src/lyrics_fetcher.cpp

   1 /***************************************************************************
   2  *   Copyright (C) 2008-2016 by Andrzej Rybczak                            *
   3  *   electricityispower@gmail.com                                          *
   4  *                                                                         *
   5  *   This program is free software; you can redistribute it and/or modify  *
   6  *   it under the terms of the GNU General Public License as published by  *
   7  *   the Free Software Foundation; either version 2 of the License, or     *
   8  *   (at your option) any later version.                                   *
   9  *                                                                         *
  10  *   This program is distributed in the hope that it will be useful,       *
  11  *   but WITHOUT ANY WARRANTY; without even the implied warranty of        *
  12  *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the         *
  13  *   GNU General Public License for more details.                          *
  14  *                                                                         *
  15  *   You should have received a copy of the GNU General Public License     *
  16  *   along with this program; if not, write to the                         *
  17  *   Free Software Foundation, Inc.,                                       *
  18  *   51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA.              *
  19  ***************************************************************************/
  20
  21 #include "config.h"
  22 #include "curl_handle.h"
  23
  24 #include <cstdlib>
  25 #include <cstring>
  26 #include <boost/algorithm/string/join.hpp>
  27 #include <boost/algorithm/string/replace.hpp>
  28 #include <boost/algorithm/string/split.hpp>
  29 #include <boost/algorithm/string/trim.hpp>
  30 #include <boost/regex.hpp>
  31
  32 #include "charset.h"
  33 #include "lyrics_fetcher.h"
  34 #include "utility/html.h"
  35 #include "utility/string.h"
  36
  37 std::istream &operator>>(std::istream &is, LyricsFetcher_ &fetcher)
  38 {
  39         std::string s;
  40         is >> s;
  41         if (s == "lyricwiki")
  42                 fetcher = std::make_unique<LyricwikiFetcher>();
  43         else if (s == "azlyrics")
  44                 fetcher = std::make_unique<AzLyricsFetcher>();
  45         else if (s == "genius")
  46                 fetcher = std::make_unique<GeniusFetcher>();
  47         else if (s == "sing365")
  48                 fetcher = std::make_unique<Sing365Fetcher>();
  49         else if (s == "lyricsmania")
  50                 fetcher = std::make_unique<LyricsmaniaFetcher>();
  51         else if (s == "metrolyrics")
  52                 fetcher = std::make_unique<MetrolyricsFetcher>();
  53         else if (s == "justsomelyrics")
  54                 fetcher = std::make_unique<JustSomeLyricsFetcher>();
  55         else if (s == "tekstowo")
  56                 fetcher = std::make_unique<TekstowoFetcher>();
  57         else if (s == "internet")
  58                 fetcher = std::make_unique<InternetLyricsFetcher>();
  59         else
  60                 is.setstate(std::ios::failbit);
  61         return is;
  62 }
  63
  64 const char LyricsFetcher::msgNotFound[] = "Not found";
  65
  66 LyricsFetcher::Result LyricsFetcher::fetch(const std::string &artist, const std::string &title)
  67 {
  68         Result result;
  69         result.first = false;
  70
  71         std::string url = urlTemplate();
  72         boost::replace_all(url, "%artist%", artist);
  73         boost::replace_all(url, "%title%", title);
  74
  75         std::string data;
  76         CURLcode code = Curl::perform(data, url);
  77
  78         if (code != CURLE_OK)
  79         {
  80                 result.second = curl_easy_strerror(code);
  81                 return result;
  82         }
  83
  84         auto lyrics = getContent(regex(), data);
  85
  86         if (lyrics.empty() || notLyrics(data))
  87         {
  88                 result.second = msgNotFound;
  89                 return result;
  90         }
  91
  92         data.clear();
  93         for (auto it = lyrics.begin(); it != lyrics.end(); ++it)
  94         {
  95                 postProcess(*it);
  96                 if (!it->empty())
  97                 {
  98                         data += *it;
  99                         if (it != lyrics.end()-1)
 100                                 data += "\n\n----------\n\n";
 101                 }
 102         }
 103
 104         result.second = data;
 105         result.first = true;
 106         return result;
 107 }
 108
 109 std::vector<std::string> LyricsFetcher::getContent(const char *regex_, const std::string &data)
 110 {
 111         std::vector<std::string> result;
 112         boost::regex rx(regex_);
 113         auto first = boost::sregex_iterator(data.begin(), data.end(), rx);
 114         auto last = boost::sregex_iterator();
 115         for (; first != last; ++first)
 116                 result.push_back(first->str(1));
 117         return result;
 118 }
 119
 120 void LyricsFetcher::postProcess(std::string &data) const
 121 {
 122         data = unescapeHtmlUtf8(data);
 123         stripHtmlTags(data);
 124         // Remove indentation from each line and collapse multiple newlines into one.
 125         std::vector<std::string> lines;
 126         boost::split(lines, data, boost::is_any_of("\n"));
 127         for (auto &line : lines)
 128                 boost::trim(line);
 129         std::unique(lines.begin(), lines.end(), [](std::string &a, std::string &b) {
 130                 return a.empty() && b.empty();
 131         });
 132         data = boost::algorithm::join(lines, "\n");
 133         boost::trim(data);
 134 }
 135
 136 /***********************************************************************/
 137
 138 LyricsFetcher::Result LyricwikiFetcher::fetch(const std::string &artist, const std::string &title)
 139 {
 140         LyricsFetcher::Result result = LyricsFetcher::fetch(artist, title);
 141         if (result.first == true)
 142         {
 143                 result.first = false;
 144
 145                 std::string data;
 146                 CURLcode code = Curl::perform(data, result.second, "", true);
 147
 148                 if (code != CURLE_OK)
 149                 {
 150                         result.second = curl_easy_strerror(code);
 151                         return result;
 152                 }
 153
 154                 auto lyrics = getContent("<div class='lyricbox'>(.*?)</div>", data);
 155
 156                 if (lyrics.empty())
 157                 {
 158                         result.second = msgNotFound;
 159                         return result;
 160                 }
 161                 std::transform(lyrics.begin(), lyrics.end(), lyrics.begin(), unescapeHtmlUtf8);
 162                 bool license_restriction = std::any_of(lyrics.begin(), lyrics.end(), [](const std::string &s) {
 163                         return s.find("Unfortunately, we are not licensed to display the full lyrics for this song at the moment.") != std::string::npos;
 164                 });
 165                 if (license_restriction)
 166                 {
 167                         result.second = "Licence restriction";
 168                         return result;
 169                 }
 170
 171                 data.clear();
 172                 for (auto it = lyrics.begin(); it != lyrics.end(); ++it)
 173                 {
 174                         stripHtmlTags(*it);
 175                         boost::trim(*it);
 176                         if (!it->empty())
 177                         {
 178                                 data += *it;
 179                                 if (it != lyrics.end()-1)
 180                                         data += "\n\n----------\n\n";
 181                         }
 182                 }
 183
 184                 result.second = data;
 185                 result.first = true;
 186         }
 187         return result;
 188 }
 189
 190 bool LyricwikiFetcher::notLyrics(const std::string &data) const
 191 {
 192         return data.find("action=edit") != std::string::npos;
 193 }
 194
 195 /**********************************************************************/
 196
 197 LyricsFetcher::Result GoogleLyricsFetcher::fetch(const std::string &artist, const std::string &title)
 198 {
 199         Result result;
 200         result.first = false;
 201
 202         std::string search_str = artist;
 203         search_str += "+";
 204         search_str += title;
 205         search_str += "+%2B";
 206         search_str += siteKeyword();
 207
 208         std::string google_url = "http://www.google.com/search?hl=en&ie=UTF-8&oe=UTF-8&q=";
 209         google_url += search_str;
 210         google_url += "&btnI=I%27m+Feeling+Lucky";
 211
 212         std::string data;
 213         CURLcode code = Curl::perform(data, google_url, google_url);
 214
 215         if (code != CURLE_OK)
 216         {
 217                 result.second = curl_easy_strerror(code);
 218                 return result;
 219         }
 220
 221         auto urls = getContent("<A HREF=\"(.*?)\">here</A>", data);
 222
 223         if (urls.empty() || !isURLOk(urls[0]))
 224         {
 225                 result.second = msgNotFound;
 226                 return result;
 227         }
 228
 229         data = unescapeHtmlUtf8(urls[0]);
 230
 231         URL = data.c_str();
 232         return LyricsFetcher::fetch("", "");
 233 }
 234
 235 bool GoogleLyricsFetcher::isURLOk(const std::string &url)
 236 {
 237         return url.find(siteKeyword()) != std::string::npos;
 238 }
 239
 240 /**********************************************************************/
 241
 242 bool MetrolyricsFetcher::isURLOk(const std::string &url)
 243 {
 244         // it sometimes return link to sitemap.xml, which is huge so we need to discard it
 245         return GoogleLyricsFetcher::isURLOk(url) && url.find("sitemap") == std::string::npos;
 246 }
 247
 248 /**********************************************************************/
 249
 250 LyricsFetcher::Result InternetLyricsFetcher::fetch(const std::string &artist, const std::string &title)
 251 {
 252         GoogleLyricsFetcher::fetch(artist, title);
 253         LyricsFetcher::Result result;
 254         result.first = false;
 255         result.second = "The following site may contain lyrics for this song: ";
 256         result.second += URL;
 257         return result;
 258 }
 259
 260 bool InternetLyricsFetcher::isURLOk(const std::string &url)
 261 {
 262         URL = url;
 263         return false;
 264 }