kworship/bible/KwBibleModuleBibleGateway.cpp

   1 /***************************************************************************
   2  *   This file is part of KWorship.                                        *
   3  *   Copyright 2008 James Hogan <james@albanarts.com>                      *
   4  *                                                                         *
   5  *   KWorship 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  *   KWorship 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 KWorship.  If not, write to the Free Software Foundation,  *
  17  *   Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.   *
  18  ***************************************************************************/
  19
  20 /**
  21  * @file KwBibleModuleBibleGateway.cpp
  22  * @brief A BibleGateway bible module.
  23  * @author James Hogan <james@albanarts.com>
  24  */
  25
  26 #include "KwBibleModuleBibleGateway.h"
  27 #include "KwBiblePassage.h"
  28
  29 #include <KIO/NetAccess>
  30 #include <KMessageBox>
  31 #include <KLocale>
  32 #include <dom/html_document.h>
  33 #include <dom/html_element.h>
  34 #include <dom/html_block.h>
  35
  36 #include <QStringList>
  37 #include <QFile>
  38
  39 /*
  40  * Constructors + destructor
  41  */
  42
  43 /// Default constructor.
  44 KwBibleModuleBibleGateway::KwBibleModuleBibleGateway(int id)
  45 : KwBibleModule()
  46 {
  47   KUrl url(QString("http://www.biblegateway.com/versions/index.php?action=getVersionInfo&vid=%1").arg(id));
  48
  49   QString tmpFile;
  50   if (KIO::NetAccess::download(url, tmpFile, 0))
  51   {
  52     QFile file(tmpFile);
  53     if (file.open(QFile::ReadOnly | QFile::Text))
  54     {
  55       QByteArray rawPage = file.readAll();
  56       file.close();
  57
  58       DOM::HTMLDocument doc;
  59       doc.loadXML(QString::fromUtf8(rawPage));
  60       DOM::Element bookList = doc.getElementById("booklist");
  61       bool tableFound = false;
  62       if (!bookList.isNull())
  63       {
  64         // Get the next table
  65         DOM::Node sibling = bookList.nextSibling();
  66         while (!sibling.isNull() && sibling.nodeType() != DOM::Node::ELEMENT_NODE)
  67         {
  68           sibling = sibling.nextSibling();
  69         }
  70         DOM::Element tableElement(sibling);
  71         if (!tableElement.isNull() && tableElement.tagName() == "table")
  72         {
  73           tableFound = true;
  74           // Each row except header is a book
  75           DOM::NodeList rows = tableElement.getElementsByTagName("tr");
  76           bool firstRow = true;
  77           for (unsigned int row = 0; row < rows.length(); ++row)
  78           {
  79             DOM::NodeList cells = DOM::Element(rows.item(row)).getElementsByTagName("td");
  80             if (cells.length() == 2)
  81             {
  82               // First cell is the name
  83               m_bookList.push_back(Book());
  84               Book* book = &m_bookList[m_bookList.size()-1];
  85               book->name = DOM::HTMLElement(cells.item(0)).innerText().string();
  86               // Also check if text is right-to-left
  87               if (firstRow)
  88               {
  89                 setRightToLeft(DOM::HTMLElement(cells.item(0)).className().string().toLower().contains("rtl"));
  90                 firstRow = false;
  91               }
  92               // Second cell is the chapter links
  93               DOM::NodeList chapterLinks = DOM::Element(cells.item(1)).getElementsByTagName("a");
  94               for (unsigned int chapter = 0; chapter < chapterLinks.length(); ++chapter)
  95               {
  96                 // Check the chapter number is right
  97                 DOM::HTMLElement link = chapterLinks.item(chapter);
  98                 bool numeric;
  99                 int check = link.innerText().string().toInt(&numeric);
 100                 if (!numeric || check != (int)chapter+1)
 101                 {
 102                   KMessageBox::error(0, i18n("Error parsing webpage: %1", i18n("Non sequential chapter list in book '%1'", book->name)));
 103                   break;
 104                 }
 105                 // Get the link
 106                 book->chapters.push_back(Chapter());
 107                 Chapter* chapter = &book->chapters[book->chapters.size()-1];
 108                 chapter->url = "http://www.biblegateway.com" + link.getAttribute("href").string();
 109                 chapter->fetched = false;
 110               }
 111             }
 112           }
 113         }
 114       }
 115       if (!tableFound)
 116       {
 117         // Book list reference node not found
 118         KMessageBox::error(0, i18n("Error parsing webpage: %1", i18n("Book list table not found")));
 119       }
 120     }
 121
 122     KIO::NetAccess::removeTempFile(tmpFile);
 123   }
 124   else
 125   {
 126     KMessageBox::error(0, KIO::NetAccess::lastErrorString());
 127   }
 128 }
 129
 130 /// Destructor.
 131 KwBibleModuleBibleGateway::~KwBibleModuleBibleGateway()
 132 {
 133 }
 134
 135 /*
 136  * Main interface
 137  */
 138
 139 QString KwBibleModuleBibleGateway::name()
 140 {
 141   return QString();
 142 }
 143
 144 QString KwBibleModuleBibleGateway::description()
 145 {
 146   return QString();
 147 }
 148
 149 int KwBibleModuleBibleGateway::numChapters(int book)
 150 {
 151   if (book >= 0 && book < m_bookList.size())
 152   {
 153     return m_bookList[book].chapters.size();
 154   }
 155   return 0;
 156 }
 157
 158 int KwBibleModuleBibleGateway::numVerses(int book, int chapter)
 159 {
 160   Chapter* chap = fetchChapter(book, chapter);
 161   if (0 != chap)
 162   {
 163     return chap->verses.size();
 164   }
 165   else
 166   {
 167     return 0;
 168   }
 169 }
 170
 171 bool KwBibleModuleBibleGateway::fillPassageVerse(int bookIndex, int chapterIndex, int verseIndex, KwBiblePassage* outPassage)
 172 {
 173   Chapter* chapter = fetchChapter(bookIndex, chapterIndex);
 174   if (0 != chapter)
 175   {
 176     outPassage->initVerse(bookIndex, 1+chapterIndex, 1+verseIndex,
 177                           chapter->verses[verseIndex].heading,
 178                           chapter->verses[verseIndex].content);
 179   }
 180   return false;
 181 }
 182
 183 /*
 184  * Protected virtual interface
 185  */
 186
 187 void KwBibleModuleBibleGateway::obtainBooks()
 188 {
 189   QStringList list;
 190   for (int book = 0; book < m_bookList.size(); ++book)
 191   {
 192     list << m_bookList[book].name;
 193   }
 194   setBooks(list);
 195 }
 196
 197 /*
 198  * Private functions
 199  */
 200
 201 /// Ensure chapter contents are fetched.
 202 KwBibleModuleBibleGateway::Chapter* KwBibleModuleBibleGateway::fetchChapter(int book, int chapter)
 203 {
 204   if (book >= 0 && book < m_bookList.size())
 205   {
 206     Book* bookObj = &m_bookList[book];
 207     if (chapter >= 0 && chapter < m_bookList[book].chapters.size())
 208     {
 209       Chapter* chap = &bookObj->chapters[chapter];
 210       if (!chap->fetched)
 211       {
 212         QString tmpFile;
 213         if (KIO::NetAccess::download(chap->url, tmpFile, 0))
 214         {
 215           QFile file(tmpFile);
 216           if (file.open(QFile::ReadOnly | QFile::Text))
 217           {
 218             QByteArray rawPage = file.readAll();
 219             file.close();
 220
 221             DOM::HTMLDocument doc;
 222             doc.loadXML(QString::fromUtf8(rawPage));
 223
 224             // Find all spans with class="sup"
 225             DOM::NodeList sups = doc.body().getElementsByClassName("sup");
 226             int verse = 0;
 227             for (unsigned int i = 0; i < sups.length(); ++i)
 228             {
 229               DOM::HTMLElement span = sups.item(i);
 230               if (span.tagName() == "span")
 231               {
 232                 // Get the verse number and validate
 233                 bool numeric;
 234                 QString verseNumber = span.innerText().string();
 235                 Verse verseInfo;
 236                 int check = verseNumber.toInt(&numeric);
 237                 if (!numeric)
 238                 {
 239                   // Its not going to be a verse if it isn't numeric
 240                   KMessageBox::error(0, i18n("Error parsing webpage: %1", i18n("Non numeric superscript encountered: '%1'. It may correspond to a verse range.", verseNumber)));
 241                   continue;
 242                 }
 243                 ++verse;
 244                 if (check != verse)
 245                 {
 246                   KMessageBox::error(0, i18n("Error parsing webpage: %1", i18n("Non sequential verse list in chapter %1 of book '%2'. Expected verse %3 but got verse %4.", (chapter+1), bookObj->name, verse, check)));
 247                   break;
 248                 }
 249
 250                 // Get any headers before it
 251                 DOM::Node sibling;
 252                 sibling = span.previousSibling();
 253                 while (!sibling.isNull())
 254                 {
 255                   DOM::Element siblingElement = sibling;
 256                   if (!siblingElement.isNull())
 257                   {
 258                     // Stop at a span class="sup"
 259                     if (siblingElement.tagName() == "span" && siblingElement.getAttribute("class") == "sup")
 260                     {
 261                       break;
 262                     }
 263                     // See if its an interesting heading
 264                     DOM::HTMLHeadingElement heading = siblingElement;
 265                     if (!heading.isNull())
 266                     {
 267                       verseInfo.heading = heading.toHTML() + verseInfo.heading;
 268                     }
 269                   }
 270                   sibling = sibling.previousSibling();
 271                 }
 272
 273                 // Get any text after it until the next span
 274                 sibling = span.nextSibling();
 275                 while (!sibling.isNull())
 276                 {
 277                   DOM::Element siblingElement = sibling;
 278                   if (!siblingElement.isNull())
 279                   {
 280                     // Stop at a span class="sup"
 281                     if (siblingElement.tagName() == "span" && siblingElement.getAttribute("class") == "sup")
 282                     {
 283                       break;
 284                     }
 285                     // Also stop at headings
 286                     DOM::HTMLHeadingElement heading = siblingElement;
 287                     if (!heading.isNull())
 288                     {
 289                       break;
 290                     }
 291                   }
 292                   verseInfo.content += sibling.toHTML();
 293                   sibling = sibling.nextSibling();
 294                 }
 295
 296                 chap->verses.push_back(verseInfo);
 297               }
 298             }
 299           }
 300
 301           KIO::NetAccess::removeTempFile(tmpFile);
 302         }
 303         else
 304         {
 305           KMessageBox::error(0, KIO::NetAccess::lastErrorString());
 306         }
 307         chap->fetched = true;
 308       }
 309
 310       return chap;
 311     }
 312   }
 313   return 0;
 314 }
 315