Add FPDF_StructElement_GetActualText
Get the actual text for a given structure element.
Change-Id: If2767b12099a21fbe7c0abcb89c307cb4fe59b02
Reviewed-on: https://pdfium-review.googlesource.com/c/pdfium/+/90431
Reviewed-by: Lei Zhang <thestig@chromium.org>
Commit-Queue: Lei Zhang <thestig@chromium.org>
diff --git a/core/fpdfdoc/cpdf_structelement.cpp b/core/fpdfdoc/cpdf_structelement.cpp
index 344b4fe..56cd410 100644
--- a/core/fpdfdoc/cpdf_structelement.cpp
+++ b/core/fpdfdoc/cpdf_structelement.cpp
@@ -53,6 +53,10 @@
return GetDict()->GetUnicodeTextFor("Alt");
}
+WideString CPDF_StructElement::GetActualText() const {
+ return GetDict()->GetUnicodeTextFor("ActualText");
+}
+
WideString CPDF_StructElement::GetTitle() const {
return GetDict()->GetUnicodeTextFor("T");
}
diff --git a/core/fpdfdoc/cpdf_structelement.h b/core/fpdfdoc/cpdf_structelement.h
index 3b7a375..5aac005 100644
--- a/core/fpdfdoc/cpdf_structelement.h
+++ b/core/fpdfdoc/cpdf_structelement.h
@@ -23,6 +23,7 @@
ByteString GetType() const { return m_Type; }
WideString GetAltText() const;
+ WideString GetActualText() const;
WideString GetTitle() const;
// Never returns nullptr.
diff --git a/fpdfsdk/fpdf_structtree.cpp b/fpdfsdk/fpdf_structtree.cpp
index f06a2c7..a162d7c 100644
--- a/fpdfsdk/fpdf_structtree.cpp
+++ b/fpdfsdk/fpdf_structtree.cpp
@@ -80,6 +80,15 @@
}
FPDF_EXPORT unsigned long FPDF_CALLCONV
+FPDF_StructElement_GetActualText(FPDF_STRUCTELEMENT struct_element,
+ void* buffer,
+ unsigned long buflen) {
+ CPDF_StructElement* elem =
+ CPDFStructElementFromFPDFStructElement(struct_element);
+ return elem ? WideStringToBuffer(elem->GetActualText(), buffer, buflen) : 0;
+}
+
+FPDF_EXPORT unsigned long FPDF_CALLCONV
FPDF_StructElement_GetID(FPDF_STRUCTELEMENT struct_element,
void* buffer,
unsigned long buflen) {
diff --git a/fpdfsdk/fpdf_structtree_embeddertest.cpp b/fpdfsdk/fpdf_structtree_embeddertest.cpp
index 6781a48..280b520 100644
--- a/fpdfsdk/fpdf_structtree_embeddertest.cpp
+++ b/fpdfsdk/fpdf_structtree_embeddertest.cpp
@@ -74,6 +74,50 @@
UnloadPage(page);
}
+TEST_F(FPDFStructTreeEmbedderTest, GetActualText) {
+ ASSERT_TRUE(OpenDocument("tagged_actual_text.pdf"));
+ FPDF_PAGE page = LoadPage(0);
+ ASSERT_TRUE(page);
+
+ {
+ ScopedFPDFStructTree struct_tree(FPDF_StructTree_GetForPage(page));
+ ASSERT_TRUE(struct_tree);
+ ASSERT_EQ(1, FPDF_StructTree_CountChildren(struct_tree.get()));
+
+ EXPECT_EQ(0U, FPDF_StructElement_GetActualText(nullptr, nullptr, 0));
+
+ FPDF_STRUCTELEMENT element =
+ FPDF_StructTree_GetChildAtIndex(struct_tree.get(), 0);
+ ASSERT_TRUE(element);
+ EXPECT_EQ(0U, FPDF_StructElement_GetActualText(element, nullptr, 0));
+
+ ASSERT_EQ(1, FPDF_StructElement_CountChildren(element));
+ FPDF_STRUCTELEMENT child_element =
+ FPDF_StructElement_GetChildAtIndex(element, 0);
+ ASSERT_TRUE(child_element);
+ EXPECT_EQ(0U, FPDF_StructElement_GetActualText(child_element, nullptr, 0));
+
+ ASSERT_EQ(1, FPDF_StructElement_CountChildren(child_element));
+ FPDF_STRUCTELEMENT gchild_element =
+ FPDF_StructElement_GetChildAtIndex(child_element, 0);
+ ASSERT_TRUE(gchild_element);
+ ASSERT_EQ(24U,
+ FPDF_StructElement_GetActualText(gchild_element, nullptr, 0));
+
+ unsigned short buffer[12] = {};
+ // Deliberately pass in a small buffer size to make sure |buffer| remains
+ // untouched.
+ ASSERT_EQ(24U, FPDF_StructElement_GetActualText(gchild_element, buffer, 1));
+ for (size_t i = 0; i < pdfium::size(buffer); ++i)
+ EXPECT_EQ(0U, buffer[i]);
+ ASSERT_EQ(24U, FPDF_StructElement_GetActualText(gchild_element, buffer,
+ sizeof(buffer)));
+ EXPECT_EQ(L"Actual Text", GetPlatformWString(buffer));
+ }
+
+ UnloadPage(page);
+}
+
TEST_F(FPDFStructTreeEmbedderTest, GetStringAttribute) {
ASSERT_TRUE(OpenDocument("tagged_table.pdf"));
FPDF_PAGE page = LoadPage(0);
diff --git a/fpdfsdk/fpdf_view_c_api_test.c b/fpdfsdk/fpdf_view_c_api_test.c
index d32e2d2..99c53a3 100644
--- a/fpdfsdk/fpdf_view_c_api_test.c
+++ b/fpdfsdk/fpdf_view_c_api_test.c
@@ -353,6 +353,7 @@
// fpdf_structtree.h
CHK(FPDF_StructElement_CountChildren);
+ CHK(FPDF_StructElement_GetActualText);
CHK(FPDF_StructElement_GetAltText);
CHK(FPDF_StructElement_GetChildAtIndex);
CHK(FPDF_StructElement_GetID);
diff --git a/public/fpdf_structtree.h b/public/fpdf_structtree.h
index 211abdf..4784bd4 100644
--- a/public/fpdf_structtree.h
+++ b/public/fpdf_structtree.h
@@ -76,6 +76,26 @@
unsigned long buflen);
// Experimental API.
+// Function: FPDF_StructElement_GetActualText
+// Get the actual text for a given element.
+// Parameters:
+// struct_element - Handle to the struct element.
+// buffer - A buffer for output the actual text. May be NULL.
+// buflen - The length of the buffer, in bytes. May be 0.
+// Return value:
+// The number of bytes in the actual text, including the terminating
+// NUL character. The number of bytes is returned regardless of the
+// |buffer| and |buflen| parameters.
+// Comments:
+// Regardless of the platform, the |buffer| is always in UTF-16LE
+// encoding. The string is terminated by a UTF16 NUL character. If
+// |buflen| is less than the required length, or |buffer| is NULL,
+// |buffer| will not be modified.
+FPDF_EXPORT unsigned long FPDF_CALLCONV
+FPDF_StructElement_GetActualText(FPDF_STRUCTELEMENT struct_element,
+ void* buffer,
+ unsigned long buflen);
+
// Function: FPDF_StructElement_GetID
// Get the ID for a given element.
// Parameters:
diff --git a/testing/resources/tagged_actual_text.in b/testing/resources/tagged_actual_text.in
new file mode 100644
index 0000000..9bb917b
--- /dev/null
+++ b/testing/resources/tagged_actual_text.in
@@ -0,0 +1,162 @@
+{{header}}
+{{object 1 0}} <<
+ /Type /Catalog
+ /Pages 2 0 R
+ /StructTreeRoot 8 0 R
+ /Lang (en-US)
+ /MarkInfo <<
+ /Marked true
+ >>
+>>
+endobj
+{{object 2 0}} <<
+ /Type /Pages
+ /Count 1
+ /Kids [3 0 R]
+>>
+endobj
+{{object 3 0}} <<
+ /Type /Page
+ /Parent 2 0 R
+ /Contents 4 0 R
+ /MediaBox [0 0 612 792]
+ /Group <<
+ /CS /DeviceRGB
+ /I true
+ /S /Transparency
+ >>
+ /Resources <<
+ /ProcSet [/PDF /ImageC /ImageI /ImageB]
+ /XObject <<
+ /Tr8 5 0 R
+ /Im7 6 0 R
+ >>
+ /ExtGState <<
+ /EGS9 7 0 R
+ >>
+ >>
+ /StructParents 0
+>>
+endobj
+{{object 4 0}} <<
+ {{streamlen}}
+>>
+stream
+0.1 w
+/Artifact
+BMC
+q
+0 0 612 792 re
+W* n
+EMC
+/Figure<</MCID 0>>
+BDC
+Q
+q
+281 685.3 50 50 re
+W* n
+q
+49.9 0 0 50 281.1 685.4 cm
+/Im7 Do
+Q
+EMC
+Q
+q
+EGS9 gs /Tr8 Do
+Q
+endstream
+endobj
+{{object 5 0}} <<
+ /Type /XObject
+ /Subtype /Form
+ /BBox [-140 395 753 395.1]
+ /Group <<
+ /CS /DeviceRGB
+ /K true
+ /S /Transparency
+ >>
+ {{streamlen}}
+>>
+stream
+endstream
+endobj
+{{object 6 0}} <<
+ /Type /XObject
+ /Subtype /Image
+ /Width 50
+ /Height 50
+ /BitsPerComponent 8
+ /ColorSpace /DeviceRGB
+ /Filter [/ASCIIHexDecode /FlateDecode]
+ {{streamlen}}
+>>
+stream
+789cedc13101000000c2a0f54fed6f06a00000000000000078031d4c0001
+endstream
+endobj
+{{object 7 0}} <<
+ /ca 0.5
+ /CA 0.5
+>>
+endobj
+{{object 8 0}} <<
+ /Type /StructTreeRoot
+ /ParentTree 9 0 R
+ /K [11 0 R]
+ /RoleMap <<
+ /Document /Document
+ /Standard /P
+ /Figure /Figure
+ >>
+>>
+endobj
+{{object 9 0}} <<
+ /Nums [0 [10 0 R]]
+>>
+endobj
+{{object 10 0}} <<
+ /Type /StructElem
+ /S /Figure
+ /A 13 0 R
+ /K [0]
+ /P 12 0 R
+ /ActualText <feff00410063007400750061006c00200054006500780074>
+ /Pg 3 0 R
+>>
+endobj
+{{object 11 0}} <<
+ /Type /StructElem
+ /S /Document
+ /K [12 0 R]
+ /P 8 0 R
+ /T (TitleText)
+ /Pg 3 0 R
+>>
+endobj
+{{object 12 0}} <<
+ /Type /StructElem
+ /S /Standard
+ /A 14 0 R
+ /K [10 0 R]
+ /P 11 0 R
+ /T <feff00730079006d0062006f006c003a0020003100300030006b>
+ /Pg 3 0 R
+>>
+endobj
+{{object 13 0}} <<
+ /O /Layout
+ /Placement /Block
+ /BBox [281.1 685.3 331.1 735.3]
+ /Width 99.9
+ /Height 99.9
+>>
+endobj
+{{object 14 0}} <<
+ /O /Layout
+ /Placement /Block
+>>
+endobj
+{{xref}}
+{{trailer}}
+{{startxref}}
+%%EOF
diff --git a/testing/resources/tagged_actual_text.pdf b/testing/resources/tagged_actual_text.pdf
new file mode 100644
index 0000000..634e333
--- /dev/null
+++ b/testing/resources/tagged_actual_text.pdf
@@ -0,0 +1,183 @@
+%PDF-1.7
+% ò¤ô
+1 0 obj <<
+ /Type /Catalog
+ /Pages 2 0 R
+ /StructTreeRoot 8 0 R
+ /Lang (en-US)
+ /MarkInfo <<
+ /Marked true
+ >>
+>>
+endobj
+2 0 obj <<
+ /Type /Pages
+ /Count 1
+ /Kids [3 0 R]
+>>
+endobj
+3 0 obj <<
+ /Type /Page
+ /Parent 2 0 R
+ /Contents 4 0 R
+ /MediaBox [0 0 612 792]
+ /Group <<
+ /CS /DeviceRGB
+ /I true
+ /S /Transparency
+ >>
+ /Resources <<
+ /ProcSet [/PDF /ImageC /ImageI /ImageB]
+ /XObject <<
+ /Tr8 5 0 R
+ /Im7 6 0 R
+ >>
+ /ExtGState <<
+ /EGS9 7 0 R
+ >>
+ >>
+ /StructParents 0
+>>
+endobj
+4 0 obj <<
+ /Length 162
+>>
+stream
+0.1 w
+/Artifact
+BMC
+q
+0 0 612 792 re
+W* n
+EMC
+/Figure<</MCID 0>>
+BDC
+Q
+q
+281 685.3 50 50 re
+W* n
+q
+49.9 0 0 50 281.1 685.4 cm
+/Im7 Do
+Q
+EMC
+Q
+q
+EGS9 gs /Tr8 Do
+Q
+endstream
+endobj
+5 0 obj <<
+ /Type /XObject
+ /Subtype /Form
+ /BBox [-140 395 753 395.1]
+ /Group <<
+ /CS /DeviceRGB
+ /K true
+ /S /Transparency
+ >>
+ /Length 0
+>>
+stream
+endstream
+endobj
+6 0 obj <<
+ /Type /XObject
+ /Subtype /Image
+ /Width 50
+ /Height 50
+ /BitsPerComponent 8
+ /ColorSpace /DeviceRGB
+ /Filter [/ASCIIHexDecode /FlateDecode]
+ /Length 61
+>>
+stream
+789cedc13101000000c2a0f54fed6f06a00000000000000078031d4c0001
+endstream
+endobj
+7 0 obj <<
+ /ca 0.5
+ /CA 0.5
+>>
+endobj
+8 0 obj <<
+ /Type /StructTreeRoot
+ /ParentTree 9 0 R
+ /K [11 0 R]
+ /RoleMap <<
+ /Document /Document
+ /Standard /P
+ /Figure /Figure
+ >>
+>>
+endobj
+9 0 obj <<
+ /Nums [0 [10 0 R]]
+>>
+endobj
+10 0 obj <<
+ /Type /StructElem
+ /S /Figure
+ /A 13 0 R
+ /K [0]
+ /P 12 0 R
+ /ActualText <feff00410063007400750061006c00200054006500780074>
+ /Pg 3 0 R
+>>
+endobj
+11 0 obj <<
+ /Type /StructElem
+ /S /Document
+ /K [12 0 R]
+ /P 8 0 R
+ /T (TitleText)
+ /Pg 3 0 R
+>>
+endobj
+12 0 obj <<
+ /Type /StructElem
+ /S /Standard
+ /A 14 0 R
+ /K [10 0 R]
+ /P 11 0 R
+ /T <feff00730079006d0062006f006c003a0020003100300030006b>
+ /Pg 3 0 R
+>>
+endobj
+13 0 obj <<
+ /O /Layout
+ /Placement /Block
+ /BBox [281.1 685.3 331.1 735.3]
+ /Width 99.9
+ /Height 99.9
+>>
+endobj
+14 0 obj <<
+ /O /Layout
+ /Placement /Block
+>>
+endobj
+xref
+0 15
+0000000000 65535 f
+0000000015 00000 n
+0000000145 00000 n
+0000000208 00000 n
+0000000556 00000 n
+0000000770 00000 n
+0000000952 00000 n
+0000001212 00000 n
+0000001253 00000 n
+0000001412 00000 n
+0000001454 00000 n
+0000001619 00000 n
+0000001730 00000 n
+0000001897 00000 n
+0000002015 00000 n
+trailer <<
+ /Root 1 0 R
+ /Size 15
+>>
+startxref
+2070
+%%EOF
