src/lyrics_fetcher.cpp

   1 /***************************************************************************
   2  *   Copyright (C) 2008-2017 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 == "jahlyrics")
  56                 fetcher = std::make_unique<JahLyricsFetcher>();
  57         else if (s == "plyrics")
  58                 fetcher = std::make_unique<PLyricsFetcher>();
  59         else if (s == "tekstowo")
  60                 fetcher = std::make_unique<TekstowoFetcher>();
  61         else if (s == "internet")
  62                 fetcher = std::make_unique<InternetLyricsFetcher>();
  63         else
  64                 is.setstate(std::ios::failbit);
  65         return is;
  66 }
  67
  68 const char LyricsFetcher::msgNotFound[] = "Not found";
  69
  70 LyricsFetcher::Result LyricsFetcher::fetch(const std::string &artist,
  71                                            const std::string &title)
  72 {
  73         Result result;
  74         result.first = false;
  75
  76         std::string url = urlTemplate();
  77         boost::replace_all(url, "%artist%", Curl::escape(artist));
  78         boost::replace_all(url, "%title%", Curl::escape(title));
  79
  80         std::string data;
  81         CURLcode code = Curl::perform(data, url);
  82
  83         if (code != CURLE_OK)
  84         {
  85                 result.second = curl_easy_strerror(code);
  86                 return result;
  87         }
  88
  89         auto lyrics = getContent(regex(), data);
  90
  91         if (lyrics.empty() || notLyrics(data))
  92         {
  93                 result.second = msgNotFound;
  94                 return result;
  95         }
  96
  97         data.clear();
  98         for (auto it = lyrics.begin(); it != lyrics.end(); ++it)
  99         {
 100                 postProcess(*it);
 101                 if (!it->empty())
 102                 {
 103                         data += *it;
 104                         if (it != lyrics.end()-1)
 105                                 data += "\n\n----------\n\n";
 106                 }
 107         }
 108
 109         result.second = data;
 110         result.first = true;
 111         return result;
 112 }
 113
 114 std::vector<std::string> LyricsFetcher::getContent(const char *regex_,
 115                                                    const std::string &data)
 116 {
 117         std::vector<std::string> result;
 118         boost::regex rx(regex_);
 119         auto first = boost::sregex_iterator(data.begin(), data.end(), rx);
 120         auto last = boost::sregex_iterator();
 121         for (; first != last; ++first)
 122                 result.push_back(first->str(1));
 123         return result;
 124 }
 125
 126 void LyricsFetcher::postProcess(std::string &data) const
 127 {
 128         data = unescapeHtmlUtf8(data);
 129         stripHtmlTags(data);
 130         // Remove indentation from each line and collapse multiple newlines into one.
 131         std::vector<std::string> lines;
 132         boost::split(lines, data, boost::is_any_of("\n"));
 133         for (auto &line : lines)
 134                 boost::trim(line);
 135         std::unique(lines.begin(), lines.end(), [](std::string &a, std::string &b) {
 136                 return a.empty() && b.empty();
 137         });
 138         data = boost::algorithm::join(lines, "\n");
 139         boost::trim(data);
 140 }
 141
 142 /***********************************************************************/
 143
 144 LyricsFetcher::Result LyricwikiFetcher::fetch(const std::string &artist,
 145                                               const std::string &title)
 146 {
 147         LyricsFetcher::Result result = LyricsFetcher::fetch(artist, title);
 148         if (result.first == true)
 149         {
 150                 result.first = false;
 151
 152                 std::string data;
 153                 CURLcode code = Curl::perform(data, result.second, "", true);
 154
 155                 if (code != CURLE_OK)
 156                 {
 157                         result.second = curl_easy_strerror(code);
 158                         return result;
 159                 }
 160
 161                 auto lyrics = getContent("<div class='lyricbox'>(.*?)</div>", data);
 162
 163                 if (lyrics.empty())
 164                 {
 165                         result.second = msgNotFound;
 166                         return result;
 167                 }
 168                 std::transform(lyrics.begin(), lyrics.end(), lyrics.begin(), unescapeHtmlUtf8);
 169                 bool license_restriction = std::any_of(lyrics.begin(), lyrics.end(), [](const std::string &s) {
 170                         return s.find("Unfortunately, we are not licensed to display the full lyrics for this song at the moment.") != std::string::npos;
 171                 });
 172                 if (license_restriction)
 173                 {
 174                         result.second = "Licence restriction";
 175                         return result;
 176                 }
 177
 178                 data.clear();
 179                 for (auto it = lyrics.begin(); it != lyrics.end(); ++it)
 180                 {
 181                         stripHtmlTags(*it);
 182                         boost::trim(*it);
 183                         if (!it->empty())
 184                         {
 185                                 data += *it;
 186                                 if (it != lyrics.end()-1)
 187                                         data += "\n\n----------\n\n";
 188                         }
 189                 }
 190
 191                 result.second = data;
 192                 result.first = true;
 193         }
 194         return result;
 195 }
 196
 197 bool LyricwikiFetcher::notLyrics(const std::string &data) const
 198 {
 199         return data.find("action=edit") != std::string::npos;
 200 }
 201
 202 /**********************************************************************/
 203
 204 LyricsFetcher::Result GoogleLyricsFetcher::fetch(const std::string &artist,
 205                                                  const std::string &title)
 206 {
 207         Result result;
 208         result.first = false;
 209
 210         std::string search_str = "site:";
 211         search_str += Curl::escape(siteKeyword());
 212         search_str += "+";
 213         search_str += Curl::escape(artist);
 214         search_str += "+";
 215         search_str += Curl::escape(title);
 216
 217         std::string google_url = "http://www.google.com/search?hl=en&ie=UTF-8&oe=UTF-8&q=";
 218         google_url += search_str;
 219         google_url += "&btnI=I%27m+Feeling+Lucky";
 220
 221         std::string data;
 222         CURLcode code = Curl::perform(data, google_url, google_url);
 223
 224         if (code != CURLE_OK)
 225         {
 226                 result.second = curl_easy_strerror(code);
 227                 return result;
 228         }
 229
 230         auto urls = getContent("<A HREF=\"(.*?)\">here</A>", data);
 231
 232         if (urls.empty() || !isURLOk(urls[0]))
 233         {
 234                 result.second = msgNotFound;
 235                 return result;
 236         }
 237
 238         data = unescapeHtmlUtf8(urls[0]);
 239
 240         URL = data.c_str();
 241         return LyricsFetcher::fetch("", "");
 242 }
 243
 244 bool GoogleLyricsFetcher::isURLOk(const std::string &url)
 245 {
 246         return url.find(siteKeyword()) != std::string::npos;
 247 }
 248
 249 /**********************************************************************/
 250
 251 bool MetrolyricsFetcher::isURLOk(const std::string &url)
 252 {
 253         // it sometimes return link to sitemap.xml, which is huge so we need to discard it
 254         return GoogleLyricsFetcher::isURLOk(url) && url.find("sitemap") == std::string::npos;
 255 }
 256
 257 /**********************************************************************/
 258
 259 LyricsFetcher::Result InternetLyricsFetcher::fetch(const std::string &artist,
 260                                                    const std::string &title)
 261 {
 262         GoogleLyricsFetcher::fetch(artist, title);
 263         LyricsFetcher::Result result;
 264         result.first = false;
 265         result.second = "The following site may contain lyrics for this song: ";
 266         result.second += URL;
 267         return result;
 268 }
 269
 270 bool InternetLyricsFetcher::isURLOk(const std::string &url)
 271 {
 272         URL = url;
 273         return false;
 274 }