Add functions to get word start/end indices in AXTextUtils.

This change adds new functions to AXTextUtils that return the
start and end indices of a string. This function will support a
new API in automation and was implemented in AXTextUtils to
maximize the possibility of shared code.

Bug: 948700
Change-Id: I4c414280910feb1cf6f3180d09e5088264959aba
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1616383
Commit-Queue: Akihiro Ota <akihiroota@chromium.org>
Reviewed-by: Nektarios Paisios <nektar@chromium.org>
Cr-Commit-Position: refs/heads/master@{#667367}

ui/accessibility/ax_text_utils.cc

@@ -6,6 +6,7 @@
 
 #include "base/i18n/break_iterator.h"
 #include "base/logging.h"
+#include "base/numerics/safe_conversions.h"
 #include "base/optional.h"
 #include "base/strings/string_util.h"
 #include "base/strings/utf_string_conversions.h"
@@ -226,4 +227,33 @@ base::string16 ActionVerbToUnlocalizedString(
   return base::string16();
 }
 
+std::vector<int> GetWordStartOffsets(const base::string16& text) {
+  std::vector<int> word_starts;
+  base::i18n::BreakIterator iter(text, base::i18n::BreakIterator::BREAK_WORD);
+  if (!iter.Init())
+    return word_starts;
+  // iter.Advance() returns false if we've run past end of the text.
+  while (iter.Advance()) {
+    if (!iter.IsWord())
+      continue;
+    word_starts.push_back(
+        base::checked_cast<int>(iter.prev()) /* start index */);
+  }
+  return word_starts;
+}
+
+std::vector<int> GetWordEndOffsets(const base::string16& text) {
+  std::vector<int> word_ends;
+  base::i18n::BreakIterator iter(text, base::i18n::BreakIterator::BREAK_WORD);
+  if (!iter.Init())
+    return word_ends;
+  // iter.Advance() returns false if we've run past end of the text.
+  while (iter.Advance()) {
+    if (!iter.IsWord())
+      continue;
+    word_ends.push_back(base::checked_cast<int>(iter.pos()) /* end index */);
+  }
+  return word_ends;
+}
+
 }  // namespace ui

ui/accessibility/ax_text_utils.h

@@ -29,7 +29,7 @@ enum TextBoundaryDirection {
 // (depending on |direction|) from the given |start_offset| until the
 // given boundary is found, and return the offset of that boundary,
 // using the vector of line break character offsets in |line_breaks|.
-size_t AX_EXPORT FindAccessibleTextBoundary(const base::string16& text,
+AX_EXPORT size_t FindAccessibleTextBoundary(const base::string16& text,
                                             const std::vector<int>& line_breaks,
                                             AXTextBoundary boundary,
                                             size_t start_offset,
@@ -37,13 +37,18 @@ size_t AX_EXPORT FindAccessibleTextBoundary(const base::string16& text,
                                             ax::mojom::TextAffinity affinity);
 
 // Returns a string ID that corresponds to the name of the given action.
-base::string16 AX_EXPORT
-ActionVerbToLocalizedString(const ax::mojom::DefaultActionVerb action_verb);
+AX_EXPORT base::string16 ActionVerbToLocalizedString(
+    const ax::mojom::DefaultActionVerb action_verb);
 
 // Returns the non-localized string representation of a supported action.
 // Some APIs on Linux and Windows need to return non-localized action names.
-base::string16 AX_EXPORT
-ActionVerbToUnlocalizedString(const ax::mojom::DefaultActionVerb action_verb);
+AX_EXPORT base::string16 ActionVerbToUnlocalizedString(
+    const ax::mojom::DefaultActionVerb action_verb);
+
+// Returns indices of all word starts in |text|.
+AX_EXPORT std::vector<int> GetWordStartOffsets(const base::string16& text);
+// Returns indices of all word ends in |text|.
+AX_EXPORT std::vector<int> GetWordEndOffsets(const base::string16& text);
 
 }  // namespace ui
 

ui/accessibility/ax_text_utils_unittest.cc

@@ -8,6 +8,7 @@
 #include <utility>
 
 #include "base/strings/utf_string_conversions.h"
+#include "testing/gmock/include/gmock/gmock.h"
 #include "testing/gtest/include/gtest/gtest.h"
 #include "ui/accessibility/ax_enums.mojom.h"
 #include "ui/accessibility/ax_text_boundary.h"
@@ -251,4 +252,29 @@ TEST(AXTextUtils, FindAccessibleTextBoundaryCharacter) {
   verify_boundaries_at_offset(18, 18L, 19UL);
 }
 
+TEST(AXTextUtils, GetWordOffsetsEmptyTest) {
+  const base::string16 text = base::UTF8ToUTF16("");
+  std::vector<int> word_starts = GetWordStartOffsets(text);
+  std::vector<int> word_ends = GetWordEndOffsets(text);
+  EXPECT_EQ(0UL, word_starts.size());
+  EXPECT_EQ(0UL, word_ends.size());
+}
+
+TEST(AXTextUtils, GetWordStartOffsetsBasicTest) {
+  const base::string16 text = base::UTF8ToUTF16("This is very simple input");
+  EXPECT_THAT(GetWordStartOffsets(text), testing::ElementsAre(0, 5, 8, 13, 20));
+}
+
+TEST(AXTextUtils, GetWordEndOffsetsBasicTest) {
+  const base::string16 text = base::UTF8ToUTF16("This is very simple input");
+  EXPECT_THAT(GetWordEndOffsets(text), testing::ElementsAre(4, 7, 12, 19, 25));
+}
+
+TEST(AXTextUtils, GetWordStartOffsetsMalformedInputTest) {
+  const base::string16 text =
+      base::UTF8ToUTF16("..we *## should parse $#@$ through bad ,,  input");
+  EXPECT_THAT(GetWordStartOffsets(text),
+              testing::ElementsAre(2, 9, 16, 27, 35, 43));
+}
+
 }  // namespace ui