You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
tellico/src/translators/bibtexhandler.cpp

320 lines
11 KiB

/***************************************************************************
copyright : (C) 2003-2006 by Robby Stephenson
email : robby@periapsis.org
***************************************************************************/
/***************************************************************************
* *
* This program is free software; you can redistribute it and/or modify *
* it under the terms of version 2 of the GNU General Public License as *
* published by the Free Software Foundation; *
* *
***************************************************************************/
#include "bibtexhandler.h"
#include "../collections/bibtexcollection.h"
#include "../entry.h"
#include "../field.h"
#include "../collection.h"
#include "../document.h"
#include "../filehandler.h"
#include "../latin1literal.h"
#include "../tellico_debug.h"
#include <kstandarddirs.h>
#include <kurl.h>
#include <kstringhandler.h>
#include <klocale.h>
#include <qstring.h>
#include <qstringlist.h>
#include <qregexp.h>
#include <qdom.h>
// don't add braces around capital letters by default
#define TELLICO_BIBTEX_BRACES 0
using Tellico::BibtexHandler;
BibtexHandler::StringListMap* BibtexHandler::s_utf8LatexMap = 0;
BibtexHandler::QuoteStyle BibtexHandler::s_quoteStyle = BibtexHandler::BRACES;
const QRegExp BibtexHandler::s_badKeyChars(QString::fromLatin1("[^0-9a-zA-Z-]"));
QStringList BibtexHandler::bibtexKeys(const Data::EntryVec& entries_) {
QStringList keys;
for(Data::EntryVec::ConstIterator it = entries_.begin(); it != entries_.end(); ++it) {
QString s = bibtexKey(it.data());
if(!s.isEmpty()) {
keys << s;
}
}
return keys;
}
QString BibtexHandler::bibtexKey(Data::ConstEntryPtr entry_) {
if(!entry_ || !entry_->collection() || entry_->collection()->type() != Data::Collection::Bibtex) {
return QString::null;
}
const Data::BibtexCollection* c = static_cast<const Data::BibtexCollection*>(entry_->collection().data());
Data::FieldPtr f = c->fieldByBibtexName(QString::fromLatin1("key"));
if(f) {
QString key = entry_->field(f->name());
if(!key.isEmpty()) {
return key;
}
}
QString author;
Data::FieldPtr authorField = c->fieldByBibtexName(QString::fromLatin1("author"));
if(authorField) {
if(authorField->flags() & Data::Field::AllowMultiple) {
// grab first author only;
QString tmp = entry_->field(authorField->name());
author = tmp.section(';', 0, 0);
} else {
author = entry_->field(authorField->name());
}
}
Data::FieldPtr titleField = c->fieldByBibtexName(QString::fromLatin1("title"));
QString title;
if(titleField) {
title = entry_->field(titleField->name());
}
Data::FieldPtr yearField = c->fieldByBibtexName(QString::fromLatin1("year"));
QString year;
if(yearField) {
year = entry_->field(yearField->name());
}
if(year.isEmpty()) {
year = entry_->field(QString::fromLatin1("pub_year"));
if(year.isEmpty()) {
year = entry_->field(QString::fromLatin1("cr_year"));
}
}
year = year.section(';', 0, 0);
return bibtexKey(author, title, year);
}
QString BibtexHandler::bibtexKey(const QString& author_, const QString& title_, const QString& year_) {
QString key;
// if no comma, take the last word
if(!author_.isEmpty()) {
if(author_.find(',') == -1) {
key += author_.section(' ', -1).lower() + '-';
} else {
// if there is a comma, take the string up to the first comma
key += author_.section(',', 0, 0).lower() + '-';
}
}
QStringList words = QStringList::split(' ', title_);
for(QStringList::ConstIterator it = words.begin(); it != words.end(); ++it) {
key += (*it).left(1).lower();
}
key += year_;
// bibtex key may only contain [0-9a-zA-Z-]
return key.replace(s_badKeyChars, QString::null);
}
void BibtexHandler::loadTranslationMaps() {
QString mapfile = locate("appdata", QString::fromLatin1("bibtex-translation.xml"));
if(mapfile.isEmpty()) {
return;
}
s_utf8LatexMap = new StringListMap();
KURL u;
u.setPath(mapfile);
// no namespace processing
QDomDocument dom = FileHandler::readXMLFile(u, false);
QDomNodeList keyList = dom.elementsByTagName(QString::fromLatin1("key"));
for(unsigned i = 0; i < keyList.count(); ++i) {
QDomNodeList strList = keyList.item(i).toElement().elementsByTagName(QString::fromLatin1("string"));
// the strList might have more than one node since there are multiple ways
// to represent a character in LaTex.
QString s = keyList.item(i).toElement().attribute(QString::fromLatin1("char"));
for(unsigned j = 0; j < strList.count(); ++j) {
(*s_utf8LatexMap)[s].append(strList.item(j).toElement().text());
// kdDebug() << "BibtexHandler::loadTranslationMaps - "
// << s << " = " << strList.item(j).toElement().text() << endl;
}
}
}
QString BibtexHandler::importText(char* text_) {
if(!s_utf8LatexMap) {
loadTranslationMaps();
}
QString str = QString::fromUtf8(text_);
for(StringListMap::Iterator it = s_utf8LatexMap->begin(); it != s_utf8LatexMap->end(); ++it) {
for(QStringList::Iterator sit = it.data().begin(); sit != it.data().end(); ++sit) {
str.replace(*sit, it.key());
}
}
// now replace capitalized letters, such as {X}
// but since we don't want to turn "... X" into "... {X}" later when exporting
// we need to lower-case any capitalized text after the first letter that is
// NOT contained in braces
QRegExp rx(QString::fromLatin1("\\{([A-Z]+)\\}"));
rx.setMinimal(true);
str.replace(rx, QString::fromLatin1("\\1"));
return str;
}
QString BibtexHandler::exportText(const QString& text_, const QStringList& macros_) {
if(!s_utf8LatexMap) {
loadTranslationMaps();
}
QChar lquote, rquote;
switch(s_quoteStyle) {
case BRACES:
lquote = '{';
rquote = '}';
break;
case QUOTES:
lquote = '"';
rquote = '"';
break;
}
QString text = text_;
for(StringListMap::Iterator it = s_utf8LatexMap->begin(); it != s_utf8LatexMap->end(); ++it) {
text.replace(it.key(), it.data()[0]);
}
if(macros_.isEmpty()) {
return lquote + addBraces(text) + rquote;
}
// Now, split the text by the character '#', and examine each token to see if it is in
// the macro list. If it is not, then add left-quote and right-quote around it. If it is, don't
// change it. Then, in case '#' occurs in a non-macro string, replace any occurrences of '}#{' with '#'
// list of new tokens
QStringList list;
// first, split the text
QStringList tokens = QStringList::split('#', text, true);
for(QStringList::Iterator it = tokens.begin(); it != tokens.end(); ++it) {
// check to see if token is a macro
if(macros_.findIndex((*it).stripWhiteSpace()) == -1) {
// the token is NOT a macro, add braces around whole words and also around capitals
list << lquote + addBraces(*it) + rquote;
} else {
list << *it;
}
}
const QChar octo = '#';
text = list.join(octo);
text.replace(QString(rquote)+octo+lquote, octo);
return text;
}
bool BibtexHandler::setFieldValue(Data::EntryPtr entry_, const QString& bibtexField_, const QString& value_) {
Data::BibtexCollection* c = static_cast<Data::BibtexCollection*>(entry_->collection().data());
Data::FieldPtr field = c->fieldByBibtexName(bibtexField_);
if(!field) {
// it was the case that the default bibliography did not have a bibtex property for keywords
// so a "keywords" field would get created in the imported collection
// but the existing collection had a field "keyword" so the values would not get imported
// here, check to see if the current collection has a field with the same bibtex name and
// use it instead of creating a new one
Data::BibtexCollection* existingColl = Data::Document::self()->collection()->type() == Data::Collection::Bibtex
? static_cast<Data::BibtexCollection*>(Data::Document::self()->collection().data())
: 0;
Data::FieldPtr existingField = existingColl ? existingColl->fieldByBibtexName(bibtexField_) : 0;
if(existingField) {
field = new Data::Field(*existingField);
} else if(value_.length() < 100) {
// arbitrarily say if the value has more than 100 chars, then it's a paragraph
QString vlower = value_.lower();
// special case, try to detect URLs
// In qt 3.1, QString::startsWith() is always case-sensitive
if(bibtexField_ == Latin1Literal("url")
|| vlower.startsWith(QString::fromLatin1("http")) // may also be https
|| vlower.startsWith(QString::fromLatin1("ftp:/"))
|| vlower.startsWith(QString::fromLatin1("file:/"))
|| vlower.startsWith(QString::fromLatin1("/"))) { // assume this indicates a local path
myDebug() << "BibtexHandler::setFieldValue() - creating a URL field for " << bibtexField_ << endl;
field = new Data::Field(bibtexField_, KStringHandler::capwords(bibtexField_), Data::Field::URL);
} else {
field = new Data::Field(bibtexField_, KStringHandler::capwords(bibtexField_), Data::Field::Line);
}
field->setCategory(i18n("Unknown"));
} else {
field = new Data::Field(bibtexField_, KStringHandler::capwords(bibtexField_), Data::Field::Para);
}
field->setProperty(QString::fromLatin1("bibtex"), bibtexField_);
c->addField(field);
}
// special case keywords, replace commas with semi-colons so they get separated
QString value = value_;
if(field->property(QString::fromLatin1("bibtex")).startsWith(QString::fromLatin1("keyword"))) {
value.replace(',', ';');
// special case refbase bibtex export, with multiple keywords fields
QString oValue = entry_->field(field);
if(!oValue.isEmpty()) {
value = oValue + "; " + value;
}
}
return entry_->setField(field, value);
}
QString& BibtexHandler::cleanText(QString& text_) {
// FIXME: need to improve this for removing all Latex entities
// QRegExp rx(QString::fromLatin1("(?=[^\\\\])\\\\.+\\{"));
QRegExp rx(QString::fromLatin1("\\\\.+\\{"));
rx.setMinimal(true);
text_.replace(rx, QString::null);
text_.replace(QRegExp(QString::fromLatin1("[{}]")), QString::null);
text_.replace('~', ' ');
return text_;
}
// add braces around capital letters
QString& BibtexHandler::addBraces(QString& text) {
#if !TELLICO_BIBTEX_BRACES
return text;
#else
int inside = 0;
uint l = text.length();
// start at first letter, but skip if only the first is capitalized
for(uint i = 0; i < l; ++i) {
const QChar c = text.at(i);
if(inside == 0 && c >= 'A' && c <= 'Z') {
uint j = i+1;
while(text.at(j) >= 'A' && text.at(j) <= 'Z' && j < l) {
++j;
}
if(i == 0 && j == 1) {
continue; // no need to do anything to first letter
}
text.insert(i, '{');
// now j should be incremented
text.insert(j+1, '}');
i = j+1;
l += 2; // the length changed
} else if(c == '{') {
++inside;
} else if(c == '}') {
--inside;
}
}
return text;
#endif
}