Add FPDFTextObj_GetText() API

Generalize CPDF_TextPage::GetTextByRect(), so that it's possible to get
the text from a text page using a predicate, that way we can easily
get the text that belongs to single text object as well.

Change-Id: Ia457af0f41184694dc1481709be72b35685bce7f
Reviewed-on: https://pdfium-review.googlesource.com/39530
Reviewed-by: Henrique Nakashima <hnakashima@chromium.org>
Reviewed-by: Lei Zhang <thestig@chromium.org>
Commit-Queue: Lei Zhang <thestig@chromium.org>

2 files changed, 20 insertions, 2 deletions

diff --git a/core/fpdftext/cpdf_textpage.cpp b/core/fpdftext/cpdf_textpage.cpp
index 2894160437..ed7f36fb6c 100644
--- a/core/fpdftext/cpdf_textpage.cpp
+++ b/core/fpdftext/cpdf_textpage.cpp
@@ -426,7 +426,8 @@ int CPDF_TextPage::GetIndexAtPos(const CFX_PointF& point,
   return pos < nCount ? pos : NearPos;
 }
 
-WideString CPDF_TextPage::GetTextByRect(const CFX_FloatRect& rect) const {
+WideString CPDF_TextPage::GetTextByPredicate(
+    const std::function<bool(const PAGECHAR_INFO&)>& predicate) const {
   if (!m_bIsParsed)
     return WideString();
 
@@ -435,7 +436,7 @@ WideString CPDF_TextPage::GetTextByRect(const CFX_FloatRect& rect) const {
   bool IsAddLineFeed = false;
   WideString strText;
   for (const auto& charinfo : m_CharList) {
-    if (IsRectIntersect(rect, charinfo.m_CharBox)) {
+    if (predicate(charinfo)) {
       if (fabs(posy - charinfo.m_Origin.y) > 0 && !IsContainPreChar &&
           IsAddLineFeed) {
         posy = charinfo.m_Origin.y;
@@ -460,6 +461,19 @@ WideString CPDF_TextPage::GetTextByRect(const CFX_FloatRect& rect) const {
   return strText;
 }
 
+WideString CPDF_TextPage::GetTextByRect(const CFX_FloatRect& rect) const {
+  return GetTextByPredicate([&rect](const PAGECHAR_INFO& charinfo) {
+    return IsRectIntersect(rect, charinfo.m_CharBox);
+  });
+}
+
+WideString CPDF_TextPage::GetTextByObject(
+    const CPDF_TextObject* pTextObj) const {
+  return GetTextByPredicate([pTextObj](const PAGECHAR_INFO& charinfo) {
+    return charinfo.m_pTextObj == pTextObj;
+  });
+}
+
 void CPDF_TextPage::GetCharInfo(int index, FPDF_CHAR_INFO* info) const {
   if (!m_bIsParsed || !pdfium::IndexInBounds(m_CharList, index))
     return;
diff --git a/core/fpdftext/cpdf_textpage.h b/core/fpdftext/cpdf_textpage.h
index 36d01854f5..90b45bd96b 100644
--- a/core/fpdftext/cpdf_textpage.h
+++ b/core/fpdftext/cpdf_textpage.h
@@ -8,6 +8,7 @@
 #define CORE_FPDFTEXT_CPDF_TEXTPAGE_H_
 
 #include <deque>
+#include <functional>
 #include <vector>
 
 #include "core/fpdfapi/page/cpdf_pageobjectlist.h"
@@ -97,6 +98,7 @@ class CPDF_TextPage {
   std::vector<CFX_FloatRect> GetRectArray(int start, int nCount) const;
   int GetIndexAtPos(const CFX_PointF& point, const CFX_SizeF& tolerance) const;
   WideString GetTextByRect(const CFX_FloatRect& rect) const;
+  WideString GetTextByObject(const CPDF_TextObject* pTextObj) const;
 
   // Returns string with the text from |m_TextBuf| that are covered by the input
   // range. |start| and |count| are in terms of the |m_CharIndex|, so the range
@@ -151,6 +153,8 @@ class CPDF_TextPage {
   TextOrientation FindTextlineFlowOrientation() const;
   void AppendGeneratedCharacter(wchar_t unicode, const CFX_Matrix& formMatrix);
   void SwapTempTextBuf(int32_t iCharListStartAppend, int32_t iBufStartAppend);
+  WideString GetTextByPredicate(
+      const std::function<bool(const PAGECHAR_INFO&)>& predicate) const;
 
   UnownedPtr<const CPDF_Page> const m_pPage;
   std::vector<uint16_t> m_CharIndex;