From ae4b26638d4987944eb1581c81e08ca9763179af Mon Sep 17 00:00:00 2001
From: Ilya Zverev <zverik@textual.ru>
Date: Tue, 27 Oct 2015 16:58:38 +0300
Subject: [PATCH] [metadata] Add GetWikiTitle and GetWikiURL methods

---
 generator/generator_tests/metadata_test.cpp |  4 ++
 indexer/feature_meta.cpp                    | 75 +++++++++++++++++++++
 indexer/feature_meta.hpp                    |  3 +
 indexer/indexer.pro                         |  1 +
 4 files changed, 83 insertions(+)

diff --git a/generator/generator_tests/metadata_test.cpp b/generator/generator_tests/metadata_test.cpp
index fd2056fb9f..0b9ce77b38 100644
--- a/generator/generator_tests/metadata_test.cpp
+++ b/generator/generator_tests/metadata_test.cpp
@@ -137,10 +137,14 @@ UNIT_TEST(Metadata_ValidateAndFormat_wikipedia)
 
   p("wikipedia", "ru:Лана Вуд");
   TEST_EQUAL(params.GetMetadata().Get(feature::Metadata::FMD_WIKIPEDIA), "ru:" + lanaWoodUrlEncoded, ("ru:"));
+  TEST_EQUAL(params.GetMetadata().GetWikiTitle(), "ru:Лана Вуд", ("ru:"));
+  TEST_EQUAL(params.GetMetadata().GetWikiURL(), "https://ru.wikipedia.org/wiki/" + lanaWoodUrlEncoded, ("ru:"));
   params.GetMetadata().Drop(feature::Metadata::FMD_WIKIPEDIA);
 
   p("wikipedia", "https://ru.wikipedia.org/wiki/" + lanaWoodUrlEncoded);
   TEST_EQUAL(params.GetMetadata().Get(feature::Metadata::FMD_WIKIPEDIA), "ru:" + lanaWoodUrlEncoded, ("https:"));
+  TEST_EQUAL(params.GetMetadata().GetWikiTitle(), "ru:Лана Вуд", ("https:"));
+  TEST_EQUAL(params.GetMetadata().GetWikiURL(), "https://ru.wikipedia.org/wiki/" + lanaWoodUrlEncoded, ("https:"));
   params.GetMetadata().Drop(feature::Metadata::FMD_WIKIPEDIA);
 
   p("wikipedia", "Test");
diff --git a/indexer/feature_meta.cpp b/indexer/feature_meta.cpp
index 5b86dee054..02d21f8287 100644
--- a/indexer/feature_meta.cpp
+++ b/indexer/feature_meta.cpp
@@ -1 +1,76 @@
 #include "indexer/feature_meta.hpp"
+
+namespace feature
+{
+string Metadata::GetWikiURL() const
+{
+  string value = this->Get(FMD_WIKIPEDIA);
+  string::size_type i = value.find(':');
+  if (i == string::npos)
+    return string();
+  return "https://" + value.substr(0, i) + ".wikipedia.org/wiki/" + value.substr(i + 1);
+}
+
+char hex_to_dec(char ch)
+{
+  if (ch >= '0' && ch <= '9')
+    return ch - '0';
+  if (ch >= 'a')
+    ch -= 32;
+  if (ch >= 'A' && ch <= 'F')
+    return ch - 'A' + 10;
+  return -1;
+}
+
+string UriDecode(string const & sSrc)
+{
+  // This code was slightly modified from
+  // http://www.codeguru.com/cpp/cpp/string/conversions/article.php/c12759
+  //
+  // Note from RFC1630: "Sequences which start with a percent
+  // sign but are not followed by two hexadecimal characters
+  // (0-9, A-F) are reserved for future extension"
+
+  const unsigned char * pSrc = (const unsigned char *)sSrc.c_str();
+  const string::size_type SRC_LEN = sSrc.length();
+  const unsigned char * const SRC_END = pSrc + SRC_LEN;
+  // last decodable '%'
+  const unsigned char * const SRC_LAST_DEC = SRC_END - 2;
+
+  char * const pStart = new char[SRC_LEN];
+  char * pEnd = pStart;
+
+  while (pSrc < SRC_LAST_DEC)
+  {
+    if (*pSrc == '%')
+    {
+      char dec1 = hex_to_dec(*(pSrc + 1));
+      char dec2 = hex_to_dec(*(pSrc + 2));
+      if (-1 != dec1 && -1 != dec2)
+      {
+        *pEnd++ = (dec1 << 4) + dec2;
+        pSrc += 2;
+      }
+    }
+    else if (*pSrc == '_')
+      *pEnd++ = ' ';
+    else
+      *pEnd++ = *pSrc;
+    pSrc++;
+  }
+
+  // the last 2- chars
+  while (pSrc < SRC_END)
+    *pEnd++ = *pSrc++;
+
+  string sResult(pStart, pEnd);
+  delete [] pStart;
+  return sResult;
+}
+
+string Metadata::GetWikiTitle() const
+{
+  string value = this->Get(FMD_WIKIPEDIA);
+  return UriDecode(value);
+}
+}  // namespace feature
diff --git a/indexer/feature_meta.hpp b/indexer/feature_meta.hpp
index 13f447e0b9..601dfd5fda 100644
--- a/indexer/feature_meta.hpp
+++ b/indexer/feature_meta.hpp
@@ -76,6 +76,9 @@ namespace feature
     inline bool Empty() const { return m_metadata.empty(); }
     inline size_t Size() const { return m_metadata.size(); }
 
+    string GetWikiURL() const;
+    string GetWikiTitle() const;
+
     template <class ArchiveT> void SerializeToMWM(ArchiveT & ar) const
     {
       for (auto const & e : m_metadata)
diff --git a/indexer/indexer.pro b/indexer/indexer.pro
index 3c65434ddc..f41ec5bf01 100644
--- a/indexer/indexer.pro
+++ b/indexer/indexer.pro
@@ -28,6 +28,7 @@ SOURCES += \
     feature_impl.cpp \
     feature_loader.cpp \
     feature_loader_base.cpp \
+    feature_meta.cpp \
     feature_utils.cpp \
     feature_visibility.cpp \
     features_offsets_table.cpp \