#include "Encoding.h"
#include "InsetIterator.h"
#include "Paragraph.h"
+#include "TocBackend.h"
#include "insets/Inset.h"
#include "insets/InsetBibitem.h"
#include "insets/InsetBibtex.h"
#include "insets/InsetInclude.h"
+#include "support/convert.h"
#include "support/docstream.h"
#include "support/gettext.h"
#include "support/lassert.h"
#include "boost/regex.hpp"
+#include <set>
+
using namespace std;
using namespace lyx::support;
namespace lyx {
-//////////////////////////////////////////////////////////////////////
-//
-// BibTeXInfo
-//
-//////////////////////////////////////////////////////////////////////
-
-BibTeXInfo::BibTeXInfo(docstring const & key, docstring const & type)
- : is_bibtex_(true), bib_key_(key), entry_type_(type), info_()
-{}
-
-
-bool BibTeXInfo::hasField(docstring const & field) const
-{
- return count(field) == 1;
-}
-
-
-docstring const & BibTeXInfo::operator[](docstring const & field) const
-{
- BibTeXInfo::const_iterator it = find(field);
- if (it != end())
- return it->second;
- static docstring const empty_value = docstring();
- return empty_value;
-}
-
-
-docstring const & BibTeXInfo::operator[](string const & field) const
-{
- return operator[](from_ascii(field));
-}
-
-
-docstring BibTeXInfo::getValueForKey(string const & key,
- BibTeXInfo const * const xref) const
-{
- docstring const ret = operator[](key);
- if (!ret.empty() || !xref)
- return ret;
- return (*xref)[key];
-}
-
+namespace {
+// gets the "family name" from an author-type string
docstring familyName(docstring const & name)
{
if (name.empty())
return retval;
}
+// converts a string containing LaTeX commands into unicode
+// for display.
+docstring convertLaTeXCommands(docstring const & str)
+{
+ docstring val = str;
+ docstring ret;
+
+ bool scanning_cmd = false;
+ bool scanning_math = false;
+ bool escaped = false; // used to catch \$, etc.
+ while (val.size()) {
+ char_type const ch = val[0];
+
+ // if we're scanning math, we output everything until we
+ // find an unescaped $, at which point we break out.
+ if (scanning_math) {
+ if (escaped)
+ escaped = false;
+ else if (ch == '\\')
+ escaped = true;
+ else if (ch == '$')
+ scanning_math = false;
+ ret += ch;
+ val = val.substr(1);
+ continue;
+ }
+
+ // if we're scanning a command name, then we just
+ // discard characters until we hit something that
+ // isn't alpha.
+ if (scanning_cmd) {
+ if (isAlphaASCII(ch)) {
+ val = val.substr(1);
+ escaped = false;
+ continue;
+ }
+ // so we're done with this command.
+ // now we fall through and check this character.
+ scanning_cmd = false;
+ }
+
+ // was the last character a \? If so, then this is something like:
+ // \\ or \$, so we'll just output it. That's probably not always right...
+ if (escaped) {
+ // exception: output \, as THIN SPACE
+ if (ch == ',')
+ ret.push_back(0x2009);
+ else
+ ret += ch;
+ val = val.substr(1);
+ escaped = false;
+ continue;
+ }
+
+ if (ch == '$') {
+ ret += ch;
+ val = val.substr(1);
+ scanning_math = true;
+ continue;
+ }
+
+ // we just ignore braces
+ if (ch == '{' || ch == '}') {
+ val = val.substr(1);
+ continue;
+ }
+
+ // we're going to check things that look like commands, so if
+ // this doesn't, just output it.
+ if (ch != '\\') {
+ ret += ch;
+ val = val.substr(1);
+ continue;
+ }
+
+ // ok, could be a command of some sort
+ // let's see if it corresponds to some unicode
+ // unicodesymbols has things in the form: \"{u},
+ // whereas we may see things like: \"u. So we'll
+ // look for that and change it, if necessary.
+ static boost::regex const reg("^\\\\\\W\\w");
+ if (boost::regex_search(to_utf8(val), reg)) {
+ val.insert(3, from_ascii("}"));
+ val.insert(2, from_ascii("{"));
+ }
+ docstring rem;
+ docstring const cnvtd = Encodings::fromLaTeXCommand(val, rem,
+ Encodings::TEXT_CMD);
+ if (!cnvtd.empty()) {
+ // it did, so we'll take that bit and proceed with what's left
+ ret += cnvtd;
+ val = rem;
+ continue;
+ }
+ // it's a command of some sort
+ scanning_cmd = true;
+ escaped = true;
+ val = val.substr(1);
+ }
+ return ret;
+}
+
+} // anon namespace
+
+
+//////////////////////////////////////////////////////////////////////
+//
+// BibTeXInfo
+//
+//////////////////////////////////////////////////////////////////////
+
+BibTeXInfo::BibTeXInfo(docstring const & key, docstring const & type)
+ : is_bibtex_(true), bib_key_(key), entry_type_(type), info_(),
+ modifier_(0)
+{}
+
+
docstring const BibTeXInfo::getAbbreviatedAuthor() const
{
if (!is_bibtex_) {
- docstring const opt = trim(operator[]("label"));
+ docstring const opt = label();
if (opt.empty())
return docstring();
return authors;
}
- docstring author = operator[]("author");
+ docstring author = convertLaTeXCommands(operator[]("author"));
if (author.empty()) {
- author = operator[]("editor");
+ author = convertLaTeXCommands(operator[]("editor"));
if (author.empty())
return bib_key_;
}
if (is_bibtex_)
return operator[]("year");
- docstring const opt = trim(operator[]("label"));
+ docstring const opt = label();
if (opt.empty())
return docstring();
}
-namespace {
-
- docstring convertLaTeXCommands(docstring const & str)
- {
- docstring val = str;
- docstring ret;
-
- bool scanning_cmd = false;
- bool scanning_math = false;
- bool escaped = false; // used to catch \$, etc.
- while (val.size()) {
- char_type const ch = val[0];
-
- // if we're scanning math, we output everything until we
- // find an unescaped $, at which point we break out.
- if (scanning_math) {
- if (escaped)
- escaped = false;
- else if (ch == '\\')
- escaped = true;
- else if (ch == '$')
- scanning_math = false;
- ret += ch;
- val = val.substr(1);
- continue;
- }
-
- // if we're scanning a command name, then we just
- // discard characters until we hit something that
- // isn't alpha.
- if (scanning_cmd) {
- if (isAlphaASCII(ch)) {
- val = val.substr(1);
- escaped = false;
- continue;
- }
- // so we're done with this command.
- // now we fall through and check this character.
- scanning_cmd = false;
- }
-
- // was the last character a \? If so, then this is something like: \\,
- // or \$, so we'll just output it. That's probably not always right...
- if (escaped) {
- ret += ch;
- val = val.substr(1);
- escaped = false;
- continue;
- }
-
- if (ch == '$') {
- ret += ch;
- val = val.substr(1);
- scanning_math = true;
- continue;
- }
-
- // we just ignore braces
- if (ch == '{' || ch == '}') {
- val = val.substr(1);
- continue;
- }
-
- // we're going to check things that look like commands, so if
- // this doesn't, just output it.
- if (ch != '\\') {
- ret += ch;
- val = val.substr(1);
- continue;
- }
-
- // ok, could be a command of some sort
- // let's see if it corresponds to some unicode
- // unicodesymbols has things in the form: \"{u},
- // whereas we may see things like: \"u. So we'll
- // look for that and change it, if necessary.
- static boost::regex const reg("^\\\\\\W\\w");
- if (boost::regex_search(to_utf8(val), reg)) {
- val.insert(3, from_ascii("}"));
- val.insert(2, from_ascii("{"));
- }
- docstring rem;
- docstring const cnvtd = Encodings::fromLaTeXCommand(val, rem);
- if (!cnvtd.empty()) {
- // it did, so we'll take that bit and proceed with what's left
- ret += cnvtd;
- val = rem;
- continue;
- }
- // it's a command of some sort
- scanning_cmd = true;
- escaped = true;
- val = val.substr(1);
- }
- return ret;
- }
-
-} // anon namespace
-
-
docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref) const
{
if (!info_.empty())
}
+docstring const & BibTeXInfo::operator[](docstring const & field) const
+{
+ BibTeXInfo::const_iterator it = find(field);
+ if (it != end())
+ return it->second;
+ static docstring const empty_value = docstring();
+ return empty_value;
+}
+
+
+docstring const & BibTeXInfo::operator[](string const & field) const
+{
+ return operator[](from_ascii(field));
+}
+
+
+docstring BibTeXInfo::getValueForKey(string const & key,
+ BibTeXInfo const * const xref) const
+{
+ docstring const ret = operator[](key);
+ if (!ret.empty() || !xref)
+ return ret;
+ return (*xref)[key];
+}
+
+
//////////////////////////////////////////////////////////////////////
//
// BiblioInfo
}
-docstring const BiblioInfo::getYear(docstring const & key) const
+docstring const BiblioInfo::getCiteNumber(docstring const & key) const
+{
+ BiblioInfo::const_iterator it = find(key);
+ if (it == end())
+ return docstring();
+ BibTeXInfo const & data = it->second;
+ return data.citeNumber();
+}
+
+
+docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier) const
{
BiblioInfo::const_iterator it = find(key);
if (it == end())
return docstring();
BibTeXInfo const & data = it->second;
docstring year = data.getYear();
- if (!year.empty())
- return year;
- // let's try the crossref
- docstring const xref = data.getXRef();
- if (xref.empty())
- return _("No year"); // no luck
- BiblioInfo::const_iterator const xrefit = find(xref);
- if (xrefit == end())
- return _("No year"); // no luck again
- BibTeXInfo const & xref_data = xrefit->second;
- return xref_data.getYear();
- return data.getYear();
+ if (year.empty()) {
+ // let's try the crossref
+ docstring const xref = data.getXRef();
+ if (xref.empty())
+ return _("No year"); // no luck
+ BiblioInfo::const_iterator const xrefit = find(xref);
+ if (xrefit == end())
+ return _("No year"); // no luck again
+ BibTeXInfo const & xref_data = xrefit->second;
+ year = xref_data.getYear();
+ }
+ if (use_modifier && data.modifier() != 0)
+ year += data.modifier();
+ return year;
}
}
+namespace {
+ // used in xhtml to sort a list of BibTeXInfo objects
+ bool lSorter(BibTeXInfo const * lhs, BibTeXInfo const * rhs)
+ {
+ docstring const lauth = lhs->getAbbreviatedAuthor();
+ docstring const rauth = rhs->getAbbreviatedAuthor();
+ docstring const lyear = lhs->getYear();
+ docstring const ryear = rhs->getYear();
+ docstring const ltitl = lhs->operator[]("title");
+ docstring const rtitl = rhs->operator[]("title");
+ return (lauth < rauth)
+ || (lauth == rauth && lyear < ryear)
+ || (lauth == rauth && lyear == ryear && ltitl < rtitl);
+ }
+}
+
+
+void BiblioInfo::collectCitedEntries(Buffer const & buf)
+{
+ cited_entries_.clear();
+ // We are going to collect all the citation keys used in the document,
+ // getting them from the TOC.
+ // FIXME We may want to collect these differently, in the first case,
+ // so that we might have them in order of appearance.
+ set<docstring> citekeys;
+ Toc const & toc = buf.tocBackend().toc("citation");
+ Toc::const_iterator it = toc.begin();
+ Toc::const_iterator const en = toc.end();
+ for (; it != en; ++it) {
+ if (it->str().empty())
+ continue;
+ vector<docstring> const keys = getVectorFromString(it->str());
+ citekeys.insert(keys.begin(), keys.end());
+ }
+ if (citekeys.empty())
+ return;
+
+ // We have a set of the keys used in this document.
+ // We will now convert it to a list of the BibTeXInfo objects used in
+ // this document...
+ vector<BibTeXInfo const *> bi;
+ set<docstring>::const_iterator cit = citekeys.begin();
+ set<docstring>::const_iterator const cen = citekeys.end();
+ for (; cit != cen; ++cit) {
+ BiblioInfo::const_iterator const bt = find(*cit);
+ if (bt == end() || !bt->second.isBibTeX())
+ continue;
+ bi.push_back(&(bt->second));
+ }
+ // ...and sort it.
+ sort(bi.begin(), bi.end(), lSorter);
+
+ // Now we can write the sorted keys
+ vector<BibTeXInfo const *>::const_iterator bit = bi.begin();
+ vector<BibTeXInfo const *>::const_iterator ben = bi.end();
+ for (; bit != ben; ++bit)
+ cited_entries_.push_back((*bit)->key());
+}
+
+
+void BiblioInfo::makeCitationLabels(Buffer const & buf)
+{
+ collectCitedEntries(buf);
+ CiteEngine const engine = buf.params().citeEngine();
+ bool const numbers =
+ (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL);
+
+ int keynumber = 0;
+ char modifier = 0;
+ // used to remember the last one we saw
+ // we'll be comparing entries to see if we need to add
+ // modifiers, like "1984a"
+ map<docstring, BibTeXInfo>::iterator last;
+
+ vector<docstring>::const_iterator it = cited_entries_.begin();
+ vector<docstring>::const_iterator const en = cited_entries_.end();
+ for (; it != en; ++it) {
+ map<docstring, BibTeXInfo>::iterator const biit = bimap_.find(*it);
+ // this shouldn't happen, but...
+ if (biit == bimap_.end())
+ // ...fail gracefully, anyway.
+ continue;
+ BibTeXInfo & entry = biit->second;
+ if (numbers) {
+ docstring const num = convert<docstring>(++keynumber);
+ entry.setCiteNumber(num);
+ } else {
+ if (it != cited_entries_.begin()
+ && entry.getAbbreviatedAuthor() == last->second.getAbbreviatedAuthor()
+ // we access the year via getYear() so as to get it from the xref,
+ // if we need to do so
+ && getYear(entry.key()) == getYear(last->second.key())) {
+ if (modifier == 0) {
+ // so the last one should have been 'a'
+ last->second.setModifier('a');
+ modifier = 'b';
+ } else if (modifier == 'z')
+ modifier = 'A';
+ else
+ modifier++;
+ } else {
+ modifier = 0;
+ }
+ entry.setModifier(modifier);
+ // remember the last one
+ last = biit;
+ }
+ }
+}
+
+
//////////////////////////////////////////////////////////////////////
//
// CitationStyle
string cite = citeCommands[s.style];
if (s.full) {
CiteStyle const * last = citeStylesFull + nCiteStylesFull;
- if (find(citeStylesFull, last, s.style) != last)
+ if (std::find(citeStylesFull, last, s.style) != last)
cite += '*';
}
if (s.forceUpperCase) {
CiteStyle const * last = citeStylesUCase + nCiteStylesUCase;
- if (find(citeStylesUCase, last, s.style) != last)
+ if (std::find(citeStylesUCase, last, s.style) != last)
cite[0] = 'C';
}