Populate Text Fields in PDFiumPage

This CL introduces a method PDFiumPage::PopulateTextFields() which
reads text form fields from the PDF document and stores relevant
information in a vector within PDFiumPage.

The CL also includes a new test file with sample text fields and a unit
test to validate the new method.

Bug: 1030242
Change-Id: I98a13e237e443f1703ac7b699cc4952cf21c5e10
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2024637
Commit-Queue: Mansi Awasthi <maawas@microsoft.com>
Reviewed-by: Lei Zhang <thestig@chromium.org>
Reviewed-by: Kevin Babbitt <kbabbitt@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#745022}

pdf/pdfium/pdfium_page.cc

@@ -183,6 +183,24 @@ bool FloatEquals(float f1, float f2) {
          kEpsilonScale * fmaxf(fmaxf(fabsf(f1), fabsf(f2)), kEpsilonScale);
 }
 
+using GetFormFieldPropertyFunction =
+    base::RepeatingCallback<unsigned long(unsigned short* buffer,
+                                          unsigned long buflen)>;
+
+// Helper method to fetch string properties of form fields.
+std::string GetFormFieldProperty(GetFormFieldPropertyFunction function) {
+  base::string16 data;
+  size_t buffer_size = function.Run(nullptr, 0);
+  if (buffer_size > 0) {
+    PDFiumAPIStringBufferSizeInBytesAdapter<base::string16> api_string_adapter(
+        &data, buffer_size, true);
+    api_string_adapter.Close(function.Run(
+        reinterpret_cast<unsigned short*>(api_string_adapter.GetData()),
+        buffer_size));
+  }
+  return base::UTF16ToUTF8(data);
+}
+
 }  // namespace
 
 PDFiumPage::LinkTarget::LinkTarget() : page(-1) {}
@@ -266,7 +284,7 @@ void PDFiumPage::CalculatePageObjectTextRunBreaks() {
     }
   }
 
-  PopulateHighlights();
+  PopulateAnnotations();
   for (const auto& highlight : highlights_) {
     if (highlight.start_char_index >= 0 &&
         highlight.start_char_index < chars_count) {
@@ -568,7 +586,7 @@ PDFiumPage::GetHighlightInfo() {
   if (!available_)
     return highlight_info;
 
-  PopulateHighlights();
+  PopulateAnnotations();
 
   highlight_info.reserve(highlights_.size());
   for (const Highlight& highlight : highlights_) {
@@ -1041,57 +1059,98 @@ void PDFiumPage::PopulateImageAltTextForStructElement(
   }
 }
 
-void PDFiumPage::PopulateHighlights() {
-  if (calculated_highlights_)
+void PDFiumPage::PopulateAnnotations() {
+  if (calculated_annotations_)
     return;
 
   FPDF_PAGE page = GetPage();
   if (!page)
     return;
 
-  calculated_highlights_ = true;
-  // Populate highlights from within the pdf page into data structures ready
-  // to be passed to mimehandler. Currently scoped to highlights only.
   int annotation_count = FPDFPage_GetAnnotCount(page);
   for (int i = 0; i < annotation_count; ++i) {
     ScopedFPDFAnnotation annot(FPDFPage_GetAnnot(page, i));
     DCHECK(annot);
     FPDF_ANNOTATION_SUBTYPE subtype = FPDFAnnot_GetSubtype(annot.get());
-    if (subtype != FPDF_ANNOT_HIGHLIGHT)
-      continue;
 
-    FS_RECTF rect;
-    if (!FPDFAnnot_GetRect(annot.get(), &rect))
-      continue;
-
-    Highlight highlight;
-    // We use the bounding box of the highlight as the bounding rect.
-    highlight.bounding_rect =
-        PageToScreen(pp::Point(), 1.0, rect.left, rect.top, rect.right,
-                     rect.bottom, PageOrientation::kOriginal);
-    GetUnderlyingTextRangeForRect(
-        pp::FloatRect(rect.left, rect.bottom, std::abs(rect.right - rect.left),
-                      std::abs(rect.bottom - rect.top)),
-        &highlight.start_char_index, &highlight.char_count);
-
-    // Retrieve the color of the highlight.
-    unsigned int color_r;
-    unsigned int color_g;
-    unsigned int color_b;
-    unsigned int color_a;
-    FPDF_PAGEOBJECT page_object = FPDFAnnot_GetObject(annot.get(), 0);
-    if (FPDFPageObj_GetFillColor(page_object, &color_r, &color_g, &color_b,
-                                 &color_a)) {
-      highlight.color = MakeARGB(color_a, color_r, color_g, color_b);
-    } else {
-      // Set the same default color as in pdfium. See calls to
-      // GetColorStringWithDefault() in CPVT_GenerateAP::Generate*AP() in
-      // pdfium.
-      highlight.color = MakeARGB(255, 255, 255, 0);
+    switch (subtype) {
+      case FPDF_ANNOT_HIGHLIGHT: {
+        PopulateHighlight(annot.get());
+        break;
+      }
+      case FPDF_ANNOT_WIDGET: {
+        // TODO(crbug.com/1030242): Populate other types of form fields too.
+        if (FPDFAnnot_GetFormFieldType(engine_->form(), annot.get()) ==
+            FPDF_FORMFIELD_TEXTFIELD) {
+          PopulateTextField(annot.get());
+        }
+        break;
+      }
+      default:
+        break;
     }
-
-    highlights_.push_back(std::move(highlight));
   }
+  calculated_annotations_ = true;
+}
+
+void PDFiumPage::PopulateHighlight(FPDF_ANNOTATION annot) {
+  DCHECK(annot);
+  DCHECK_EQ(FPDFAnnot_GetSubtype(annot), FPDF_ANNOT_HIGHLIGHT);
+
+  FS_RECTF rect;
+  if (!FPDFAnnot_GetRect(annot, &rect))
+    return;
+
+  Highlight highlight;
+  // We use the bounding box of the highlight as the bounding rect.
+  highlight.bounding_rect =
+      PageToScreen(pp::Point(), 1.0, rect.left, rect.top, rect.right,
+                   rect.bottom, PageOrientation::kOriginal);
+  GetUnderlyingTextRangeForRect(
+      pp::FloatRect(rect.left, rect.bottom, std::abs(rect.right - rect.left),
+                    std::abs(rect.bottom - rect.top)),
+      &highlight.start_char_index, &highlight.char_count);
+
+  // Retrieve the color of the highlight.
+  unsigned int color_r;
+  unsigned int color_g;
+  unsigned int color_b;
+  unsigned int color_a;
+  FPDF_PAGEOBJECT page_object = FPDFAnnot_GetObject(annot, 0);
+  if (FPDFPageObj_GetFillColor(page_object, &color_r, &color_g, &color_b,
+                               &color_a)) {
+    highlight.color = MakeARGB(color_a, color_r, color_g, color_b);
+  } else {
+    // Set the same default color as in pdfium. See calls to
+    // GetColorStringWithDefault() in CPVT_GenerateAP::Generate*AP() in
+    // pdfium.
+    highlight.color = MakeARGB(255, 255, 255, 0);
+  }
+
+  highlights_.push_back(std::move(highlight));
+}
+
+void PDFiumPage::PopulateTextField(FPDF_ANNOTATION annot) {
+  DCHECK(annot);
+  FPDF_FORMHANDLE form_handle = engine_->form();
+  DCHECK_EQ(FPDFAnnot_GetFormFieldType(form_handle, annot),
+            FPDF_FORMFIELD_TEXTFIELD);
+
+  FS_RECTF rect;
+  if (!FPDFAnnot_GetRect(annot, &rect))
+    return;
+
+  TextField text_field;
+  // We use the bounding box of the text field as the bounding rect.
+  text_field.bounding_rect =
+      PageToScreen(pp::Point(), 1.0, rect.left, rect.top, rect.right,
+                   rect.bottom, PageOrientation::kOriginal);
+  text_field.value = GetFormFieldProperty(
+      base::BindRepeating(FPDFAnnot_GetFormFieldValue, form_handle, annot));
+  text_field.name = GetFormFieldProperty(
+      base::BindRepeating(FPDFAnnot_GetFormFieldName, form_handle, annot));
+  text_field.flags = FPDFAnnot_GetFormFieldFlags(form_handle, annot);
+  text_fields_.push_back(std::move(text_field));
 }
 
 bool PDFiumPage::GetUnderlyingTextRangeForRect(const pp::FloatRect& rect,
@@ -1244,6 +1303,12 @@ PDFiumPage::Highlight::Highlight(const Highlight& that) = default;
 
 PDFiumPage::Highlight::~Highlight() = default;
 
+PDFiumPage::TextField::TextField() = default;
+
+PDFiumPage::TextField::TextField(const TextField& that) = default;
+
+PDFiumPage::TextField::~TextField() = default;
+
 int ToPDFiumRotation(PageOrientation orientation) {
   // Could static_cast<int>(orientation), but using an exhaustive switch will
   // trigger an error if we ever change the definition of PageOrientation.

pdf/pdfium/pdfium_page.h

@@ -170,6 +170,7 @@ class PDFiumPage {
   FRIEND_TEST_ALL_PREFIXES(PDFiumPageImageTest, TestImageAltText);
   FRIEND_TEST_ALL_PREFIXES(PDFiumPageLinkTest, TestLinkGeneration);
   FRIEND_TEST_ALL_PREFIXES(PDFiumPageHighlightTest, TestPopulateHighlights);
+  FRIEND_TEST_ALL_PREFIXES(PDFiumPageTextFieldTest, TestPopulateTextFields);
 
   // Returns a link index if the given character index is over a link, or -1
   // otherwise.
@@ -182,8 +183,12 @@ class PDFiumPage {
   void PopulateAnnotationLinks();
   // Calculate the locations of images on the page.
   void CalculateImages();
-  // Populate highlights on the page.
-  void PopulateHighlights();
+  // Populate annotations like highlight and text field on the page.
+  void PopulateAnnotations();
+  // Populate |highlights_| with |annot|.
+  void PopulateHighlight(FPDF_ANNOTATION annot);
+  // Populate |text_fields_| with |annot|.
+  void PopulateTextField(FPDF_ANNOTATION annot);
   // Returns link type and fills target associated with a link. Returns
   // NONSELECTABLE_AREA if link detection failed.
   Area GetLinkTarget(FPDF_LINK link, LinkTarget* target);
@@ -275,6 +280,20 @@ class PDFiumPage {
     uint32_t color;
   };
 
+  // Represents a text field within the page.
+  struct TextField {
+    TextField();
+    TextField(const TextField& other);
+    ~TextField();
+
+    // Represents the name of form field as defined in the field dictionary.
+    std::string name;
+    std::string value;
+    pp::Rect bounding_rect;
+    // Represents the flags of form field as defined in the field dictionary.
+    int flags;
+  };
+
   PDFiumEngine* engine_;
   ScopedFPDFPage page_;
   ScopedFPDFTextPage text_page_;
@@ -285,8 +304,9 @@ class PDFiumPage {
   std::vector<Link> links_;
   bool calculated_images_ = false;
   std::vector<Image> images_;
-  bool calculated_highlights_ = false;
+  bool calculated_annotations_ = false;
   std::vector<Highlight> highlights_;
+  std::vector<TextField> text_fields_;
   bool calculated_page_object_text_run_breaks_ = false;
   // The set of character indices on which text runs need to be broken for page
   // objects.

pdf/pdfium/pdfium_page_unittest.cc

@@ -372,7 +372,7 @@ TEST_F(PDFiumPageHighlightTest, TestPopulateHighlights) {
 
   PDFiumPage* page = GetPDFiumPageForTest(engine.get(), 0);
   ASSERT_TRUE(page);
-  page->PopulateHighlights();
+  page->PopulateAnnotations();
   ASSERT_EQ(base::size(kExpectedHighlights), page->highlights_.size());
 
   for (size_t i = 0; i < page->highlights_.size(); ++i) {
@@ -386,4 +386,41 @@ TEST_F(PDFiumPageHighlightTest, TestPopulateHighlights) {
   }
 }
 
+using PDFiumPageTextFieldTest = PDFiumTestBase;
+
+TEST_F(PDFiumPageTextFieldTest, TestPopulateTextFields) {
+  struct ExpectedTextField {
+    const char* name;
+    const char* value;
+    pp::Rect bounding_rect;
+    int flags;
+  };
+
+  static const ExpectedTextField kExpectedTextFields[] = {
+      {"Text Box", "Text", {138, 230, 135, 41}, 0},
+      {"ReadOnly", "Elephant", {138, 163, 135, 41}, 1},
+      {"Required", "Required Field", {138, 303, 135, 34}, 2},
+      {"Password", "", {138, 356, 135, 35}, 8192}};
+
+  TestClient client;
+  std::unique_ptr<PDFiumEngine> engine =
+      InitializeEngine(&client, FILE_PATH_LITERAL("form_text_fields.pdf"));
+  ASSERT_TRUE(engine);
+  ASSERT_EQ(1, engine->GetNumberOfPages());
+
+  PDFiumPage* page = GetPDFiumPageForTest(engine.get(), 0);
+  ASSERT_TRUE(page);
+  page->PopulateAnnotations();
+  size_t text_fields_count = page->text_fields_.size();
+  ASSERT_EQ(base::size(kExpectedTextFields), text_fields_count);
+
+  for (size_t i = 0; i < text_fields_count; ++i) {
+    EXPECT_EQ(kExpectedTextFields[i].name, page->text_fields_[i].name);
+    EXPECT_EQ(kExpectedTextFields[i].value, page->text_fields_[i].value);
+    CompareRect(kExpectedTextFields[i].bounding_rect,
+                page->text_fields_[i].bounding_rect);
+    EXPECT_EQ(kExpectedTextFields[i].flags, page->text_fields_[i].flags);
+  }
+}
+
 }  // namespace chrome_pdf

pdf/test/data/form_text_fields.in

@@ -0,0 +1,105 @@
+{{header}}
+{{object 1 0}} <<
+  /Type /Catalog
+  /Pages 2 0 R
+  /AcroForm <<
+    /Fields [ 7 0 R 8 0 R 9 0 R 10 0 R ]
+    /DR 4 0 R
+  >>
+>>
+endobj
+{{object 2 0}} <<
+  /Count 1
+  /Kids [ 3 0 R ]
+  /Type /Pages
+>>
+endobj
+{{object 3 0}} <<
+  /Type /Page
+  /Parent 2 0 R
+  /Resources 4 0 R
+  /MediaBox [ 0 0 300 300 ]
+  /Contents 6 0 R
+  /Annots [ 7 0 R 8 0 R 9 0 R 10 0 R ]
+>>
+endobj
+{{object 4 0}} <<
+  /Font <<
+    /F1 5 0 R
+  >>
+>>
+endobj
+{{object 5 0}} <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Helvetica
+>>
+endobj
+{{object 6 0}} <<
+  {{streamlen}}
+>>
+stream
+BT
+/F1 12 Tf
+100 200 Td
+(Test Form) Tj
+/F1 12 Tf
+-80 -40 Td
+(Read Only:) Tj
+/F1 12 Tf
+0 -50 Td
+(Sample Text) Tj
+/F1 12 Tf
+200 -55 Td
+(*required field) Tj
+/F1 12 Tf
+-200 -35 Td
+(Password:) Tj
+ET
+endstream
+endobj
+{{object 7 0}} <<
+  /Type /Annot
+  /Subtype /Widget
+  /FT /Tx
+  /T (Text Box)
+  /V (Text)
+  /DA (0 0 0 rg /F1 12 Tf)
+  /Rect [ 100 100 200 130 ]
+>>
+endobj
+{{object 8 0}} <<
+  /Type /Annot
+  /Subtype /Widget
+  /FT /Tx
+  /Ff 1
+  /T (ReadOnly)
+  /V (Elephant)
+  /DA (0 0 0 rg /F1 12 Tf)
+  /Rect [ 100 150 200 180 ]
+>>
+endobj
+{{object 9 0}} <<
+  /Type /Annot
+  /Subtype /Widget
+  /FT /Tx
+  /Ff 2
+  /T (Required)
+  /V (Required Field)
+  /DA (0 0 0 rg /F1 12 Tf)
+  /Rect [ 100 50 200 75 ]
+>>
+{{object 10 0}} <<
+  /Type /Annot
+  /Subtype /Widget
+  /FT /Tx
+  /Ff 8192
+  /T (Password)
+  /DA (0 0 0 rg /F1 12 Tf)
+  /Rect [ 100 10 200 35 ]
+>>
+endobj
+{{xref}}
+{{trailer}}
+{{startxref}}
+%%EOF

pdf/test/data/form_text_fields.pdf

@@ -0,0 +1,122 @@
+%PDF-1.7
+%<25><><EFBFBD><EFBFBD>
+1 0 obj <<
+  /Type /Catalog
+  /Pages 2 0 R
+  /AcroForm <<
+    /Fields [ 7 0 R 8 0 R 9 0 R 10 0 R ]
+    /DR 4 0 R
+  >>
+>>
+endobj
+2 0 obj <<
+  /Count 1
+  /Kids [ 3 0 R ]
+  /Type /Pages
+>>
+endobj
+3 0 obj <<
+  /Type /Page
+  /Parent 2 0 R
+  /Resources 4 0 R
+  /MediaBox [ 0 0 300 300 ]
+  /Contents 6 0 R
+  /Annots [ 7 0 R 8 0 R 9 0 R 10 0 R ]
+>>
+endobj
+4 0 obj <<
+  /Font <<
+    /F1 5 0 R
+  >>
+>>
+endobj
+5 0 obj <<
+  /Type /Font
+  /Subtype /Type1
+  /BaseFont /Helvetica
+>>
+endobj
+6 0 obj <<
+  /Length 194
+>>
+stream
+BT
+/F1 12 Tf
+100 200 Td
+(Test Form) Tj
+/F1 12 Tf
+-80 -40 Td
+(Read Only:) Tj
+/F1 12 Tf
+0 -50 Td
+(Sample Text) Tj
+/F1 12 Tf
+200 -55 Td
+(*required field) Tj
+/F1 12 Tf
+-200 -35 Td
+(Password:) Tj
+ET
+endstream
+endobj
+7 0 obj <<
+  /Type /Annot
+  /Subtype /Widget
+  /FT /Tx
+  /T (Text Box)
+  /V (Text)
+  /DA (0 0 0 rg /F1 12 Tf)
+  /Rect [ 100 100 200 130 ]
+>>
+endobj
+8 0 obj <<
+  /Type /Annot
+  /Subtype /Widget
+  /FT /Tx
+  /Ff 1
+  /T (ReadOnly)
+  /V (Elephant)
+  /DA (0 0 0 rg /F1 12 Tf)
+  /Rect [ 100 150 200 180 ]
+>>
+endobj
+9 0 obj <<
+  /Type /Annot
+  /Subtype /Widget
+  /FT /Tx
+  /Ff 2
+  /T (Required)
+  /V (Required Field)
+  /DA (0 0 0 rg /F1 12 Tf)
+  /Rect [ 100 50 200 75 ]
+>>
+10 0 obj <<
+  /Type /Annot
+  /Subtype /Widget
+  /FT /Tx
+  /Ff 8192
+  /T (Password)
+  /DA (0 0 0 rg /F1 12 Tf)
+  /Rect [ 100 10 200 35 ]
+>>
+endobj
+xref
+0 11
+0000000000 65535 f 
+0000000015 00000 n 
+0000000143 00000 n 
+0000000208 00000 n 
+0000000363 00000 n 
+0000000414 00000 n 
+0000000490 00000 n 
+0000000736 00000 n 
+0000000884 00000 n 
+0000001044 00000 n 
+0000001201 00000 n 
+trailer <<
+  /Root 1 0 R
+  /Size 11
+>>
+startxref
+1347
+%%EOF