goldendict-ng/dsl.cc

2396 lines
69 KiB
C++
Raw Normal View History

2012-02-20 21:47:14 +00:00
/* This file is (c) 2008-2012 Konstantin Isakov <ikm@goldendict.org>
* Part of GoldenDict. Licensed under GPLv3 or later, see the LICENSE file */
#include "dsl.hh"
#include "dsl_details.hh"
#include "btreeidx.hh"
#include "folding.hh"
#include "utf8.hh"
#include "chunkedstorage.hh"
#include "dictzip.h"
#include "htmlescape.hh"
#include "iconv.hh"
#include "filetype.hh"
#include "fsencoding.hh"
#include "audiolink.hh"
2009-04-22 21:37:32 +00:00
#include "langcoder.hh"
#include "wstring_qt.hh"
#include "zipfile.hh"
#include "indexedzip.hh"
#include "gddebug.hh"
#include "tiff.hh"
2014-04-16 16:18:28 +00:00
#include "fulltextsearch.hh"
#include "ftshelpers.hh"
#include "language.hh"
2009-04-22 21:37:32 +00:00
#include <zlib.h>
#include <map>
#include <set>
#include <string>
#include <vector>
#include <list>
#include <wctype.h>
#ifdef _MSC_VER
#include <stub_msvc.h>
#endif
#include <QSemaphore>
#include <QThreadPool>
#include <QAtomicInt>
#include <QUrl>
#include <QDir>
#include <QFileInfo>
#include <QPainter>
2014-04-16 16:18:28 +00:00
#include <QMap>
#include <QStringList>
#include <QRegularExpression>
// For TIFF conversion
#include <QImage>
#include <QByteArray>
#include <QBuffer>
2013-05-25 12:07:49 +00:00
// For SVG handling
#include <QtSvg/QSvgRenderer>
2022-06-19 12:24:34 +00:00
#include <QtConcurrent>
#include "utils.hh"
namespace Dsl {
using namespace Details;
using std::map;
using std::multimap;
using std::pair;
using std::set;
using std::string;
using gd::wstring;
using gd::wchar;
using std::vector;
using std::list;
2021-11-06 08:26:30 +00:00
using Utf8::Encoding;
using BtreeIndexing::WordArticleLink;
using BtreeIndexing::IndexedWords;
using BtreeIndexing::IndexInfo;
namespace {
DEF_EX_STR( exCantReadFile, "Can't read file", Dictionary::Ex )
2014-04-16 16:18:28 +00:00
DEF_EX( exUserAbort, "User abort", Dictionary::Ex )
2014-04-24 18:50:47 +00:00
DEF_EX_STR( exDictzipError, "DICTZIP error", Dictionary::Ex )
enum
{
Signature = 0x584c5344, // DSLX on little-endian, XLSD on big-endian
CurrentFormatVersion = 23 + BtreeIndexing::FormatVersion + Folding::Version,
CurrentZipSupportVersion = 2,
CurrentFtsIndexVersion = 7
};
struct IdxHeader
{
uint32_t signature; // First comes the signature, DSLX
uint32_t formatVersion; // File format version (CurrentFormatVersion)
uint32_t zipSupportVersion; // Zip support version -- narrows down reindexing
// when it changes only for dictionaries with the
// zip files
int dslEncoding; // Which encoding is used for the file indexed
uint32_t chunksOffset; // The offset to chunks' storage
uint32_t hasAbrv; // Non-zero means file has abrvs at abrvAddress
uint32_t abrvAddress; // Address of abrv map in the chunked storage
uint32_t indexBtreeMaxElements; // Two fields from IndexInfo
uint32_t indexRootOffset;
uint32_t articleCount; // Number of articles this dictionary has
uint32_t wordCount; // Number of headwords this dictionary has
2009-04-22 21:37:32 +00:00
uint32_t langFrom; // Source language
uint32_t langTo; // Target language
uint32_t hasZipFile; // Non-zero means there's a zip file with resources
// present
uint32_t hasSoundDictionaryName;
uint32_t zipIndexBtreeMaxElements; // Two fields from IndexInfo of the zip
// resource index.
uint32_t zipIndexRootOffset;
}
#ifndef _MSC_VER
__attribute__((packed))
#endif
;
struct InsidedCard
{
uint32_t offset;
uint32_t size;
QVector< wstring > headwords;
InsidedCard( uint32_t _offset, uint32_t _size, QVector< wstring > const & words ) :
offset( _offset ), size( _size ), headwords( words )
{}
InsidedCard() {}
};
bool indexIsOldOrBad( string const & indexFile, bool hasZipFile )
{
File::Class idx( indexFile, "rb" );
IdxHeader header;
return idx.readRecords( &header, sizeof( header ), 1 ) != 1 ||
header.signature != Signature ||
header.formatVersion != CurrentFormatVersion ||
(bool) header.hasZipFile != hasZipFile ||
( hasZipFile && header.zipSupportVersion != CurrentZipSupportVersion );
}
class DslDictionary: public BtreeIndexing::BtreeDictionary
{
Mutex idxMutex;
File::Class idx;
IdxHeader idxHeader;
sptr< ChunkedStorage::Reader > chunks;
string dictionaryName;
string preferredSoundDictionary;
map< string, string > abrv;
Mutex dzMutex;
dictData * dz;
Mutex resourceZipMutex;
IndexedZip resourceZip;
BtreeIndex resourceZipIndex;
QAtomicInt deferredInitDone;
Mutex deferredInitMutex;
bool deferredInitRunnableStarted;
QSemaphore deferredInitRunnableExited;
string initError;
int optionalPartNom;
quint8 articleNom;
2012-12-07 11:59:29 +00:00
int maxPictureWidth;
wstring currentHeadword;
string resourceDir1, resourceDir2;
public:
DslDictionary( string const & id, string const & indexFile,
2012-12-07 11:59:29 +00:00
vector< string > const & dictionaryFiles,
int maxPictureWidth_ );
void deferredInit() override;
~DslDictionary();
string getName() noexcept override
{ return dictionaryName; }
map< Dictionary::Property, string > getProperties() noexcept override
{ return map< Dictionary::Property, string >(); }
unsigned long getArticleCount() noexcept override
{ return idxHeader.articleCount; }
unsigned long getWordCount() noexcept override
{ return idxHeader.wordCount; }
inline quint32 getLangFrom() const override
2009-04-22 21:37:32 +00:00
{ return idxHeader.langFrom; }
inline quint32 getLangTo() const override
2009-04-22 21:37:32 +00:00
{ return idxHeader.langTo; }
inline virtual string getResourceDir1() const
{ return resourceDir1; }
inline virtual string getResourceDir2() const
{ return resourceDir2; }
2021-11-17 14:19:32 +00:00
sptr< Dictionary::DataRequest > getArticle( wstring const &,
vector< wstring > const & alts,
wstring const &,
bool ignoreDiacritics ) override
2022-01-09 08:35:07 +00:00
;
sptr< Dictionary::DataRequest > getResource( string const & name ) override
2022-01-09 08:35:07 +00:00
;
sptr< Dictionary::DataRequest > getSearchResults( QString const & searchString,
2014-04-16 16:18:28 +00:00
int searchMode, bool matchCase,
int distanceBetweenWords,
int maxResults,
bool ignoreWordsOrder,
bool ignoreDiacritics ) override;
QString const& getDescription() override;
2012-09-07 16:17:44 +00:00
QString getMainFilename() override;
void getArticleText( uint32_t articleAddress, QString & headword, QString & text ) override;
2014-04-16 16:18:28 +00:00
void makeFTSIndex(QAtomicInt & isCancelled, bool firstIteration ) override;
2014-04-16 16:18:28 +00:00
void setFTSParameters( Config::FullTextSearch const & fts ) override
{
if( ensureInitDone().size() )
return;
can_FTS = fts.enabled
&& !fts.disabledTypes.contains( "DSL", Qt::CaseInsensitive )
&& ( fts.maxDictionarySize == 0 || getArticleCount() <= fts.maxDictionarySize );
}
uint32_t getFtsIndexVersion() override
{ return CurrentFtsIndexVersion; }
protected:
void loadIcon() noexcept override;
private:
string const & ensureInitDone() override;
void doDeferredInit();
/// Loads the article. Does not process the DSL language.
void loadArticle( uint32_t address,
wstring const & requestedHeadwordFolded,
bool ignoreDiacritics,
wstring & tildeValue,
wstring & displayedHeadword,
unsigned & headwordIndex,
wstring & articleText );
/// Converts DSL language to an Html.
string dslToHtml( wstring const &, wstring const & headword = wstring() );
// Parts of dslToHtml()
string nodeToHtml( ArticleDom::Node const & );
string processNodeChildren( ArticleDom::Node const & node );
bool hasHiddenZones() /// Return true if article has hidden zones
{ return optionalPartNom != 0; }
friend class DslArticleRequest;
friend class DslResourceRequest;
2014-04-16 16:18:28 +00:00
friend class DslFTSResultsRequest;
friend class DslDeferredInitRunnable;
};
DslDictionary::DslDictionary( string const & id,
string const & indexFile,
2012-12-07 11:59:29 +00:00
vector< string > const & dictionaryFiles,
int maxPictureWidth_ ):
BtreeDictionary( id, dictionaryFiles ),
idx( indexFile, "rb" ),
idxHeader( idx.read< IdxHeader >() ),
dz( 0 ),
deferredInitRunnableStarted( false ),
optionalPartNom( 0 ),
2012-12-07 11:59:29 +00:00
articleNom( 0 ),
maxPictureWidth( maxPictureWidth_ )
{
2014-04-16 16:18:28 +00:00
can_FTS = true;
ftsIdxName = indexFile + Dictionary::getFtsSuffix();
2014-04-16 16:18:28 +00:00
if( !Dictionary::needToRebuildIndex( dictionaryFiles, ftsIdxName )
&& !FtsHelpers::ftsIndexIsOldOrBad( ftsIdxName, this ) )
FTS_index_completed.ref();
2014-04-16 16:18:28 +00:00
// Read the dictionary name
idx.seek( sizeof( idxHeader ) );
vector< char > dName( idx.read< uint32_t >() );
2019-01-17 14:53:13 +00:00
if( dName.size() > 0 )
{
idx.read( &dName.front(), dName.size() );
dictionaryName = string( &dName.front(), dName.size() );
}
vector< char > sName( idx.read< uint32_t >() );
2019-01-17 14:53:13 +00:00
if( sName.size() > 0 )
{
idx.read( &sName.front(), sName.size() );
preferredSoundDictionary = string( &sName.front(), sName.size() );
}
resourceDir1 = getDictionaryFilenames()[ 0 ] + ".files" + FsEncoding::separator();
QString s = QString::fromStdString( getDictionaryFilenames()[ 0 ] );
if( s.endsWith( QString::fromLatin1( ".dz" ), Qt::CaseInsensitive ) )
s.chop( 3 );
resourceDir2 = s.toStdString() + ".files" + FsEncoding::separator();
// Everything else would be done in deferred init
}
DslDictionary::~DslDictionary()
{
Mutex::Lock _( deferredInitMutex );
// Wait for init runnable to complete if it was ever started
2022-04-10 01:33:55 +00:00
// if ( deferredInitRunnableStarted )
// deferredInitRunnableExited.acquire();
if ( dz )
dict_data_close( dz );
}
//////// DslDictionary::deferredInit()
void DslDictionary::deferredInit()
{
if ( !Utils::AtomicInt::loadAcquire( deferredInitDone ) )
{
Mutex::Lock _( deferredInitMutex );
if ( Utils::AtomicInt::loadAcquire( deferredInitDone ) )
return;
if ( !deferredInitRunnableStarted )
{
QThreadPool::globalInstance()->start( [ this ]() { this->doDeferredInit(); }, -1000 );
deferredInitRunnableStarted = true;
}
}
}
string const & DslDictionary::ensureInitDone()
{
// Simple, really.
doDeferredInit();
return initError;
}
void DslDictionary::doDeferredInit()
{
if ( !Utils::AtomicInt::loadAcquire( deferredInitDone ) )
{
Mutex::Lock _( deferredInitMutex );
if ( Utils::AtomicInt::loadAcquire( deferredInitDone ) )
return;
// Do deferred init
try
{
// Don't lock index file - no one should be working with it until
// the init is complete.
//Mutex::Lock _( idxMutex );
chunks = std::shared_ptr<ChunkedStorage::Reader>(new ChunkedStorage::Reader(idx, idxHeader.chunksOffset));
// Open the .dsl file
2014-04-24 18:50:47 +00:00
DZ_ERRORS error;
dz = dict_data_open( getDictionaryFilenames()[ 0 ].c_str(), &error, 0 );
if ( !dz )
2014-04-24 18:50:47 +00:00
throw exDictzipError( string( dz_error_str( error ) )
+ "(" + getDictionaryFilenames()[ 0 ] + ")" );
// Read the abrv, if any
if ( idxHeader.hasAbrv )
{
vector< char > chunk;
char * abrvBlock = chunks->getBlock( idxHeader.abrvAddress, chunk );
uint32_t total;
memcpy( &total, abrvBlock, sizeof( uint32_t ) );
abrvBlock += sizeof( uint32_t );
GD_DPRINTF( "Loading %u abbrv\n", total );
while( total-- )
{
uint32_t keySz;
memcpy( &keySz, abrvBlock, sizeof( uint32_t ) );
abrvBlock += sizeof( uint32_t );
char * key = abrvBlock;
abrvBlock += keySz;
uint32_t valueSz;
memcpy( &valueSz, abrvBlock, sizeof( uint32_t ) );
abrvBlock += sizeof( uint32_t );
abrv[ string( key, keySz ) ] = string( abrvBlock, valueSz );
abrvBlock += valueSz;
}
}
// Initialize the index
openIndex( IndexInfo( idxHeader.indexBtreeMaxElements,
idxHeader.indexRootOffset ),
idx, idxMutex );
// Open a resource zip file, if there's one
if ( idxHeader.hasZipFile &&
( idxHeader.zipIndexBtreeMaxElements ||
idxHeader.zipIndexRootOffset ) )
{
resourceZip.openIndex( IndexInfo( idxHeader.zipIndexBtreeMaxElements,
idxHeader.zipIndexRootOffset ),
idx, idxMutex );
QString zipName = QDir::fromNativeSeparators( getDictionaryFilenames().back().c_str() );
if ( zipName.endsWith( ".zip", Qt::CaseInsensitive ) ) // Sanity check
resourceZip.openZipFile( zipName );
}
}
catch( std::exception & e )
{
initError = e.what();
}
catch( ... )
{
initError = "Unknown error";
}
deferredInitDone.ref();
}
}
void DslDictionary::loadIcon() noexcept
{
if ( dictionaryIconLoaded )
return;
QString fileName = QDir::fromNativeSeparators( getDictionaryFilenames()[ 0 ].c_str() );
// Remove the extension
if ( fileName.endsWith( ".dsl.dz", Qt::CaseInsensitive ) )
fileName.chop( 6 );
else
fileName.chop( 3 );
if ( !loadIconFromFile( fileName ) )
{
// Load failed -- use default icons
dictionaryIcon = QIcon(":/icons/icon32_dsl.png");
dictionaryNativeIcon = QIcon(":/icons/icon_dsl_native.png");
}
dictionaryIconLoaded = true;
}
/// Determines whether or not this char is treated as whitespace for dsl
/// parsing or not. We can't rely on any Unicode standards here, since the
/// only standard that matters here is the original Dsl compiler's insides.
/// Some dictionaries, for instance, are known to specifically use a non-
/// breakable space (0xa0) to indicate that a headword begins with a space,
/// so nbsp is not a whitespace character for Dsl compiler.
/// For now we have only space and tab, since those are most likely the only
/// ones recognized as spaces by that compiler.
bool isDslWs( wchar ch )
{
switch( ch )
{
case ' ':
case '\t':
return true;
default:
return false;
}
}
void DslDictionary::loadArticle( uint32_t address,
wstring const & requestedHeadwordFolded,
bool ignoreDiacritics,
wstring & tildeValue,
wstring & displayedHeadword,
unsigned & headwordIndex,
wstring & articleText )
{
wstring articleData;
{
vector< char > chunk;
2009-04-21 19:03:16 +00:00
char * articleProps;
{
Mutex::Lock _( idxMutex );
2009-04-21 19:03:16 +00:00
articleProps = chunks->getBlock( address, chunk );
}
2009-04-21 19:03:16 +00:00
uint32_t articleOffset, articleSize;
2009-04-21 19:03:16 +00:00
memcpy( &articleOffset, articleProps, sizeof( articleOffset ) );
memcpy( &articleSize, articleProps + sizeof( articleOffset ),
sizeof( articleSize ) );
GD_DPRINTF( "offset = %x\n", articleOffset );
char * articleBody;
{
Mutex::Lock _( dzMutex );
articleBody = dict_data_read_( dz, articleOffset, articleSize, 0, 0 );
}
2009-04-21 19:03:16 +00:00
if ( !articleBody )
{
// throw exCantReadFile( getDictionaryFilenames()[ 0 ] );
articleData = U"\n\r\t" + gd::toWString( QString( "DICTZIP error: " ) + dict_error_str( dz ) );
}
else
{
try
{
articleData =
Iconv::toWstring(
2021-11-06 08:26:30 +00:00
Utf8::getEncodingNameFor( Encoding( idxHeader.dslEncoding ) ),
articleBody, articleSize );
free( articleBody );
// Strip DSL comments
bool b = false;
stripComments( articleData, b );
}
catch( ... )
{
free( articleBody );
throw;
}
}
}
size_t pos = 0;
bool hadFirstHeadword = false;
bool foundDisplayedHeadword = false;
// Check is we retrieve insided card
bool insidedCard = isDslWs( articleData.at( 0 ) );
2009-05-10 16:28:20 +00:00
wstring tildeValueWithUnsorted; // This one has unsorted parts left
for( headwordIndex = 0; ; )
{
size_t begin = pos;
pos = articleData.find_first_of( U"\n\r" , begin );
if ( pos == wstring::npos )
pos = articleData.size();
if ( !foundDisplayedHeadword )
{
// Process the headword
2009-05-10 16:28:20 +00:00
wstring rawHeadword = wstring( articleData, begin, pos - begin );
if( insidedCard && !rawHeadword.empty() && isDslWs( rawHeadword[ 0 ] ) )
{
// Headword of the insided card
wstring::size_type hpos = rawHeadword.find( L'@' );
if( hpos != string::npos )
{
wstring head = Folding::trimWhitespace( rawHeadword.substr( hpos + 1 ) );
hpos = head.find( L'~' );
while( hpos != string::npos )
{
if( hpos == 0 || head[ hpos ] != L'\\' )
break;
hpos = head.find( L'~', hpos + 1 );
}
if( hpos == string::npos )
rawHeadword = head;
else
rawHeadword.clear();
}
}
if( !rawHeadword.empty() )
{
if ( !hadFirstHeadword )
{
// We need our tilde expansion value
tildeValue = rawHeadword;
list< wstring > lst;
2009-05-10 16:28:20 +00:00
expandOptionalParts( tildeValue, &lst );
2009-05-10 16:28:20 +00:00
if ( lst.size() ) // Should always be
tildeValue = lst.front();
tildeValueWithUnsorted = tildeValue;
processUnsortedParts( tildeValue, false );
}
wstring str = rawHeadword;
2009-05-10 16:28:20 +00:00
if ( hadFirstHeadword )
expandTildes( str, tildeValueWithUnsorted );
processUnsortedParts( str, true );
str = Folding::applySimpleCaseOnly( str );
list< wstring > lst;
expandOptionalParts( str, &lst );
// Does one of the results match the requested word? If so, we'd choose
// it as our headword.
2009-05-10 16:28:20 +00:00
for( list< wstring >::iterator i = lst.begin(); i != lst.end(); ++i )
{
unescapeDsl( *i );
normalizeHeadword( *i );
2009-05-10 16:28:20 +00:00
bool found;
if( ignoreDiacritics )
found = Folding::applyDiacriticsOnly( Folding::trimWhitespace( *i ) ) == Folding::applyDiacriticsOnly( requestedHeadwordFolded );
else
found = Folding::trimWhitespace( *i ) == requestedHeadwordFolded;
if ( found )
{
// Found it. Now we should make a displayed headword for it.
if ( hadFirstHeadword )
expandTildes( rawHeadword, tildeValueWithUnsorted );
processUnsortedParts( rawHeadword, false );
displayedHeadword = rawHeadword;
foundDisplayedHeadword = true;
break;
}
}
if ( !foundDisplayedHeadword )
{
++headwordIndex;
hadFirstHeadword = true;
}
}
}
if ( pos == articleData.size() )
break;
// Skip \n\r
if ( articleData[ pos ] == '\r' )
++pos;
if ( pos != articleData.size() )
{
if ( articleData[ pos ] == '\n' )
++pos;
}
if ( pos == articleData.size() )
{
// Ok, it's end of article
break;
}
if( isDslWs( articleData[ pos ] ) )
{
// Check for begin article text
if( insidedCard )
{
// Check for next insided headword
wstring::size_type hpos = articleData.find_first_of( U"\n\r" , pos );
if ( hpos == wstring::npos )
hpos = articleData.size();
wstring str = wstring( articleData, pos, hpos - pos );
hpos = str.find( L'@');
if( hpos == wstring::npos || str[ hpos - 1 ] == L'\\' || !isAtSignFirst( str ) )
break;
}
else
break;
}
}
if ( !foundDisplayedHeadword )
{
// This is strange. Anyway, use tilde expansion value, it's better
// than nothing (or requestedHeadwordFolded for insided card.
if( insidedCard )
displayedHeadword = requestedHeadwordFolded;
else
displayedHeadword = tildeValue;
}
if ( pos != articleData.size() )
articleText = wstring( articleData, pos );
else
articleText.clear();
}
string DslDictionary::dslToHtml( wstring const & str, wstring const & headword )
{
// Normalize the string
wstring normalizedStr = gd::normalize( str );
currentHeadword = headword;
ArticleDom dom( normalizedStr, getName(), headword );
optionalPartNom = 0;
string html = processNodeChildren( dom.root );
2013-10-06 15:24:19 +00:00
return html;
}
string DslDictionary::processNodeChildren( ArticleDom::Node const & node )
{
string result;
for( ArticleDom::Node::const_iterator i = node.begin(); i != node.end();
++i )
result += nodeToHtml( *i );
return result;
}
string DslDictionary::nodeToHtml( ArticleDom::Node const & node )
{
2013-10-06 15:24:19 +00:00
string result;
if ( !node.isTag )
2013-10-06 15:24:19 +00:00
{
result = Html::escape( Utf8::encode( node.text ) );
2013-10-06 15:24:19 +00:00
// Handle all end-of-line
string::size_type n;
// Strip all '\r'
while( ( n = result.find( '\r' ) ) != string::npos )
result.erase( n, 1 );
// Replace all '\n'
while( ( n = result.find( '\n' ) ) != string::npos )
result.replace( n, 1, "<p></p>" );
return result;
}
if( node.tagName == U"b" )
result += "<b class=\"dsl_b\">" + processNodeChildren( node ) + "</b>";
else if( node.tagName == U"i" )
result += "<i class=\"dsl_i\">" + processNodeChildren( node ) + "</i>";
else if( node.tagName == U"u" )
{
string nodeText = processNodeChildren( node );
if ( nodeText.size() && isDslWs( nodeText[ 0 ] ) )
result.push_back( ' ' ); // Fix a common problem where in "foo[i] bar[/i]"
// the space before "bar" gets underlined.
result += "<span class=\"dsl_u\">" + nodeText + "</span>";
}
else if( node.tagName == U"c" )
{
if( node.tagAttrs.empty() )
{
result += "<span class=\"c_default_color\">"
+ processNodeChildren( node ) + "</span>";
}
else
{
result += "<font color=\""
+ Html::escape( Utf8::encode( node.tagAttrs ) ) + "\">"
+ processNodeChildren( node ) + "</font>";
}
}
else if( node.tagName == U"*" )
{
string id = "O" + getId().substr( 0, 7 ) + "_" +
QString::number( articleNom ).toStdString() +
"_opt_" + QString::number( optionalPartNom++ ).toStdString();
result += R"(<span class="dsl_opt" id=")" + id + "\">" + processNodeChildren( node ) + "</span>";
}
else if( node.tagName == U"m" )
result += "<div class=\"dsl_m\">" + processNodeChildren( node ) + "</div>";
else
if ( node.tagName.size() == 2 && node.tagName[ 0 ] == L'm' &&
iswdigit( node.tagName[ 1 ] ) )
result += "<div class=\"dsl_" + Utf8::encode( node.tagName ) + "\">" + processNodeChildren( node ) + "</div>";
else if( node.tagName == U"trn" )
result += "<span class=\"dsl_trn\">" + processNodeChildren( node ) + "</span>";
else if( node.tagName == U"ex" )
result += "<span class=\"dsl_ex\">" + processNodeChildren( node ) + "</span>";
else if( node.tagName == U"com" )
result += "<span class=\"dsl_com\">" + processNodeChildren( node ) + "</span>";
else if( node.tagName == U"s" || node.tagName == U"video" )
{
2019-09-26 16:40:39 +00:00
string filename = Filetype::simplifyString( Utf8::encode( node.renderAsText() ), false );
string n = resourceDir1 + filename;
if ( Filetype::isNameOfSound( filename ) )
{
// If we have the file here, do the exact reference to this dictionary.
// Otherwise, make a global 'search' one.
2009-04-21 19:03:16 +00:00
bool search = !File::exists( n ) && !File::exists( resourceDir2 + filename )
&& !File::exists( getContainingFolder().toStdString() + FsEncoding::separator() + filename )
&& ( !resourceZip.isOpen() || !resourceZip.hasFile( Utf8::decode( filename ) ) );
QUrl url;
url.setScheme( "gdau" );
url.setHost( QString::fromUtf8( search ? "search" : getId().c_str() ) );
url.setPath( Utils::Url::ensureLeadingSlash( QString::fromUtf8( filename.c_str() ) ) );
if( search && idxHeader.hasSoundDictionaryName )
Utils::Url::setFragment( url, QString::fromUtf8( preferredSoundDictionary.c_str() ) );
string ref = string( "\"" ) + url.toEncoded().data() + "\"";
result += addAudioLink( ref, getId() );
result += "<span class=\"dsl_s_wav\"><a href=" + ref + R"(><img src="qrc:///icons/playsound.png" border="0" align="absmiddle" alt="Play"/></a></span>)";
}
else
if ( Filetype::isNameOfPicture( filename ) )
{
QUrl url;
url.setScheme( "bres" );
url.setHost( QString::fromUtf8( getId().c_str() ) );
url.setPath( Utils::Url::ensureLeadingSlash( QString::fromUtf8( filename.c_str() ) ) );
2012-12-07 11:59:29 +00:00
vector< char > imgdata;
bool resize = false;
try
{
File::loadFromFile( n, imgdata );
}
catch( File::exCantOpen & )
{
try
2012-12-07 11:59:29 +00:00
{
n = resourceDir2 + filename;
File::loadFromFile( n, imgdata );
}
catch( File::exCantOpen & )
{
try
{
n = getContainingFolder().toStdString() + FsEncoding::separator() + filename;
File::loadFromFile( n, imgdata );
}
catch( File::exCantOpen & )
{
// Try reading from zip file
if ( resourceZip.isOpen() )
{
Mutex::Lock _( resourceZipMutex );
resourceZip.loadFile( Utf8::decode( filename ), imgdata );
}
}
2012-12-07 11:59:29 +00:00
}
}
catch(...)
{
}
if( !imgdata.empty() )
{
2013-05-25 12:07:49 +00:00
if( Filetype::isNameOfSvg( filename ) )
{
// We don't need to render svg file now
QSvgRenderer svg;
svg.load( QByteArray::fromRawData( imgdata.data(), imgdata.size() ) );
if( svg.isValid() )
{
QSize imgsize = svg.defaultSize();
resize = maxPictureWidth > 0
&& imgsize.width() > maxPictureWidth;
}
}
else
{
QImage img = QImage::fromData( (unsigned char *) &imgdata.front(),
imgdata.size() );
2014-02-11 17:27:24 +00:00
2013-05-25 12:07:49 +00:00
resize = maxPictureWidth > 0
&& img.width() > maxPictureWidth;
}
2012-12-07 11:59:29 +00:00
}
if( resize )
{
string link( url.toEncoded().data() );
link.replace( 0, 4, "gdpicture" );
result += string( "<a href=\"" ) + link + "\">"
+ "<img src=\"" + url.toEncoded().data()
+ "\" alt=\"" + Html::escape( filename ) + "\""
2012-12-07 16:08:09 +00:00
+ "width=\"" + QString::number( maxPictureWidth).toStdString() + "\"/>"
2012-12-07 11:59:29 +00:00
+ "</a>";
}
else
result += string( "<img src=\"" ) + url.toEncoded().data()
+ "\" alt=\"" + Html::escape( filename ) + "\"/>";
}
else
2013-06-22 16:36:25 +00:00
if ( Filetype::isNameOfVideo( filename ) ) {
QUrl url;
url.setScheme( "gdvideo" );
url.setHost( QString::fromUtf8( getId().c_str() ) );
url.setPath( Utils::Url::ensureLeadingSlash( QString::fromUtf8( filename.c_str() ) ) );
2013-06-22 16:36:25 +00:00
result += string( R"(<a class="dsl_s dsl_video" href=")" ) + url.toEncoded().data() + "\">"
2013-06-22 16:36:25 +00:00
+ "<span class=\"img\"></span>"
+ "<span class=\"filename\">" + processNodeChildren( node ) + "</span>" + "</a>";
}
else
{
// Unknown file type, downgrade to a hyperlink
QUrl url;
url.setScheme( "bres" );
url.setHost( QString::fromUtf8( getId().c_str() ) );
url.setPath( Utils::Url::ensureLeadingSlash( QString::fromUtf8( filename.c_str() ) ) );
result += string( R"(<a class="dsl_s" href=")" ) + url.toEncoded().data()
+ "\">" + processNodeChildren( node ) + "</a>";
}
}
else if( node.tagName == U"url" )
{
2019-09-26 16:40:39 +00:00
string link = Html::escape( Filetype::simplifyString( Utf8::encode( node.renderAsText() ), false ) );
if( QUrl::fromEncoded( link.c_str() ).scheme().isEmpty() )
link = "http://" + link;
QUrl url( QString::fromUtf8( link.c_str() ) );
if( url.isLocalFile() && url.host().isEmpty() )
{
// Convert relative links to local files to absolute ones
QString name = QFileInfo( getMainFilename() ).absolutePath();
name += url.toLocalFile();
QFileInfo info( name );
if( info.isFile() )
{
name = info.canonicalFilePath();
url.setPath( Utils::Url::ensureLeadingSlash( QUrl::fromLocalFile( name ).path() ) );
link = string( url.toEncoded().data() );
}
}
result += R"(<a class="dsl_url" href=")" + link +"\">" + processNodeChildren( node ) + "</a>";
}
else if( node.tagName == U"!trs" )
2014-04-16 16:18:28 +00:00
{
result += "<span class=\"dsl_trs\">" + processNodeChildren( node ) + "</span>";
2014-04-16 16:18:28 +00:00
}
else if( node.tagName == U"p" )
{
result += "<span class=\"dsl_p\"";
string val = Utf8::encode( node.renderAsText() );
// If we have such a key, display a title
map< string, string >::const_iterator i = abrv.find( val );
if ( i != abrv.end() )
{
string title;
if ( Utf8::decode( i->second ).size() < 70 )
{
2009-05-14 22:00:56 +00:00
// Replace all spaces with non-breakable ones, since that's how
// Lingvo shows tooltips
title.reserve( i->second.size() );
for( char const * c = i->second.c_str(); *c; ++c )
{
if ( *c == ' ' || *c == '\t' )
{
// u00A0 in utf8
title.push_back( 0xC2 );
title.push_back( 0xA0 );
}
else
if( *c == '-' ) // Change minus to non-breaking hyphen (uE28091 in utf8)
{
title.push_back( 0xE2 );
title.push_back( 0x80 );
title.push_back( 0x91 );
}
else
title.push_back( *c );
}
}
else
title = i->second;
result += " title=\"" + Html::escape( title ) + "\"";
}
result += ">" + processNodeChildren( node ) + "</span>";
}
else if( node.tagName == U"'" )
{
// There are two ways to display the stress: by adding an accent sign or via font styles.
// We generate two spans, one with accented data and another one without it, so the
// user could pick up the best suitable option.
string data = processNodeChildren( node );
result += R"(<span class="dsl_stress"><span class="dsl_stress_without_accent">)" + data + "</span>"
+ "<span class=\"dsl_stress_with_accent\">" + data + Utf8::encode( wstring( 1, 0x301 ) )
+ "</span></span>";
}
else if( node.tagName == U"lang" )
{
result += "<span class=\"dsl_lang\"";
if( !node.tagAttrs.empty() )
{
// Find ISO 639-1 code
string langcode;
QString attr = gd::toQString( node.tagAttrs );
int n = attr.indexOf( "id=" );
if( n >= 0 )
{
int id = attr.mid( n + 3 ).toInt();
if( id )
langcode = findCodeForDslId( id );
}
else
{
n = attr.indexOf( "name=\"" );
if( n >= 0 )
{
int n2 = attr.indexOf( '\"', n + 6 );
if( n2 > 0 )
{
quint32 id = dslLanguageToId( gd::toWString( attr.mid( n + 6, n2 - n - 6 ) ) );
langcode = LangCoder::intToCode2( id ).toStdString();
}
}
}
if( !langcode.empty() )
result += " lang=\"" + langcode + "\"";
}
result += ">" + processNodeChildren( node ) + "</span>";
}
else if( node.tagName == U"ref" )
{
QUrl url;
url.setScheme( "gdlookup" );
url.setHost( "localhost" );
2022-09-05 23:49:55 +00:00
wstring nodeStr = node.renderAsText();
normalizeHeadword( nodeStr );
url.setPath( Utils::Url::ensureLeadingSlash( gd::toQString( nodeStr ) ) );
if( !node.tagAttrs.empty() )
{
QString attr = gd::toQString( node.tagAttrs ).remove( '\"' );
int n = attr.indexOf( '=' );
if( n > 0 )
{
QList< QPair< QString, QString > > query;
query.append( QPair< QString, QString >( attr.left( n ), attr.mid( n + 1 ) ) );
Utils::Url::setQueryItems( url, query );
}
}
result += string( R"(<a class="dsl_ref" href=")" ) + url.toEncoded().data() +"\">"
2014-04-16 16:18:28 +00:00
+ processNodeChildren( node ) + "</a>";
}
else if( node.tagName == U"@" )
{
// Special case - insided card header was not parsed
QUrl url;
url.setScheme( "gdlookup" );
url.setHost( "localhost" );
wstring nodeStr = node.renderAsText();
2013-06-07 12:58:10 +00:00
normalizeHeadword( nodeStr );
url.setPath( Utils::Url::ensureLeadingSlash( gd::toQString( nodeStr ) ) );
result += string( R"(<a class="dsl_ref" href=")" ) + url.toEncoded().data() +"\">"
2014-04-16 16:18:28 +00:00
+ processNodeChildren( node ) + "</a>";
}
else if( node.tagName == U"sub" )
{
result += "<sub>" + processNodeChildren( node ) + "</sub>";
}
else if( node.tagName == U"sup" )
{
result += "<sup>" + processNodeChildren( node ) + "</sup>";
}
else if( node.tagName == U"t" )
{
result += "<span class=\"dsl_t\">" + processNodeChildren( node ) + "</span>";
}
else if( node.tagName == U"br" )
2013-06-07 12:58:10 +00:00
{
result += "<br />";
}
else
{
gdWarning( R"(DSL: Unknown tag "%s" with attributes "%s" found in "%s", article "%s".)",
gd::toQString( node.tagName ).toUtf8().data(), gd::toQString( node.tagAttrs ).toUtf8().data(),
getName().c_str(), gd::toQString( currentHeadword ).toUtf8().data() );
2020-10-09 17:49:35 +00:00
result += "<span class=\"dsl_unknown\">[" + string( gd::toQString( node.tagName ).toUtf8().data() );
if( !node.tagAttrs.empty() )
result += " " + string( gd::toQString( node.tagAttrs ).toUtf8().data() );
result += "]" + processNodeChildren( node ) + "</span>";
}
return result;
}
2012-09-07 16:17:44 +00:00
QString const& DslDictionary::getDescription()
{
if( !dictionaryDescription.isEmpty() )
return dictionaryDescription;
dictionaryDescription = "NONE";
QString fileName = QDir::fromNativeSeparators( getDictionaryFilenames()[ 0 ].c_str() );
2012-09-07 16:17:44 +00:00
// Remove the extension
if ( fileName.endsWith( ".dsl.dz", Qt::CaseInsensitive ) )
fileName.chop( 6 );
else
fileName.chop( 3 );
fileName += "ann";
QFileInfo info( fileName );
if ( info.exists() )
{
QFile annFile( fileName );
if( !annFile.open( QFile::ReadOnly | QFile::Text ) )
return dictionaryDescription;
QTextStream annStream( &annFile );
QString data, str;
str = annStream.readLine();
if( str.left( 10 ).compare( "#LANGUAGE " ) != 0 )
{
annStream.seek( 0 );
dictionaryDescription = annStream.readAll();
}
else
{
// Multilanguage annotation
qint32 gdLang, annLang;
QString langStr;
gdLang = LangCoder::code2toInt( QLocale::system().name().left( 2 ).toLatin1().data() );
2012-09-07 16:17:44 +00:00
for(;;)
{
data.clear();
langStr = str.mid( 10 ).replace( '\"', ' ').trimmed();
annLang = LangCoder::findIdForLanguage( gd::toWString( langStr ) );
do
{
str = annStream.readLine();
if( str.left( 10 ).compare( "#LANGUAGE " ) == 0 )
break;
if( !str.endsWith( '\n' ) )
str.append( '\n' );
data += str;
}
while ( !annStream.atEnd() );
if( dictionaryDescription.compare( "NONE ") == 0 || langStr.compare( "English", Qt::CaseInsensitive ) == 0 || gdLang == annLang )
dictionaryDescription = data.trimmed();
if( gdLang == annLang || annStream.atEnd() )
break;
}
}
}
return dictionaryDescription;
}
QString DslDictionary::getMainFilename() { return getDictionaryFilenames()[ 0 ].c_str(); }
2014-04-16 16:18:28 +00:00
void DslDictionary::makeFTSIndex( QAtomicInt & isCancelled, bool firstIteration )
{
2014-04-16 16:18:28 +00:00
if( !( Dictionary::needToRebuildIndex( getDictionaryFilenames(), ftsIdxName )
|| FtsHelpers::ftsIndexIsOldOrBad( ftsIdxName, this ) ) )
2014-04-16 16:18:28 +00:00
FTS_index_completed.ref();
2014-04-16 16:18:28 +00:00
if( haveFTSIndex() )
return;
2014-04-16 16:18:28 +00:00
if( ensureInitDone().size() )
return;
if( firstIteration && getArticleCount() > FTS::MaxDictionarySizeForFastSearch )
return;
gdDebug( "Dsl: Building the full-text index for dictionary: %s\n",
getName().c_str() );
try
{
FtsHelpers::makeFTSIndex( this, isCancelled );
FTS_index_completed.ref();
2014-04-16 16:18:28 +00:00
}
catch( std::exception &ex )
{
2014-04-16 16:18:28 +00:00
gdWarning( "DSL: Failed building full-text search index for \"%s\", reason: %s\n", getName().c_str(), ex.what() );
QFile::remove( ftsIdxName.c_str() );
2014-04-16 16:18:28 +00:00
}
}
2014-04-16 16:18:28 +00:00
void DslDictionary::getArticleText( uint32_t articleAddress, QString & headword, QString & text )
{
headword.clear();
text.clear();
2014-04-16 16:18:28 +00:00
vector< char > chunk;
2014-04-16 16:18:28 +00:00
char * articleProps;
wstring articleData;
{
2014-04-16 16:18:28 +00:00
Mutex::Lock _( idxMutex );
articleProps = chunks->getBlock( articleAddress, chunk );
}
uint32_t articleOffset, articleSize;
2014-04-16 16:18:28 +00:00
memcpy( &articleOffset, articleProps, sizeof( articleOffset ) );
memcpy( &articleSize, articleProps + sizeof( articleOffset ),
sizeof( articleSize ) );
2014-04-16 16:18:28 +00:00
char * articleBody;
{
Mutex::Lock _( dzMutex );
articleBody = dict_data_read_( dz, articleOffset, articleSize, 0, 0 );
}
2014-04-16 16:18:28 +00:00
if ( !articleBody )
{
return;
}
else
{
try
{
2014-04-16 16:18:28 +00:00
articleData =
Iconv::toWstring(
2021-11-06 08:26:30 +00:00
getEncodingNameFor( Encoding( idxHeader.dslEncoding ) ),
2014-04-16 16:18:28 +00:00
articleBody, articleSize );
free( articleBody );
// Strip DSL comments
bool b = false;
stripComments( articleData, b );
}
catch( ... )
{
2014-04-16 16:18:28 +00:00
free( articleBody );
return;
}
}
2014-04-16 16:18:28 +00:00
// Skip headword
size_t pos = 0;
wstring articleHeadword, tildeValue;
2014-04-16 16:18:28 +00:00
// Check if we retrieve insided card
bool insidedCard = isDslWs( articleData.at( 0 ) );
2014-04-16 16:18:28 +00:00
for( ; ; )
{
size_t begin = pos;
pos = articleData.find_first_of( U"\n\r" , begin );
2014-04-16 16:18:28 +00:00
if ( articleHeadword.empty() )
2014-04-16 16:18:28 +00:00
{
// Process the headword
articleHeadword = wstring( articleData, begin, pos - begin );
2014-04-16 16:18:28 +00:00
if( insidedCard && !articleHeadword.empty() && isDslWs( articleHeadword[ 0 ] ) )
{
// Headword of the insided card
wstring::size_type hpos = articleHeadword.find( L'@' );
if( hpos != string::npos )
{
wstring head = Folding::trimWhitespace( articleHeadword.substr( hpos + 1 ) );
hpos = head.find( L'~' );
while( hpos != string::npos )
{
if( hpos == 0 || head[ hpos ] != L'\\' )
break;
hpos = head.find( L'~', hpos + 1 );
}
if( hpos == string::npos )
articleHeadword = head;
else
articleHeadword.clear();
}
}
2014-04-16 16:18:28 +00:00
if( !articleHeadword.empty() )
{
list< wstring > lst;
2014-04-16 16:18:28 +00:00
tildeValue = articleHeadword;
processUnsortedParts( articleHeadword, true );
expandOptionalParts( articleHeadword, &lst );
2014-04-16 16:18:28 +00:00
if ( lst.size() ) // Should always be
articleHeadword = lst.front();
}
2014-04-16 16:18:28 +00:00
}
if ( pos == articleData.size() )
break;
// Skip \n\r
if ( articleData[ pos ] == '\r' )
++pos;
if ( pos != articleData.size() )
{
if ( articleData[ pos ] == '\n' )
++pos;
}
if ( pos == articleData.size() )
2014-04-16 16:18:28 +00:00
{
// Ok, it's end of article
2014-04-16 16:18:28 +00:00
break;
}
if( isDslWs( articleData[ pos ] ) )
{
// Check for begin article text
if( insidedCard )
{
// Check for next insided headword
wstring::size_type hpos = articleData.find_first_of( U"\n\r" , pos );
if ( hpos == wstring::npos )
hpos = articleData.size();
wstring str = wstring( articleData, pos, hpos - pos );
hpos = str.find( L'@');
if( hpos == wstring::npos || str[ hpos - 1 ] == L'\\' || !isAtSignFirst( str ) )
break;
}
else
break;
}
}
if( !articleHeadword.empty() )
{
unescapeDsl( articleHeadword );
normalizeHeadword( articleHeadword );
headword = gd::toQString( articleHeadword );
}
2014-04-16 16:18:28 +00:00
wstring articleText;
if ( pos != articleData.size() )
articleText = wstring( articleData, pos );
else
articleText.clear();
if( !tildeValue.empty() )
{
list< wstring > lst;
processUnsortedParts( tildeValue, false );
expandOptionalParts( tildeValue, &lst );
if ( lst.size() ) // Should always be
expandTildes( articleText, lst.front() );
}
if( !articleText.empty() )
{
text = gd::toQString( articleText ).normalized( QString::NormalizationForm_C );
articleText.clear();
// Parse article text
// Strip some areas
const int stripTagsNumber = 5;
static QString stripTags[ stripTagsNumber ] =
{
"s",
"url",
"!trs",
"video",
"preview"
};
static QString stripEndTags[ stripTagsNumber ] =
{
"[/s]",
"[/url]",
"[/!trs]",
"[/video]",
"[/preview]"
};
int pos = 0;
while( pos >= 0 )
{
pos = text.indexOf( '[', pos, Qt::CaseInsensitive );
if( pos >= 0 )
{
if( ( pos > 0 && text[ pos - 1 ] == '\\' && ( pos < 2 || text[ pos - 2 ] != '\\' ) )
|| ( pos > text.size() - 2 || text[ pos + 1 ] == '/' ) )
{
pos += 1;
continue;
}
int pos2 = text.indexOf( ']', pos + 1, Qt::CaseInsensitive );
if( pos2 < 0 )
break;
QString tag = text.mid( pos + 1, pos2 - pos - 1 );
int n;
for( n = 0; n < stripTagsNumber; n++ )
{
if( tag.compare( stripTags[ n ], Qt::CaseInsensitive ) == 0 )
{
pos2 = text.indexOf( stripEndTags[ n ] , pos + stripTags[ n ].size() + 2, Qt::CaseInsensitive );
text.replace( pos, pos2 > 0 ? pos2 - pos + stripEndTags[ n ].length() : text.length() - pos, " " );
break;
}
}
if( n >= stripTagsNumber )
pos += 1;
}
}
// Strip tags
2014-04-16 16:18:28 +00:00
text.replace( QRegularExpression( R"(\[(|/)(p|trn|ex|com|\*|t|br|m[0-9]?)\])" ), " " );
text.replace( QRegularExpression( R"(\[(|/)lang(\s[^\]]*)?\])" ), " " );
text.remove( QRegularExpression( R"(\[[^\\\[\]]+\])" ) );
2021-11-19 13:47:22 +00:00
text.remove( QString::fromLatin1( "<<" ) );
text.remove( QString::fromLatin1( ">>" ) );
2014-04-16 16:18:28 +00:00
// Chech for insided cards
bool haveInsidedCards = false;
pos = 0;
while( pos >= 0 )
{
pos = text.indexOf( "@", pos );
if( pos > 0 && text.at( pos - 1 ) != '\\' )
{
haveInsidedCards = true;
break;
}
2014-04-16 16:18:28 +00:00
if( pos >= 0 )
pos += 1;
}
2014-04-16 16:18:28 +00:00
if( haveInsidedCards )
{
// Use base DSL parser for articles with insided cards
ArticleDom dom( gd::toWString( text ), getName(), articleHeadword );
text = gd::toQString( dom.root.renderAsText( true ) );
}
else
{
// Unescape DSL symbols
pos = 0;
while( pos >= 0 )
{
pos = text.indexOf( '\\', pos );
if( pos >= 0 )
{
if( text[ pos + 1 ] == '\\' )
pos += 1;
2014-04-16 16:18:28 +00:00
text.remove( pos, 1 );
}
}
}
}
}
/// DslDictionary::getArticle()
class DslArticleRequest: public Dictionary::DataRequest
{
wstring word;
vector< wstring > alts;
DslDictionary & dict;
bool ignoreDiacritics;
QAtomicInt isCancelled;
QSemaphore hasExited;
2022-06-19 12:24:34 +00:00
QFuture< void > f;
public:
DslArticleRequest( wstring const & word_,
vector< wstring > const & alts_,
DslDictionary & dict_, bool ignoreDiacritics_ ):
word( word_ ), alts( alts_ ), dict( dict_ ), ignoreDiacritics( ignoreDiacritics_ )
{
2022-06-19 12:24:34 +00:00
f = QtConcurrent::run( [ this ]() { this->run(); } );
// QThreadPool::globalInstance()->start( [ this ]() { this->run(); } );
}
void run();
void cancel() override
{
isCancelled.ref();
}
2009-04-21 19:03:16 +00:00
~DslArticleRequest()
{
isCancelled.ref();
2022-06-19 12:24:34 +00:00
f.waitForFinished();
//hasExited.acquire();
}
};
void DslArticleRequest::run()
{
if ( Utils::AtomicInt::loadAcquire( isCancelled ) )
{
finish();
return;
}
if ( dict.ensureInitDone().size() )
{
setErrorString( QString::fromUtf8( dict.ensureInitDone().c_str() ) );
finish();
return;
}
vector< WordArticleLink > chain = dict.findArticles( word, ignoreDiacritics );
for( unsigned x = 0; x < alts.size(); ++x )
{
/// Make an additional query for each alt
vector< WordArticleLink > altChain = dict.findArticles( alts[ x ], ignoreDiacritics );
chain.insert( chain.end(), altChain.begin(), altChain.end() );
}
// Some synonyms make it that the articles appear several times. We combat
// this by only allowing them to appear once. Dsl treats different headwords
// of the same article as different articles, so we also include headword
// index here.
set< pair< uint32_t, unsigned > > articlesIncluded;
wstring wordCaseFolded = Folding::applySimpleCaseOnly( word );
for( unsigned x = 0; x < chain.size(); ++x )
{
// Check if we're cancelled occasionally
if ( Utils::AtomicInt::loadAcquire( isCancelled ) )
{
finish();
return;
}
// Grab that article
wstring tildeValue;
wstring displayedHeadword;
wstring articleBody;
unsigned headwordIndex;
2013-09-24 13:56:47 +00:00
string articleText, articleAfter;
2013-09-24 13:56:47 +00:00
try
{
dict.loadArticle( chain[ x ].articleOffset, wordCaseFolded, ignoreDiacritics, tildeValue,
2013-09-24 13:56:47 +00:00
displayedHeadword, headwordIndex, articleBody );
2013-09-24 13:56:47 +00:00
if ( !articlesIncluded.insert( std::make_pair( chain[ x ].articleOffset,
headwordIndex ) ).second )
continue; // We already have this article in the body.
2013-09-24 13:56:47 +00:00
dict.articleNom += 1;
2013-09-24 13:56:47 +00:00
if( displayedHeadword.empty() || isDslWs( displayedHeadword[ 0 ] ) )
displayedHeadword = word; // Special case - insided card
articleText += "<div class=\"dsl_article\">";
2013-09-24 13:56:47 +00:00
articleText += "<div class=\"dsl_headwords\"";
if( dict.isFromLanguageRTL() )
articleText += " dir=\"rtl\"";
2013-10-06 17:40:06 +00:00
articleText += "><p>";
2013-09-24 13:56:47 +00:00
if( displayedHeadword.size() == 1 && displayedHeadword[0] == '<' ) // Fix special case - "<" header
articleText += "<"; // dslToHtml can't handle it correctly.
else
articleText += dict.dslToHtml( displayedHeadword, displayedHeadword );
2013-09-24 13:56:47 +00:00
/// After this may be expand button will be inserted
2013-10-06 17:40:06 +00:00
articleAfter += "</p></div>";
2013-09-24 13:56:47 +00:00
expandTildes( articleBody, tildeValue );
2013-09-24 13:56:47 +00:00
articleAfter += "<div class=\"dsl_definition\"";
if( dict.isToLanguageRTL() )
articleAfter += " dir=\"rtl\"";
articleAfter += ">";
articleAfter += dict.dslToHtml( articleBody, displayedHeadword );
2013-09-24 13:56:47 +00:00
articleAfter += "</div>";
articleAfter += "</div>";
2013-09-24 13:56:47 +00:00
if( dict.hasHiddenZones() )
{
string prefix = "O" + dict.getId().substr( 0, 7 ) + "_" + QString::number( dict.articleNom ).toStdString();
string id1 = prefix + "_expand";
string id2 = prefix + "_opt_";
string button = R"( <img src="qrc:///icons/expand_opt.png" class="hidden_expand_opt" id=")" + id1 +
2013-09-24 13:56:47 +00:00
"\" onclick=\"gdExpandOptPart('" + id1 + "','" + id2 +"')\" alt=\"[+]\"/>";
if( articleText.compare( articleText.size() - 4, 4, "</p>" ) == 0 )
articleText.insert( articleText.size() - 4, " " + button );
else
articleText += button;
}
2013-09-24 13:56:47 +00:00
articleText += articleAfter;
}
catch( std::exception &ex )
{
gdWarning( "DSL: Failed loading article from \"%s\", reason: %s\n", dict.getName().c_str(), ex.what() );
2013-09-24 13:56:47 +00:00
articleText = string( "<span class=\"dsl_article\">" )
+ QObject::tr( "Article loading error" ).toStdString()
2013-09-24 13:56:47 +00:00
+ "</span>";
}
Mutex::Lock _( dataMutex );
data.resize( data.size() + articleText.size() );
memcpy( &data.front() + data.size() - articleText.size(),
articleText.data(), articleText.size() );
hasAnyData = true;
}
finish();
}
sptr< Dictionary::DataRequest > DslDictionary::getArticle( wstring const & word,
vector< wstring > const & alts,
wstring const &,
bool ignoreDiacritics )
2022-01-09 08:35:07 +00:00
{
return std::make_shared<DslArticleRequest>( word, alts, *this, ignoreDiacritics );
}
//// DslDictionary::getResource()
class DslResourceRequest: public Dictionary::DataRequest
{
DslDictionary & dict;
string resourceName;
QAtomicInt isCancelled;
QSemaphore hasExited;
2022-06-19 12:24:34 +00:00
QFuture< void > f;
public:
DslResourceRequest( DslDictionary & dict_,
string const & resourceName_ ):
dict( dict_ ),
resourceName( resourceName_ )
{
2022-06-19 12:24:34 +00:00
f = QtConcurrent::run( [ this ]() { this->run(); } );
// QThreadPool::globalInstance()->start( [ this ]() { this->run(); } );
}
void run();
void cancel() override
{
isCancelled.ref();
}
2009-04-21 19:03:16 +00:00
~DslResourceRequest()
{
isCancelled.ref();
2022-06-19 12:24:34 +00:00
f.waitForFinished();
//hasExited.acquire();
}
};
void DslResourceRequest::run()
{
// Some runnables linger enough that they are cancelled before they start
if ( Utils::AtomicInt::loadAcquire( isCancelled ) )
{
finish();
return;
}
if ( dict.ensureInitDone().size() )
{
setErrorString( QString::fromUtf8( dict.ensureInitDone().c_str() ) );
finish();
return;
}
string n = dict.getContainingFolder().toStdString() + FsEncoding::separator() + resourceName;
GD_DPRINTF( "n is %s\n", n.c_str() );
try
{
try
{
Mutex::Lock _( dataMutex );
File::loadFromFile( n, data );
}
catch( File::exCantOpen & )
{
n = dict.getResourceDir1() + resourceName;
try {
Mutex::Lock _( dataMutex );
File::loadFromFile( n, data );
}
catch( File::exCantOpen & )
{
n = dict.getResourceDir2() + resourceName;
try
{
Mutex::Lock _( dataMutex );
File::loadFromFile( n, data );
}
catch( File::exCantOpen & )
{
// Try reading from zip file
if ( dict.resourceZip.isOpen() )
{
Mutex::Lock _( dict.resourceZipMutex );
Mutex::Lock __( dataMutex );
if ( !dict.resourceZip.loadFile( Utf8::decode( resourceName ), data ) )
throw; // Make it fail since we couldn't read the archive
}
else
throw;
}
}
}
if ( Filetype::isNameOfTiff( resourceName ) )
{
// Convert it
2022-04-05 13:25:07 +00:00
Mutex::Lock _( dataMutex );
GdTiff::tiff2img( data );
}
Mutex::Lock _( dataMutex );
hasAnyData = true;
}
2013-09-24 13:56:47 +00:00
catch( std::exception &ex )
{
gdWarning( "DSL: Failed loading resource \"%s\" for \"%s\", reason: %s\n",
resourceName.c_str(), dict.getName().c_str(), ex.what() );
2013-09-24 13:56:47 +00:00
// Resource not loaded -- we don't set the hasAnyData flag then
}
finish();
}
sptr< Dictionary::DataRequest > DslDictionary::getResource( string const & name )
2022-01-09 08:35:07 +00:00
{
return std::make_shared<DslResourceRequest>( *this, name );
}
2014-04-16 16:18:28 +00:00
sptr< Dictionary::DataRequest > DslDictionary::getSearchResults( QString const & searchString,
int searchMode, bool matchCase,
int distanceBetweenWords,
int maxResults,
bool ignoreWordsOrder,
bool ignoreDiacritics )
2014-04-16 16:18:28 +00:00
{
return std::make_shared<FtsHelpers::FTSResultsRequest>( *this, searchString,searchMode, matchCase, distanceBetweenWords, maxResults, ignoreWordsOrder, ignoreDiacritics );
2014-04-16 16:18:28 +00:00
}
} // anonymous namespace
/// makeDictionaries
vector< sptr< Dictionary::Class > > makeDictionaries(
vector< string > const & fileNames,
string const & indicesDir,
2012-12-07 11:59:29 +00:00
Dictionary::Initializing & initializing,
int maxPictureWidth, unsigned int maxHeadwordSize )
2022-01-09 08:35:07 +00:00
{
vector< sptr< Dictionary::Class > > dictionaries;
for( vector< string >::const_iterator i = fileNames.begin(); i != fileNames.end();
++i )
{
// Try .dsl and .dsl.dz suffixes
bool uncompressedDsl = ( i->size() >= 4 &&
strcasecmp( i->c_str() + ( i->size() - 4 ), ".dsl" ) == 0 );
if ( !uncompressedDsl &&
( i->size() < 7 ||
strcasecmp( i->c_str() + ( i->size() - 7 ), ".dsl.dz" ) != 0 ) )
continue;
// Make sure it's not an abbreviation file
int extSize = ( uncompressedDsl ? 4 : 7 );
if ( i->size() - extSize >= 5 &&
strncasecmp( i->c_str() + i->size() - extSize - 5, "_abrv", 5 ) == 0 )
{
// It is, skip it
continue;
}
unsigned atLine = 0; // Indicates current line in .dsl, for debug purposes
try
{
vector< string > dictFiles( 1, *i );
// Check if there is an 'abrv' file present
string baseName = ( (*i)[ i->size() - 4 ] == '.' ) ?
string( *i, 0, i->size() - 4 ) : string( *i, 0, i->size() - 7 );
string abrvFileName;
if ( File::tryPossibleName( baseName + "_abrv.dsl", abrvFileName ) ||
File::tryPossibleName( baseName + "_abrv.dsl.dz", abrvFileName ) ||
File::tryPossibleName( baseName + "_ABRV.DSL", abrvFileName ) ||
File::tryPossibleName( baseName + "_ABRV.DSL.DZ", abrvFileName ) ||
File::tryPossibleName( baseName + "_ABRV.DSL.dz", abrvFileName ) )
dictFiles.push_back( abrvFileName );
string dictId = Dictionary::makeDictionaryId( dictFiles );
// See if there's a zip file with resources present. If so, include it.
string zipFileName;
2017-04-24 14:42:01 +00:00
if ( File::tryPossibleZipName( baseName + ".dsl.files.zip", zipFileName ) ||
File::tryPossibleZipName( baseName + ".dsl.dz.files.zip", zipFileName ) ||
File::tryPossibleZipName( baseName + ".DSL.FILES.ZIP", zipFileName ) ||
File::tryPossibleZipName( baseName + ".DSL.DZ.FILES.ZIP", zipFileName ) )
dictFiles.push_back( zipFileName );
string indexFile = indicesDir + dictId;
if ( Dictionary::needToRebuildIndex( dictFiles, indexFile ) ||
indexIsOldOrBad( indexFile, zipFileName.size() ) )
{
DslScanner scanner( *i );
try { // Here we intercept any errors during the read to save line at
// which the incident happened. We need alive scanner for that.
if( scanner.getDictionaryName() == U"Abbrev" )
continue; // For now just skip abbreviations
// Building the index
initializing.indexingDictionary( Utf8::encode( scanner.getDictionaryName() ) );
gdDebug( "Dsl: Building the index for dictionary: %s\n",
gd::toQString( scanner.getDictionaryName() ).toUtf8().data() );
File::Class idx( indexFile, "wb" );
IdxHeader idxHeader;
memset( &idxHeader, 0, sizeof( idxHeader ) );
// We write a dummy header first. At the end of the process the header
// will be rewritten with the right values.
idx.write( idxHeader );
string dictionaryName = Utf8::encode( scanner.getDictionaryName() );
idx.write( (uint32_t)dictionaryName.size() );
idx.write( dictionaryName.data(), dictionaryName.size() );
string soundDictName = Utf8::encode( scanner.getSoundDictionaryName() );
if( !soundDictName.empty() )
{
idxHeader.hasSoundDictionaryName = 1;
idx.write( (uint32_t)soundDictName.size() );
idx.write( soundDictName.data(), soundDictName.size() );
}
idxHeader.dslEncoding = scanner.getEncoding();
IndexedWords indexedWords;
ChunkedStorage::Writer chunks( idx );
// Read the abbreviations
if ( abrvFileName.size() )
{
try
{
DslScanner abrvScanner( abrvFileName );
map< string, string > abrv;
wstring curString;
size_t curOffset;
for( ; ; )
{
// Skip any whitespace
2021-10-23 04:05:16 +00:00
if ( !abrvScanner.readNextLineWithoutComments( curString, curOffset, true ) )
break;
if ( curString.empty() || isDslWs( curString[ 0 ] ) )
continue;
list< wstring > keys;
bool eof = false;
// Insert the key and read more, or get to the definition
for( ; ; )
{
processUnsortedParts( curString, true );
if ( keys.size() )
expandTildes( curString, keys.front() );
expandOptionalParts( curString, &keys );
if ( !abrvScanner.readNextLineWithoutComments( curString, curOffset ) || curString.empty() )
{
2017-06-22 15:02:04 +00:00
gdWarning( "Premature end of file %s\n", abrvFileName.c_str() );
eof = true;
break;
}
if ( isDslWs( curString[ 0 ] ) )
break;
}
if ( eof )
break;
curString.erase( 0, curString.find_first_not_of( U" \t" ) );
if ( keys.size() )
expandTildes( curString, keys.front() );
// If the string has any dsl markup, we strip it
string value = Utf8::encode( ArticleDom( curString ).root.renderAsText() );
for( list< wstring >::iterator i = keys.begin(); i != keys.end();
++i )
{
unescapeDsl( *i );
normalizeHeadword( *i );
abrv[ Utf8::encode( Folding::trimWhitespace( *i ) ) ] = value;
}
}
idxHeader.hasAbrv = 1;
idxHeader.abrvAddress = chunks.startNewBlock();
uint32_t sz = abrv.size();
chunks.addToBlock( &sz, sizeof( uint32_t ) );
for( map< string, string >::const_iterator i = abrv.begin();
i != abrv.end(); ++i )
{
// GD_DPRINTF( "%s:%s\n", i->first.c_str(), i->second.c_str() );
sz = i->first.size();
chunks.addToBlock( &sz, sizeof( uint32_t ) );
chunks.addToBlock( i->first.data(), sz );
sz = i->second.size();
chunks.addToBlock( &sz, sizeof( uint32_t ) );
chunks.addToBlock( i->second.data(), sz );
}
}
catch( std::exception & e )
{
gdWarning( "Error reading abrv file \"%s\", error: %s. Skipping it.\n",
abrvFileName.c_str(), e.what() );
}
}
bool hasString = false;
wstring curString;
size_t curOffset;
uint32_t articleCount = 0, wordCount = 0;
for( ; ; )
{
// Find the main headword
2021-10-23 04:05:16 +00:00
if ( !hasString && !scanner.readNextLineWithoutComments( curString, curOffset, true) )
break; // Clean end of file
hasString = false;
// The line read should either consist of pure whitespace, or be a headword
// skip too long headword,it can never be headword.
if( curString.empty() || curString.size() > 100 )
continue;
if ( isDslWs( curString[ 0 ] ) )
{
// The first character is blank. Let's make sure that all other
// characters are blank, too.
for( size_t x = 1; x < curString.size(); ++x )
{
if ( !isDslWs( curString[ x ] ) )
{
2017-06-22 15:02:04 +00:00
gdWarning( "Garbage string in %s at offset 0x%lX\n", i->c_str(), (unsigned long) curOffset );
break;
}
}
continue;
}
// Ok, got the headword
list< wstring > allEntryWords;
processUnsortedParts( curString, true );
expandOptionalParts( curString, &allEntryWords );
uint32_t articleOffset = curOffset;
//GD_DPRINTF( "Headword: %ls\n", curString.c_str() );
// More headwords may follow
for( ; ; )
{
if ( ! ( hasString = scanner.readNextLineWithoutComments( curString, curOffset ) ) )
{
2017-06-22 15:02:04 +00:00
gdWarning( "Premature end of file %s\n", i->c_str() );
break;
}
// Lingvo skips empty strings between the headwords
if ( curString.empty() )
continue;
if ( isDslWs( curString[ 0 ] ) )
break; // No more headwords
#ifdef QT_DEBUG
qDebug() << "Alt headword" << gd::toQString( curString );
#endif
processUnsortedParts( curString, true );
expandTildes( curString, allEntryWords.front() );
expandOptionalParts( curString, &allEntryWords );
}
if ( !hasString )
break;
// Insert new entry
uint32_t descOffset = chunks.startNewBlock();
chunks.addToBlock( &articleOffset, sizeof( articleOffset ) );
for( list< wstring >::iterator j = allEntryWords.begin();
j != allEntryWords.end(); ++j )
{
unescapeDsl( *j );
normalizeHeadword( *j );
indexedWords.addWord( *j, descOffset, maxHeadwordSize );
}
++articleCount;
wordCount += allEntryWords.size();
int insideInsided = 0;
wstring headword;
QVector< InsidedCard > insidedCards;
uint32_t offset = curOffset;
QVector< wstring > insidedHeadwords;
unsigned linesInsideCard = 0;
2017-07-03 15:12:22 +00:00
int dogLine = 0;
bool wasEmptyLine = false;
int headwordLine = scanner.getLinesRead() - 2;
bool noSignificantLines = Folding::applyWhitespaceOnly( curString ).empty();
bool haveLine = !noSignificantLines;
// Skip the article's body
for( ; ; )
{
2021-10-23 04:05:16 +00:00
hasString = haveLine ? true : scanner.readNextLineWithoutComments( curString, curOffset);
haveLine = false;
if ( !hasString || ( curString.size() && !isDslWs( curString[ 0 ] ) ) )
{
if( insideInsided )
2017-07-03 15:12:22 +00:00
{
gdWarning( "Unclosed tag '@' at line %i", dogLine );
insidedCards.append( InsidedCard( offset, curOffset - offset, insidedHeadwords ) );
2017-07-03 15:12:22 +00:00
}
if( noSignificantLines )
gdWarning( "Orphan headword at line %i", headwordLine );
break;
}
// Check for orphan strings
if( curString.empty() )
{
wasEmptyLine = true;
continue;
}
else
{
if( wasEmptyLine && !Folding::applyWhitespaceOnly( curString ).empty() )
gdWarning( "Orphan string at line %i", scanner.getLinesRead() - 1 );
}
if( noSignificantLines )
noSignificantLines = Folding::applyWhitespaceOnly( curString ).empty();
// Find embedded cards
wstring::size_type n = curString.find( L'@' );
if( n == wstring::npos || curString[ n - 1 ] == L'\\' )
{
if( insideInsided )
linesInsideCard++;
continue;
}
2017-07-03 15:12:22 +00:00
else
{
// Embedded card tag must be placed at first position in line after spaces
if( !isAtSignFirst( curString ) )
2017-07-03 15:12:22 +00:00
{
gdWarning( "Unescaped '@' symbol at line %i", scanner.getLinesRead() - 1 );
2017-07-03 18:32:13 +00:00
if( insideInsided )
linesInsideCard++;
2017-07-03 15:12:22 +00:00
continue;
}
}
dogLine = scanner.getLinesRead() - 1;
// Handle embedded card
if( insideInsided )
{
if( linesInsideCard )
{
insidedCards.append( InsidedCard( offset, curOffset - offset, insidedHeadwords ) );
insidedHeadwords.clear();
linesInsideCard = 0;
offset = curOffset;
}
}
else
{
offset = curOffset;
linesInsideCard = 0;
}
headword = Folding::trimWhitespace( curString.substr( n + 1 ) );
if( !headword.empty() )
{
processUnsortedParts( headword, true );
expandTildes( headword, allEntryWords.front() );
insidedHeadwords.append( headword );
insideInsided = true;
}
else
insideInsided = false;
}
// Now that we're having read the first string after the article
// itself, we can use its offset to calculate the article's size.
// An end of file works here, too.
uint32_t articleSize = ( curOffset - articleOffset );
chunks.addToBlock( &articleSize, sizeof( articleSize ) );
for( QVector< InsidedCard >::iterator i = insidedCards.begin(); i != insidedCards.end(); ++i )
{
uint32_t descOffset = chunks.startNewBlock();
chunks.addToBlock( &(*i).offset, sizeof( (*i).offset ) );
chunks.addToBlock( &(*i).size, sizeof( (*i).size ) );
for( int x = 0; x < (*i).headwords.size(); x++ )
2013-06-06 16:37:33 +00:00
{
allEntryWords.clear();
expandOptionalParts( (*i).headwords[ x ], &allEntryWords );
for( list< wstring >::iterator j = allEntryWords.begin();
j != allEntryWords.end(); ++j )
{
unescapeDsl( *j );
normalizeHeadword( *j );
indexedWords.addWord( *j, descOffset, maxHeadwordSize );
}
wordCount += allEntryWords.size();
}
++articleCount;
}
if ( !hasString )
break;
}
// Finish with the chunks
idxHeader.chunksOffset = chunks.finish();
// Build index
IndexInfo idxInfo = BtreeIndexing::buildIndex( indexedWords, idx );
idxHeader.indexBtreeMaxElements = idxInfo.btreeMaxElements;
idxHeader.indexRootOffset = idxInfo.rootOffset;
indexedWords.clear(); // Release memory -- no need for this data
// If there was a zip file, index it too
if ( zipFileName.size() )
{
GD_DPRINTF( "Indexing zip file\n" );
idxHeader.hasZipFile = 1;
IndexedWords zipFileNames;
IndexedZip zipFile;
if ( zipFile.openZipFile( QDir::fromNativeSeparators( zipFileName.c_str() ) ) )
zipFile.indexFile( zipFileNames );
if( !zipFileNames.empty() )
{
// Build the resulting zip file index
IndexInfo idxInfo = BtreeIndexing::buildIndex( zipFileNames, idx );
idxHeader.zipIndexBtreeMaxElements = idxInfo.btreeMaxElements;
idxHeader.zipIndexRootOffset = idxInfo.rootOffset;
}
else
{
// Bad zip file -- no index (though the mark that we have one
// remains)
idxHeader.zipIndexBtreeMaxElements = 0;
idxHeader.zipIndexRootOffset = 0;
}
}
else
idxHeader.hasZipFile = 0;
// That concludes it. Update the header.
idxHeader.signature = Signature;
idxHeader.formatVersion = CurrentFormatVersion;
idxHeader.zipSupportVersion = CurrentZipSupportVersion;
idxHeader.articleCount = articleCount;
idxHeader.wordCount = wordCount;
idxHeader.langFrom = dslLanguageToId( scanner.getLangFrom() );
idxHeader.langTo = dslLanguageToId( scanner.getLangTo() );
2009-04-22 21:37:32 +00:00
idx.rewind();
idx.write( &idxHeader, sizeof( idxHeader ) );
} // In-place try for saving line count
catch( ... )
{
atLine = scanner.getLinesRead();
throw;
}
} // if need to rebuild
dictionaries.push_back( std::make_shared<DslDictionary>( dictId,
indexFile,
2012-12-07 11:59:29 +00:00
dictFiles,
maxPictureWidth ) );
}
catch( std::exception & e )
{
gdWarning( "DSL dictionary reading failed: %s:%u, error: %s\n",
i->c_str(), atLine, e.what() );
}
}
return dictionaries;
}
}