Pdf a11y: Populate AXNode with new text style info

This change fills the font-family, font-size, font-weight, italic, bold
and color attributes of the AXNode with exposed values from the PDFium
APIs.

I've also added a PDF accessibility tree dump test using the filter
properties to dump those attributes in the expected file. The test file
is "text-style.pdf".

Somehow, I added in a previous change text-format-expected-blink.txt.
This file should have never been committed. I removed it in this CL.

Bug: 985604
Change-Id: I70a4729269aea69e184e61710d7977425b431704
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1755412
Reviewed-by: Lei Zhang <thestig@chromium.org>
Reviewed-by: Tom Sepez <tsepez@chromium.org>
Reviewed-by: Bill Budge <bbudge@chromium.org>
Reviewed-by: Ian Prest <iapres@microsoft.com>
Commit-Queue: Lei Zhang <thestig@chromium.org>
Cr-Commit-Position: refs/heads/master@{#711942}

chrome/browser/pdf/pdf_extension_test.cc

@@ -2622,3 +2622,7 @@ IN_PROC_BROWSER_TEST_P(PDFExtensionAccessibilityTreeDumpTest,
                        TextRunStyleHeuristic) {
   RunPDFTest(FILE_PATH_LITERAL("text-run-style-heuristic.pdf"));
 }
+
+IN_PROC_BROWSER_TEST_P(PDFExtensionAccessibilityTreeDumpTest, TextStyle) {
+  RunPDFTest(FILE_PATH_LITERAL("text-style.pdf"));
+}

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

@@ -1,15 +0,0 @@
-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-style-expected-auralinux.txt

@@ -0,0 +1,11 @@
+[embedded component]
+++[document frame]
+++++[landmark] name='Page 1'
+++++++[paragraph]
+++++++++[text] name='Case 1<newline>'
+++++++[paragraph]
+++++++++[text] name='Case 2<newline>'
+++++++[paragraph]
+++++++++[text] name='Case 3<newline>'
+++++++[paragraph]
+++++++++[text] name='Case 4'

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

@@ -0,0 +1,15 @@
+embeddedObject color=-16777216
+++document restriction=readOnly
+++++region name='Page 1' restriction=readOnly isPageBreakingObject=true
+++++++paragraph restriction=readOnly
+++++++++staticText name='Case 1<newline>' restriction=readOnly
+++++++++++inlineTextBox fontFamily='Helvetica-Bold-Italic' name='Case 1<newline>' restriction=readOnly color=-65536 fontSize=38.00 fontWeight=800.00
+++++++paragraph restriction=readOnly
+++++++++staticText name='Case 2<newline>' restriction=readOnly
+++++++++++inlineTextBox fontFamily='Helvetica-Bold-Italic' name='Case 2<newline>' restriction=readOnly color=-16711936 fontSize=38.00 fontWeight=800.00
+++++++paragraph restriction=readOnly
+++++++++staticText name='Case 3<newline>' restriction=readOnly
+++++++++++inlineTextBox fontFamily='Helvetica-Bold-Italic' name='Case 3<newline>' restriction=readOnly color=-65536 fontSize=38.00 fontWeight=800.00
+++++++paragraph restriction=readOnly
+++++++++staticText name='Case 4' restriction=readOnly
+++++++++++inlineTextBox fontFamily='Helvetica-Bold-Italic' name='Case 4' restriction=readOnly fontSize=38.00 fontWeight=800.00

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

@@ -0,0 +1,11 @@
+AXGroup AXDescription='Page 1'
+++AXGroup
+++++AXGroup AXDescription='Page 1'
+++++++AXGroup
+++++++++AXStaticText AXValue='Case 1<newline>'
+++++++AXGroup
+++++++++AXStaticText AXValue='Case 2<newline>'
+++++++AXGroup
+++++++++AXStaticText AXValue='Case 3<newline>'
+++++++AXGroup
+++++++++AXStaticText AXValue='Case 4'

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

@@ -0,0 +1,11 @@
+group
+++document
+++++region Name='Page 1'
+++++++group
+++++++++description Name='Case 1<newline>'
+++++++group
+++++++++description Name='Case 2<newline>'
+++++++group
+++++++++description Name='Case 3<newline>'
+++++++group
+++++++++description Name='Case 4'

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

@@ -0,0 +1,11 @@
+ROLE_SYSTEM_GROUPING FOCUSABLE
+++ROLE_SYSTEM_DOCUMENT READONLY FOCUSABLE
+++++IA2_ROLE_LANDMARK name='Page 1' READONLY
+++++++IA2_ROLE_PARAGRAPH READONLY
+++++++++ROLE_SYSTEM_STATICTEXT name='Case 1<newline>' READONLY
+++++++IA2_ROLE_PARAGRAPH READONLY
+++++++++ROLE_SYSTEM_STATICTEXT name='Case 2<newline>' READONLY
+++++++IA2_ROLE_PARAGRAPH READONLY
+++++++++ROLE_SYSTEM_STATICTEXT name='Case 3<newline>' READONLY
+++++++IA2_ROLE_PARAGRAPH READONLY
+++++++++ROLE_SYSTEM_STATICTEXT name='Case 4' READONLY

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

@@ -0,0 +1,92 @@
+{{header}}
+%@BLINK-ALLOW:fontFamily='Helvetica-Bold-Italic'
+%@BLINK-ALLOW:fontSize=38.00
+%@BLINK-ALLOW:fontWeight=800.00
+%@BLINK-ALLOW:color=*
+{{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
+    >>
+  >>
+  /Contents 6 0 R
+>>
+endobj
+{{object 4 0}} <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Helvetica-Bold-Italic
+  /FontDescriptor 5 0 R
+>>
+{{object 5 0}} <<
+  /Type /FontDescriptor
+  /FontName /Helvetica-Bold-Italic
+  /FontWeight 800
+  /StemV 165
+  /ItalicAngle 45
+  /Ascent 776
+  /Flags 64
+  /FontBBox [ -250 -236 2827 1000 ]
+  /CapHeight 763
+  /Descent -223
+>>
+endobj
+{{object 6 0}} <<
+  {{streamlen}}
+>>
+stream
+BT
+20 150 Td
+/F1 38 Tf
+0 1 0 RG
+1 0 0 rg
+0 Tr
+(Case 1) Tj
+ET
+
+BT
+20 110 Td
+/F1 38 Tf
+0 1 0 RG
+1 0 0 rg
+1 Tr
+(Case 2) Tj
+ET
+
+BT
+20 70 Td
+/F1 38 Tf
+0 1 0 RG
+1 0 0 rg
+2 Tr
+(Case 3) Tj
+ET
+
+BT
+20 30 Td
+/F1 38 Tf
+0 1 0 RG
+1 0 0 rg
+3 Tr
+(Case 4) Tj
+ET
+endstream
+endobj
+{{xref}}
+{{trailer}}
+{{startxref}}
+%%EOF

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

@@ -0,0 +1,105 @@
+%PDF-1.7
+%<25><><EFBFBD><EFBFBD>
+%@BLINK-ALLOW:fontFamily='Helvetica-Bold-Italic'
+%@BLINK-ALLOW:fontSize=38.00
+%@BLINK-ALLOW:fontWeight=800.00
+%@BLINK-ALLOW:color=*
+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
+    >>
+  >>
+  /Contents 6 0 R
+>>
+endobj
+4 0 obj <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Helvetica-Bold-Italic
+  /FontDescriptor 5 0 R
+>>
+5 0 obj <<
+  /Type /FontDescriptor
+  /FontName /Helvetica-Bold-Italic
+  /FontWeight 800
+  /StemV 165
+  /ItalicAngle 45
+  /Ascent 776
+  /Flags 64
+  /FontBBox [ -250 -236 2827 1000 ]
+  /CapHeight 763
+  /Descent -223
+>>
+endobj
+6 0 obj <<
+  /Length 245
+>>
+stream
+BT
+20 150 Td
+/F1 38 Tf
+0 1 0 RG
+1 0 0 rg
+0 Tr
+(Case 1) Tj
+ET
+
+BT
+20 110 Td
+/F1 38 Tf
+0 1 0 RG
+1 0 0 rg
+1 Tr
+(Case 2) Tj
+ET
+
+BT
+20 70 Td
+/F1 38 Tf
+0 1 0 RG
+1 0 0 rg
+2 Tr
+(Case 3) Tj
+ET
+
+BT
+20 30 Td
+/F1 38 Tf
+0 1 0 RG
+1 0 0 rg
+3 Tr
+(Case 4) Tj
+ET
+endstream
+endobj
+xref
+0 7
+0000000000 65535 f 
+0000000147 00000 n 
+0000000200 00000 n 
+0000000293 00000 n 
+0000000419 00000 n 
+0000000524 00000 n 
+0000000748 00000 n 
+trailer <<
+  /Root 1 0 R
+  /Size 7
+>>
+startxref
+1045
+%%EOF

components/pdf/renderer/pdf_accessibility_tree.cc

@@ -206,6 +206,30 @@ bool IsObjectInTextRun(const std::vector<T>& objects,
           objects[object_index].text_run_index <= text_run_index);
 }
 
+bool IsTextRenderModeFill(const PP_TextRenderingMode& mode) {
+  switch (mode) {
+    case PP_TEXTRENDERINGMODE_FILL:
+    case PP_TEXTRENDERINGMODE_FILLSTROKE:
+    case PP_TEXTRENDERINGMODE_FILLCLIP:
+    case PP_TEXTRENDERINGMODE_FILLSTROKECLIP:
+      return true;
+    default:
+      return false;
+  }
+}
+
+bool IsTextRenderModeStroke(const PP_TextRenderingMode& mode) {
+  switch (mode) {
+    case PP_TEXTRENDERINGMODE_STROKE:
+    case PP_TEXTRENDERINGMODE_FILLSTROKE:
+    case PP_TEXTRENDERINGMODE_STROKECLIP:
+    case PP_TEXTRENDERINGMODE_FILLSTROKECLIP:
+      return true;
+    default:
+      return false;
+  }
+}
+
 }  // namespace
 
 PdfAccessibilityTree::PdfAccessibilityTree(content::RendererPpapiHost* host,
@@ -699,6 +723,24 @@ ui::AXNodeData* PdfAccessibilityTree::CreateInlineTextBoxNode(
                                            chars_utf8);
   inline_text_box_node->AddIntAttribute(ax::mojom::IntAttribute::kTextDirection,
                                         text_run.direction);
+  inline_text_box_node->AddStringAttribute(
+      ax::mojom::StringAttribute::kFontFamily, text_run.style.font_name);
+  inline_text_box_node->AddFloatAttribute(ax::mojom::FloatAttribute::kFontSize,
+                                          text_run.style.font_size);
+  inline_text_box_node->AddFloatAttribute(
+      ax::mojom::FloatAttribute::kFontWeight, text_run.style.font_weight);
+  if (text_run.style.is_italic)
+    inline_text_box_node->AddTextStyle(ax::mojom::TextStyle::kItalic);
+  if (text_run.style.is_bold)
+    inline_text_box_node->AddTextStyle(ax::mojom::TextStyle::kBold);
+  if (IsTextRenderModeFill(text_run.style.render_mode)) {
+    inline_text_box_node->AddIntAttribute(ax::mojom::IntAttribute::kColor,
+                                          text_run.style.fill_color);
+  } else if (IsTextRenderModeStroke(text_run.style.render_mode)) {
+    inline_text_box_node->AddIntAttribute(ax::mojom::IntAttribute::kColor,
+                                          text_run.style.stroke_color);
+  }
+
   inline_text_box_node->relative_bounds.bounds =
       ToGfxRectF(text_run.bounds) + page_bounds.OffsetFromOrigin();
   std::vector<int32_t> char_offsets =

pdf/pdfium/pdfium_assert_matching_enums.cc

@@ -10,6 +10,7 @@
 #include "third_party/pdfium/public/fpdf_edit.h"
 #include "third_party/pdfium/public/fpdf_fwlevent.h"
 #include "third_party/pdfium/public/fpdf_sysfontinfo.h"
+#include "third_party/pdfium/public/fpdfview.h"
 #include "ui/events/keycodes/keyboard_codes.h"
 
 #define STATIC_ASSERT_ENUM(a, b)                            \
@@ -214,6 +215,20 @@ STATIC_ASSERT_ENUM(PP_PRIVATEDUPLEXMODE_SIMPLEX, Simplex);
 STATIC_ASSERT_ENUM(PP_PRIVATEDUPLEXMODE_SHORT_EDGE, DuplexFlipShortEdge);
 STATIC_ASSERT_ENUM(PP_PRIVATEDUPLEXMODE_LONG_EDGE, DuplexFlipLongEdge);
 
+STATIC_ASSERT_ENUM(PP_TEXTRENDERINGMODE_FILL, FPDF_TEXTRENDERMODE_FILL);
+STATIC_ASSERT_ENUM(PP_TEXTRENDERINGMODE_STROKE, FPDF_TEXTRENDERMODE_STROKE);
+STATIC_ASSERT_ENUM(PP_TEXTRENDERINGMODE_FILLSTROKE,
+                   FPDF_TEXTRENDERMODE_FILL_STROKE);
+STATIC_ASSERT_ENUM(PP_TEXTRENDERINGMODE_INVISIBLE,
+                   FPDF_TEXTRENDERMODE_INVISIBLE);
+STATIC_ASSERT_ENUM(PP_TEXTRENDERINGMODE_FILLCLIP,
+                   FPDF_TEXTRENDERMODE_FILL_CLIP);
+STATIC_ASSERT_ENUM(PP_TEXTRENDERINGMODE_STROKECLIP,
+                   FPDF_TEXTRENDERMODE_STROKE_CLIP);
+STATIC_ASSERT_ENUM(PP_TEXTRENDERINGMODE_FILLSTROKECLIP,
+                   FPDF_TEXTRENDERMODE_FILL_STROKE_CLIP);
+STATIC_ASSERT_ENUM(PP_TEXTRENDERINGMODE_CLIP, FPDF_TEXTRENDERMODE_CLIP);
+
 #if defined(OS_WIN)
 STATIC_ASSERT_ENUM(chrome_pdf::kEmf, FPDF_PRINTMODE_EMF);
 STATIC_ASSERT_ENUM(chrome_pdf::kTextOnly, FPDF_PRINTMODE_TEXTONLY);

pdf/pdfium/pdfium_page.cc

@@ -323,7 +323,12 @@ void PDFiumPage::CalculateTextRunStyleInfo(
     style_info->stroke_color = MakeARGB(0xff, 0, 0, 0);
   }
 
-  style_info->render_mode = FPDFText_GetTextRenderMode(text_page, char_index);
+  int render_mode = FPDFText_GetTextRenderMode(text_page, char_index);
+  if (render_mode < 0 || render_mode > PP_TEXTRENDERINGMODE_LAST) {
+    style_info->render_mode = PP_TEXTRENDERINGMODE_UNKNOWN;
+  } else {
+    style_info->render_mode = static_cast<PP_TextRenderingMode>(render_mode);
+  }
 }
 
 bool PDFiumPage::AreTextStyleEqual(

pdf/pdfium/pdfium_page_unittest.cc

@@ -207,9 +207,17 @@ TEST_F(PDFiumPageTextTest, GetTextRunInfo) {
   int current_char_index = 0;
 
   pp::PDF::PrivateAccessibilityTextStyleInfo expected_style_1 = {
-      "Times-Roman", 0, 0, 12, 0xff000000, 0xff000000, false, false};
+      "Times-Roman",
+      0,
+      PP_TEXTRENDERINGMODE_FILL,
+      12,
+      0xff000000,
+      0xff000000,
+      false,
+      false};
   pp::PDF::PrivateAccessibilityTextStyleInfo expected_style_2 = {
-      "Helvetica", 0, 0, 16, 0xff000000, 0xff000000, false, false};
+      "Helvetica", 0,    PP_TEXTRENDERINGMODE_FILL, 16, 0xff000000, 0xff000000,
+      false,       false};
   // The links span from [7, 22], [52, 66] and [92, 108] with 16, 15 and 17
   // text run lengths respectively. There are text runs preceding and
   // succeeding them.

ppapi/c/private/ppb_pdf.h

@@ -67,6 +67,20 @@ struct PP_PrivateAccessibilityPageInfo {
   uint32_t image_count;
 };
 
+// See PDF Reference 1.7, page 402, table 5.3.
+typedef enum {
+  PP_TEXTRENDERINGMODE_UNKNOWN = -1,
+  PP_TEXTRENDERINGMODE_FILL = 0,
+  PP_TEXTRENDERINGMODE_STROKE = 1,
+  PP_TEXTRENDERINGMODE_FILLSTROKE = 2,
+  PP_TEXTRENDERINGMODE_INVISIBLE = 3,
+  PP_TEXTRENDERINGMODE_FILLCLIP = 4,
+  PP_TEXTRENDERINGMODE_STROKECLIP = 5,
+  PP_TEXTRENDERINGMODE_FILLSTROKECLIP = 6,
+  PP_TEXTRENDERINGMODE_CLIP = 7,
+  PP_TEXTRENDERINGMODE_LAST = PP_TEXTRENDERINGMODE_CLIP
+} PP_TextRenderingMode;
+
 // This holds the text style information provided by the PDF and will be used
 // in accessibility to provide the text style information. Needs to stay in
 // sync with C++ version. (PrivateAccessibilityTextStyleInfo and
@@ -75,7 +89,7 @@ struct PP_PrivateAccessibilityTextStyleInfo {
   const char* font_name;
   uint32_t font_name_length;
   int font_weight;
-  int render_mode;
+  PP_TextRenderingMode render_mode;
   double font_size;
   // Colors are ARGB.
   uint32_t fill_color;

ppapi/cpp/private/pdf.h

@@ -27,7 +27,7 @@ class PDF {
   struct PrivateAccessibilityTextStyleInfo {
     std::string font_name;
     int font_weight;
-    int render_mode;
+    PP_TextRenderingMode render_mode;
     double font_size;
     // Colors are ARGB.
     uint32_t fill_color;

ppapi/proxy/ppapi_messages.h

@@ -153,6 +153,7 @@ IPC_ENUM_TRAITS_MAX_VALUE(PP_HardwareAcceleration, PP_HARDWAREACCELERATION_LAST)
 IPC_ENUM_TRAITS_MAX_VALUE(PP_AudioProfile, PP_AUDIOPROFILE_MAX)
 IPC_ENUM_TRAITS_MAX_VALUE(PP_VideoProfile, PP_VIDEOPROFILE_MAX)
 IPC_ENUM_TRAITS_MAX_VALUE(PP_PrivateDirection, PP_PRIVATEDIRECTION_LAST)
+IPC_ENUM_TRAITS_MAX_VALUE(PP_TextRenderingMode, PP_TEXTRENDERINGMODE_LAST)
 IPC_ENUM_TRAITS_MAX_VALUE(PP_PdfAccessibilityAction,
                           PP_PDF_ACCESSIBILITYACTION_LAST)
 IPC_ENUM_TRAITS_MAX_VALUE(PP_PdfAccessibilityScrollAlignment,

ppapi/shared_impl/pdf_accessibility_shared.h

@@ -25,7 +25,7 @@ struct PPAPI_SHARED_EXPORT PdfAccessibilityTextStyleInfo {
 
   std::string font_name;
   int font_weight;
-  int render_mode;
+  PP_TextRenderingMode render_mode;
   double font_size;
   // Colors are ARGB.
   uint32_t fill_color;