr
+( one) Tj
+
+0 0 Td
+/F4 8 Tf
+1 0 0 1 154 30 Tm
+1 0 0 rg
+0 1 0 RG
+2 Tr
+( text) Tj
+
+0 0 Td
+/F4 8 Tf
+1 0 0 1 170 30 Tm
+0 0 1 rg
+1 0 0 RG
+1 Tr
+( run.) Tj
+ET
+endstream
+endobj
+xref
+0 11
+0000000000 65535 f 
+0000000015 00000 n 
+0000000068 00000 n 
+0000000161 00000 n 
+0000000463 00000 n 
+0000000541 00000 n 
+0000000617 00000 n 
+0000000722 00000 n 
+0000000934 00000 n 
+0000001046 00000 n 
+0000001267 00000 n 
+trailer <<
+  /Root 1 0 R
+  /Size 11
+>>
+startxref
+1789
+%%EOF

pdf/pdf_engine.h

@@ -376,9 +376,9 @@ class PDFEngine {
   // Get a given unicode character on a given page.
   virtual uint32_t GetCharUnicode(int page_index, int char_index) = 0;
   // Given a start char index, find the longest continuous run of text that's
-  // in a single direction and with the same style and font size. Return a
-  // filled out pp::PDF::PrivateAccessibilityTextRunInfo on success or
-  // base::nullopt on failure. e.g. When |start_char_index| is out of bounds.
+  // in a single direction and with the same text style. Return a filled out
+  // pp::PDF::PrivateAccessibilityTextRunInfo on success or base::nullopt on
+  // failure. e.g. When |start_char_index| is out of bounds.
   virtual base::Optional<pp::PDF::PrivateAccessibilityTextRunInfo>
   GetTextRunInfo(int page_index, int start_char_index) = 0;
   // Gets the number of links on a given page.

pdf/pdfium/pdfium_page.cc

@@ -326,6 +326,21 @@ void PDFiumPage::CalculateTextRunStyleInfo(
   style_info->render_mode = FPDFText_GetTextRenderMode(text_page, char_index);
 }
 
+bool PDFiumPage::AreTextStyleEqual(
+    int char_index,
+    const pp::PDF::PrivateAccessibilityTextStyleInfo& style) {
+  pp::PDF::PrivateAccessibilityTextStyleInfo char_style;
+  CalculateTextRunStyleInfo(char_index, &char_style);
+  return char_style.font_name == style.font_name &&
+         char_style.font_weight == style.font_weight &&
+         char_style.render_mode == style.render_mode &&
+         DoubleEquals(char_style.font_size, style.font_size) &&
+         char_style.fill_color == style.fill_color &&
+         char_style.stroke_color == style.stroke_color &&
+         char_style.is_italic == style.is_italic &&
+         char_style.is_bold == style.is_bold;
+}
+
 base::Optional<pp::PDF::PrivateAccessibilityTextRunInfo>
 PDFiumPage::GetTextRunInfo(int start_char_index) {
   FPDF_PAGE page = GetPage();
@@ -406,11 +421,9 @@ PDFiumPage::GetTextRunInfo(int start_char_index) {
         GetFloatCharRectInPixels(page, text_page, char_index);
 
     if (!base::IsUnicodeWhitespace(character)) {
-      // TODO (bebeaudr): add text run break heuristic to break on style.
-
-      // Heuristic: End the text run if different font size is encountered.
-      double font_size = FPDFText_GetFontSize(text_page, char_index);
-      if (!DoubleEquals(font_size, text_run_font_size))
+      // Heuristic: End the text run if the text style of the current character
+      // is different from the text run's style.
+      if (!AreTextStyleEqual(char_index, info.style))
         break;
 
       // Heuristic: End text run if character isn't going in the same direction.

pdf/pdfium/pdfium_page.h

@@ -201,6 +201,11 @@ class PDFiumPage {
   void CalculateTextRunStyleInfo(
       int char_index,
       pp::PDF::PrivateAccessibilityTextStyleInfo* style_info);
+  // Returns a boolean indicating if the character at index |char_index| has the
+  // same text style as the text run.
+  bool AreTextStyleEqual(
+      int char_index,
+      const pp::PDF::PrivateAccessibilityTextStyleInfo& style);
 
   // Key    :  Marked content id for the image element as specified in the
   //           struct tree.