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
  28 #include <KIO/NetAccess>
  29 #include <KMessageBox>
  30 #include <KLocale>
  31 #include <dom/html_document.h>
  32 #include <dom/html_element.h>
  33 #include <dom/html_block.h>
  34
  35 #include <QStringList>
  36 #include <QFile>
  37
  38 /*
  39  * Constructors + destructor
  40  */
  41
  42 /// Default constructor.
  43 KwBibleModuleBibleGateway::KwBibleModuleBibleGateway(int id)
  44 : KwBibleModule()
  45 {
  46   KUrl url(QString("http://www.biblegateway.com/versions/index.php?action=getVersionInfo&vid=%1").arg(id));
  47
  48   QString tmpFile;
  49   if (KIO::NetAccess::download(url, tmpFile, 0))
  50   {
  51     QFile file(tmpFile);
  52     if (file.open(QFile::ReadOnly | QFile::Text))
  53     {
  54       QByteArray rawPage = file.readAll();
  55       file.close();
  56
  57       DOM::HTMLDocument doc;
  58       doc.loadXML(QString::fromUtf8(rawPage));
  59       DOM::Element bookList = doc.getElementById("booklist");
  60       bool tableFound = false;
  61       if (!bookList.isNull())
  62       {
  63         // Get the next table
  64         DOM::Node sibling = bookList.nextSibling();
  65         while (!sibling.isNull() && sibling.nodeType() != DOM::Node::ELEMENT_NODE)
  66         {
  67           sibling = sibling.nextSibling();
  68         }
  69         DOM::Element tableElement(sibling);
  70         if (!tableElement.isNull() && tableElement.tagName() == "table")
  71         {
  72           tableFound = true;
  73           // Each row except header is a book
  74           DOM::NodeList rows = tableElement.getElementsByTagName("tr");
  75           for (unsigned int row = 0; row < rows.length(); ++row)
  76           {
  77             DOM::NodeList cells = DOM::Element(rows.item(row)).getElementsByTagName("td");
  78             if (cells.length() == 2)
  79             {
  80               // First cell is the name
  81               m_bookList.push_back(Book());
  82               Book* book = &m_bookList[m_bookList.size()-1];
  83               book->name = DOM::HTMLElement(cells.item(0)).innerText().string();
  84               // Second cell is the chapter links
  85               DOM::NodeList chapterLinks = DOM::Element(cells.item(1)).getElementsByTagName("a");
  86               for (unsigned int chapter = 0; chapter < chapterLinks.length(); ++chapter)
  87               {
  88                 // Check the chapter number is right
  89                 DOM::HTMLElement link = chapterLinks.item(chapter);
  90                 bool numeric;
  91                 int check = link.innerText().string().toInt(&numeric);
  92                 if (!numeric || check != (int)chapter+1)
  93                 {
  94                   KMessageBox::error(0, i18n("Error parsing webpage: %1", i18n("Non sequential chapter list in book '%1'", book->name)));
  95                   break;
  96                 }
  97                 // Get the link
  98                 book->chapters.push_back(Chapter());
  99                 Chapter* chapter = &book->chapters[book->chapters.size()-1];
 100                 chapter->url = "http://www.biblegateway.com" + link.getAttribute("href").string();
 101                 chapter->fetched = false;
 102               }
 103             }
 104           }
 105         }
 106       }
 107       if (!tableFound)
 108       {
 109         // Book list reference node not found
 110         KMessageBox::error(0, i18n("Error parsing webpage: %1", i18n("Book list table not found")));
 111       }
 112     }
 113
 114     KIO::NetAccess::removeTempFile(tmpFile);
 115   }
 116   else
 117   {
 118     KMessageBox::error(0, KIO::NetAccess::lastErrorString());
 119   }
 120 }
 121
 122 /// Destructor.
 123 KwBibleModuleBibleGateway::~KwBibleModuleBibleGateway()
 124 {
 125 }
 126
 127 /*
 128  * Main interface
 129  */
 130
 131 QString KwBibleModuleBibleGateway::name()
 132 {
 133   return QString();
 134 }
 135
 136 QString KwBibleModuleBibleGateway::description()
 137 {
 138   return QString();
 139 }
 140
 141 int KwBibleModuleBibleGateway::numChapters(int book)
 142 {
 143   if (book >= 0 && book < m_bookList.size())
 144   {
 145     return m_bookList[book].chapters.size();
 146   }
 147   return 0;
 148 }
 149
 150 int KwBibleModuleBibleGateway::numVerses(int book, int chapter)
 151 {
 152   Chapter* chap = fetchChapter(book, chapter);
 153   if (0 != chap)
 154   {
 155     return chap->verses.size();
 156   }
 157   else
 158   {
 159     return 0;
 160   }
 161 }
 162
 163 bool KwBibleModuleBibleGateway::fillPassage(const Key& key, KwBiblePassage* outPassage) const
 164 {
 165   return false;
 166 }
 167
 168 QString KwBibleModuleBibleGateway::renderText(const KwBibleModule::Key& key)
 169 {
 170   QString result;
 171   Chapter* chapter = fetchChapter(key.start.book, key.start.chapter);
 172   if (0 != chapter)
 173   {
 174     if (key.start.verse >= 0 && key.start.verse < chapter->verses.size())
 175     {
 176       return chapter->verses[key.start.verse];
 177     }
 178     else
 179     {
 180       return chapter->verses.join("");
 181     }
 182   }
 183   return result;
 184 }
 185
 186 /*
 187  * Protected virtual interface
 188  */
 189
 190 void KwBibleModuleBibleGateway::obtainBooks()
 191 {
 192   QStringList list;
 193   for (int book = 0; book < m_bookList.size(); ++book)
 194   {
 195     list << m_bookList[book].name;
 196   }
 197   setBooks(list);
 198 }
 199
 200 /*
 201  * Private functions
 202  */
 203
 204 /// Ensure chapter contents are fetched.
 205 KwBibleModuleBibleGateway::Chapter* KwBibleModuleBibleGateway::fetchChapter(int book, int chapter)
 206 {
 207   if (book >= 0 && book < m_bookList.size())
 208   {
 209     Book* bookObj = &m_bookList[book];
 210     if (chapter >= 0 && chapter < m_bookList[book].chapters.size())
 211     {
 212       Chapter* chap = &bookObj->chapters[chapter];
 213       if (!chap->fetched)
 214       {
 215         QString tmpFile;
 216         if (KIO::NetAccess::download(chap->url, tmpFile, 0))
 217         {
 218           QFile file(tmpFile);
 219           if (file.open(QFile::ReadOnly | QFile::Text))
 220           {
 221             QByteArray rawPage = file.readAll();
 222             file.close();
 223
 224             DOM::HTMLDocument doc;
 225             doc.loadXML(QString::fromUtf8(rawPage));
 226
 227             // Find all spans with class="sup"
 228             DOM::NodeList sups = doc.body().getElementsByClassName("sup");
 229             int verse = 0;
 230             for (unsigned int i = 0; i < sups.length(); ++i)
 231             {
 232               DOM::HTMLElement span = sups.item(i);
 233               if (span.tagName() == "span")
 234               {
 235                 // Get the verse number and validate
 236                 bool numeric;
 237                 QString verseNumber = span.innerText().string();
 238                 int check = verseNumber.toInt(&numeric);
 239                 if (!numeric)
 240                 {
 241                   // Its not going to be a verse if it isn't numeric
 242                   KMessageBox::error(0, i18n("Error parsing webpage: %1", i18n("Non numeric superscript encountered: '%1'. It may correspond to a verse range.", verseNumber)));
 243                   continue;
 244                 }
 245                 ++verse;
 246                 if (check != verse)
 247                 {
 248                   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)));
 249                   break;
 250                 }
 251                 QString content;
 252
 253                 // Get any headers before it
 254                 DOM::Node sibling;
 255                 sibling = span.previousSibling();
 256                 while (!sibling.isNull())
 257                 {
 258                   DOM::Element siblingElement = sibling;
 259                   if (!siblingElement.isNull())
 260                   {
 261                     // Stop at a span class="sup"
 262                     if (siblingElement.tagName() == "span" && siblingElement.getAttribute("class") == "sup")
 263                     {
 264                       break;
 265                     }
 266                     // See if its an interesting heading
 267                     DOM::HTMLHeadingElement heading = siblingElement;
 268                     if (!heading.isNull())
 269                     {
 270                       content = heading.toHTML() + content;
 271                     }
 272                   }
 273                   sibling = sibling.previousSibling();
 274                 }
 275
 276                 // Add verse number
 277                 content += QString("<sup>%1</sup>").arg(verse);
 278
 279                 // Get any text after it until the next span
 280                 sibling = span.nextSibling();
 281                 while (!sibling.isNull())
 282                 {
 283                   DOM::Element siblingElement = sibling;
 284                   if (!siblingElement.isNull())
 285                   {
 286                     // Stop at a span class="sup"
 287                     if (siblingElement.tagName() == "span" && siblingElement.getAttribute("class") == "sup")
 288                     {
 289                       break;
 290                     }
 291                     // Also stop at headings
 292                     DOM::HTMLHeadingElement heading = siblingElement;
 293                     if (!heading.isNull())
 294                     {
 295                       break;
 296                     }
 297                   }
 298                   content += sibling.toHTML();
 299                   sibling = sibling.nextSibling();
 300                 }
 301
 302                 chap->verses.push_back(content);
 303               }
 304             }
 305           }
 306
 307           KIO::NetAccess::removeTempFile(tmpFile);
 308         }
 309         else
 310         {
 311           KMessageBox::error(0, KIO::NetAccess::lastErrorString());
 312         }
 313         chap->fetched = true;
 314       }
 315
 316       return chap;
 317     }
 318   }
 319   return 0;
 320 }
 321