From 2929b89ced4dbc23279a6fc568b7a9ec5d1c0d7f Mon Sep 17 00:00:00 2001
From: Yury Melnichek <melnichek@gmail.com>
Date: Sun, 29 May 2011 16:35:06 +0200
Subject: [PATCH] [search] Factor out SplitAndNormalizeAndSimplifyString().

---
 search/keyword_matcher.cpp | 52 ++++++++++++++++++++------------------
 search/keyword_matcher.hpp |  1 +
 search/query.cpp           | 10 ++++----
 search/string_match.hpp    |  8 +++++-
 4 files changed, 40 insertions(+), 31 deletions(-)

diff --git a/search/keyword_matcher.cpp b/search/keyword_matcher.cpp
index 711516abda..37e0be3fc0 100644
--- a/search/keyword_matcher.cpp
+++ b/search/keyword_matcher.cpp
@@ -1,6 +1,8 @@
 #include "keyword_matcher.hpp"
 #include "delimiters.hpp"
+#include "string_match.hpp"
 #include "../base/string_utils.hpp"
+#include "../std/bind.hpp"
 #include "../std/numeric.hpp"
 
 namespace search
@@ -25,36 +27,36 @@ KeywordMatcher::KeywordMatcher(strings::UniString * pKeywords,
 
 void KeywordMatcher::ProcessName(string const & name)
 {
-  search::Delimiters delims;
-  for (strings::TokenizeIterator<search::Delimiters> iter(name, delims); iter; ++iter)
+  SplitAndNormalizeAndSimplifyString(
+        name, bind(&KeywordMatcher::ProcessNameToken, this, cref(name), _1), Delimiters());
+}
+
+void KeywordMatcher::ProcessNameToken(string const & name, strings::UniString const & s)
+{
+  for (size_t i = 0; i < m_minKeywordMatchCost.size(); ++i)
   {
-    strings::UniString const s = strings::MakeLowerCase(iter.GetUniString());
+    m_minKeywordMatchCost[i] = min(m_minKeywordMatchCost[i],
+                                   m_keywordMatchFn(&m_pKewords[i][0], m_pKewords[i].size(),
+                                                    &s[0], s.size(),
+                                                    m_minKeywordMatchCost[i]));
+  }
 
-    for (size_t i = 0; i < m_minKeywordMatchCost.size(); ++i)
+  if (!m_prefix.empty())
+  {
+    uint32_t const matchCost = m_prefixMatchFn(&m_prefix[0], m_prefix.size(),
+                                               &s[0], s.size(), m_minPrefixMatchCost);
+    if (matchCost < m_minPrefixMatchCost)
     {
-      m_minKeywordMatchCost[i] = min(m_minKeywordMatchCost[i],
-                                     m_keywordMatchFn(&m_pKewords[i][0], m_pKewords[i].size(),
-                                                      &s[0], s.size(),
-                                                      m_minKeywordMatchCost[i]));
+      m_bestPrefixMatch = name;
+      m_minPrefixMatchCost = matchCost;
     }
-
-    if (!m_prefix.empty())
+  }
+  else
+  {
+    if (m_bestPrefixMatch.empty())
     {
-      uint32_t const matchCost = m_prefixMatchFn(&m_prefix[0], m_prefix.size(),
-                                                 &s[0], s.size(), m_minPrefixMatchCost);
-      if (matchCost < m_minPrefixMatchCost)
-      {
-        m_bestPrefixMatch = name;
-        m_minPrefixMatchCost = matchCost;
-      }
-    }
-    else
-    {
-      if (m_bestPrefixMatch.empty())
-      {
-        m_bestPrefixMatch = name;
-        m_minPrefixMatchCost = 0;
-      }
+      m_bestPrefixMatch = name;
+      m_minPrefixMatchCost = 0;
     }
   }
 }
diff --git a/search/keyword_matcher.hpp b/search/keyword_matcher.hpp
index 0d9e2ee549..7bb0457dd9 100644
--- a/search/keyword_matcher.hpp
+++ b/search/keyword_matcher.hpp
@@ -33,6 +33,7 @@ public:
                  StringMatchFn keywordMatchFn, StringMatchFn prefixMatchFn);
 
   void ProcessName(string const & name);
+  void ProcessNameToken(string const & name, strings::UniString const & token);
 
   // Useful for FeatureType.ForEachName(), calls ProcessName() and always returns true.
   bool operator () (int /*lang*/, string const & name)
diff --git a/search/query.cpp b/search/query.cpp
index ca3a76a7eb..6390f0917a 100644
--- a/search/query.cpp
+++ b/search/query.cpp
@@ -2,6 +2,7 @@
 #include "delimiters.hpp"
 #include "keyword_matcher.hpp"
 #include "string_match.hpp"
+#include "../base/stl_add.hpp"
 
 namespace search
 {
@@ -43,12 +44,11 @@ Query::Query(string const & query, m2::RectD const & rect, IndexType const * pIn
   : m_queryText(query), m_rect(rect), m_pIndex(pIndex)
 {
   search::Delimiters delims;
-  for (strings::TokenizeIterator<search::Delimiters> iter(query, delims); iter; ++iter)
+  SplitAndNormalizeAndSimplifyString(query, MakeBackInsertFunctor(m_keywords), delims);
+  if (!m_keywords.empty() && !delims(strings::LastUniChar(query)))
   {
-    if (iter.IsLast() && !delims(strings::LastUniChar(query)))
-      m_prefix = strings::MakeLowerCase(iter.GetUniString());
-    else
-      m_keywords.push_back(strings::MakeLowerCase(iter.GetUniString()));
+    m_prefix.swap(m_keywords.back());
+    m_keywords.pop_back();
   }
 }
 
diff --git a/search/string_match.hpp b/search/string_match.hpp
index 4e780383b7..97e31f808b 100644
--- a/search/string_match.hpp
+++ b/search/string_match.hpp
@@ -1,5 +1,4 @@
 #pragma once
-
 #include "../base/base.hpp"
 #include "../base/buffer_vector.hpp"
 #include "../base/string_utils.hpp"
@@ -13,6 +12,13 @@ using strings::UniChar;
 namespace impl
 {
 
+template <class DelimsT, typename F>
+void SplitAndNormalizeAndSimplifyString(string const & s, F f, DelimsT const & delims)
+{
+  for (strings::TokenizeIterator<DelimsT> iter(s, delims); iter; ++iter)
+    f(strings::MakeLowerCase(iter.GetUniString()));
+}
+
 struct MatchCostData
 {
   uint32_t m_A, m_B;