X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FBiblioInfo.cpp;h=420eaa2177c87f9d1d05f92d939565fd60313136;hb=f5123ccfa82215030c70561af881175d92244492;hp=c68e18ae4c2d00d871428b24427dc8543b9d9e90;hpb=f50c47d4be8d511e485ff6a63bd62e21e904efe9;p=lyx.git

diff --git a/src/BiblioInfo.cpp b/src/BiblioInfo.cpp
index c68e18ae4c..420eaa2177 100644
--- a/src/BiblioInfo.cpp
+++ b/src/BiblioInfo.cpp
@@ -18,20 +18,27 @@
 #include "buffer_funcs.h"
 #include "Encoding.h"
 #include "InsetIterator.h"
+#include "Language.h"
 #include "Paragraph.h"
+#include "TextClass.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/debug.h"
 #include "support/docstream.h"
 #include "support/gettext.h"
 #include "support/lassert.h"
 #include "support/lstrings.h"
 #include "support/textutils.h"
 
-#include "boost/regex.hpp"
+#include "support/regex.h"
+
+#include <set>
 
 using namespace std;
 using namespace lyx::support;
@@ -129,10 +136,14 @@ docstring convertLaTeXCommands(docstring const & str)
 			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...
+		// 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;
+			// exception: output \, as THIN SPACE
+			if (ch == ',')
+				ret.push_back(0x2009);
+			else
+				ret += ch;
 			val = val.substr(1);
 			escaped = false;
 			continue;
@@ -164,13 +175,14 @@ docstring convertLaTeXCommands(docstring const & str)
 		// 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)) {
+		static lyx::regex const reg("^\\\\\\W\\w");
+		if (lyx::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);
+		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;
@@ -195,16 +207,11 @@ docstring convertLaTeXCommands(docstring const & str)
 //////////////////////////////////////////////////////////////////////
 
 BibTeXInfo::BibTeXInfo(docstring const & key, docstring const & type)
-	: is_bibtex_(true), bib_key_(key), entry_type_(type), info_()
+	: is_bibtex_(true), bib_key_(key), entry_type_(type), info_(),
+	  modifier_(0)
 {}
 
 
-bool BibTeXInfo::hasField(docstring const & field) const
-{
-	return count(field) == 1;
-}
-
-
 docstring const BibTeXInfo::getAbbreviatedAuthor() const
 {
 	if (!is_bibtex_) {
@@ -213,17 +220,24 @@ docstring const BibTeXInfo::getAbbreviatedAuthor() const
 			return docstring();
 
 		docstring authors;
-		split(opt, authors, '(');
+		docstring const remainder = trim(split(opt, authors, '('));
+		if (remainder.empty())
+			// in this case, we didn't find a "(", 
+			// so we don't have author (year)
+			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_;
 	}
 
+	// FIXME Move this to a separate routine that can
+	// be called from elsewhere.
+	// 
 	// OK, we've got some names. Let's format them.
 	// Try to split the author list on " and "
 	vector<docstring> const authors =
@@ -250,9 +264,12 @@ docstring const BibTeXInfo::getYear() const
 		return docstring();
 
 	docstring authors;
-	docstring const tmp = split(opt, authors, '(');
+	docstring tmp = split(opt, authors, '(');
+	if (tmp.empty()) 
+		// we don't have author (year)
+		return docstring();
 	docstring year;
-	split(tmp, year, ')');
+	tmp = split(tmp, year, ')');
 	return year;
 }
 
@@ -265,7 +282,247 @@ docstring const BibTeXInfo::getXRef() const
 }
 
 
-docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref) const
+namespace {
+	string parseOptions(string const & format, string & optkey, 
+			string & ifpart, string & elsepart);
+
+	// Calls parseOptions to deal with an embedded option, such as:
+	//   {%number%[[, no.~%number%]]}
+	// which must appear at the start of format. ifelsepart gets the
+	// whole of the option, and we return what's left after the option.
+	// we return format if there is an error.
+	string parseEmbeddedOption(string const & format, string & ifelsepart)
+	{
+		LASSERT(format[0] == '{' && format[1] == '%', return format);
+		string optkey;
+		string ifpart;
+		string elsepart;
+		string const rest = parseOptions(format, optkey, ifpart, elsepart);
+		if (format == rest) { // parse error
+			LYXERR0("ERROR! Couldn't parse `" << format <<"'.");
+			return format;
+		}
+		LASSERT(rest.size() <= format.size(), /* */);
+		ifelsepart = format.substr(0, format.size() - rest.size());
+		return rest;
+	}
+	
+	
+	// Gets a "clause" from a format string, where the clause is 
+	// delimited by '[[' and ']]'. Returns what is left after the
+	// clause is removed, and returns format if there is an error.
+	string getClause(string const & format, string & clause)
+	{
+		string fmt = format;
+		// remove '[['
+		fmt = fmt.substr(2);
+		// we'll remove characters from the front of fmt as we 
+		// deal with them
+		while (fmt.size()) { 
+			if (fmt[0] == ']' && fmt.size() > 1 && fmt[1] == ']') {
+				// that's the end
+				fmt = fmt.substr(2);
+				break;
+			}
+			// check for an embedded option
+			if (fmt[0] == '{' && fmt.size() > 1 && fmt[1] == '%') {
+				string part;
+				string const rest = parseEmbeddedOption(fmt, part);
+				if (fmt == rest) {
+					LYXERR0("ERROR! Couldn't parse embedded option in `" << format <<"'.");
+					return format;
+				}
+				clause += part;
+				fmt = rest;
+			} else { // it's just a normal character
+				clause += fmt[0];
+				fmt = fmt.substr(1);
+			}
+		}
+		return fmt;
+	}
+
+
+	// parse an options string, which must appear at the start of the
+	// format parameter. puts the parsed bits in optkey, ifpart, and
+	// elsepart and returns what's left after the option is removed.
+	// if there's an error, it returns format itself.
+	string parseOptions(string const & format, string & optkey, 
+			string & ifpart, string & elsepart) 
+	{
+		LASSERT(format[0] == '{' && format[1] == '%', return format);
+		// strip '{%'
+		string fmt = format.substr(2);
+		size_t pos = fmt.find('%'); // end of key
+		if (pos == string::npos) {
+			LYXERR0("Error parsing  `" << format <<"'. Can't find end of key.");
+			return format;
+		}
+		optkey = fmt.substr(0,pos);
+		fmt = fmt.substr(pos + 1);
+		// [[format]] should be next
+		if (fmt[0] != '[' || fmt[1] != '[') {
+			LYXERR0("Error parsing  `" << format <<"'. Can't find '[[' after key.");
+			return format;
+		}
+
+		string curfmt = fmt;
+		fmt = getClause(curfmt, ifpart);
+		if (fmt == curfmt) {
+			LYXERR0("Error parsing  `" << format <<"'. Couldn't get if clause.");
+			return format;
+		}
+
+		if (fmt[0] == '}') // we're done, no else clause
+			return fmt.substr(1);
+	
+		// else part should follow
+		if (fmt[0] != '[' || fmt[1] != '[') {
+			LYXERR0("Error parsing  `" << format <<"'. Can't find else clause.");
+			return format;
+		}
+		
+		curfmt = fmt;
+		fmt = getClause(curfmt, elsepart);
+		// we should be done
+		if (fmt == curfmt || fmt[0] != '}') {
+			LYXERR0("Error parsing  `" << format <<"'. Can't find end of option.");
+			return format;
+		}
+		return fmt.substr(1);
+}
+
+} // anon namespace
+
+
+docstring BibTeXInfo::expandFormat(string const & format, 
+		BibTeXInfo const * const xref, int & counter, Buffer const & buf, 
+		bool richtext) const
+{
+	// incorrect use of macros could put us in an infinite loop
+	static int max_passes = 5000;
+	docstring ret; // return value
+	string key;
+	bool scanning_key = false;
+	bool scanning_rich = false;
+
+	string fmt = format;
+	// we'll remove characters from the front of fmt as we 
+	// deal with them
+	while (fmt.size()) {
+		if (counter++ > max_passes) {
+			LYXERR0("Recursion limit reached while parsing `" 
+			        << format << "'.");
+			return _("ERROR!");
+		}
+				
+		char_type thischar = fmt[0];
+		if (thischar == '%') { 
+			// beginning or end of key
+			if (scanning_key) { 
+				// end of key
+				scanning_key = false;
+				// so we replace the key with its value, which may be empty
+				if (key[0] == '!') {
+					// macro
+					string const val = 
+						buf.params().documentClass().getCiteMacro(key);
+					fmt = val + fmt.substr(1);
+					continue;
+				} else if (key[0] == '_') {
+					// a translatable bit
+					string const val = 
+						buf.params().documentClass().getCiteMacro(key);
+					docstring const trans = 
+						translateIfPossible(from_utf8(val), buf.params().language->code());
+					ret += trans;
+				} else {
+					docstring const val = getValueForKey(key, xref);
+					ret += val;
+				}
+			} else {
+				// beginning of key
+				key.clear();
+				scanning_key = true;
+			}
+		}
+		else if (thischar == '{') { 
+			// beginning of option?
+			if (scanning_key) {
+				LYXERR0("ERROR: Found `{' when scanning key in `" << format << "'.");
+				return _("ERROR!");
+			}
+			if (fmt.size() > 1) {
+				if (fmt[1] == '%') {
+					// it is the beginning of an optional format
+					string optkey;
+					string ifpart;
+					string elsepart;
+					string const newfmt = 
+						parseOptions(fmt, optkey, ifpart, elsepart);
+					if (newfmt == fmt) // parse error
+						return _("ERROR!");
+					fmt = newfmt;
+					docstring const val = getValueForKey(optkey, xref);
+					if (!val.empty())
+						ret += expandFormat(ifpart, xref, counter, buf, richtext);
+					else if (!elsepart.empty())
+						ret += expandFormat(elsepart, xref, counter, buf, richtext);
+					// fmt will have been shortened for us already
+					continue; 
+				}
+				if (fmt[1] == '!') {
+					// beginning of rich text
+					scanning_rich = true;
+					fmt = fmt.substr(2);
+					continue;
+				}
+			}
+			// we are here if '{' was not followed by % or !. 
+			// So it's just a character.
+			ret += thischar;
+		}
+		else if (scanning_rich && thischar == '!' 
+		         && fmt.size() > 1 && fmt[1] == '}') {
+			// end of rich text
+			scanning_rich = false;
+			fmt = fmt.substr(2);
+			continue;
+		}
+		else if (scanning_key)
+			key += char(thischar);
+		else if (richtext) {
+			if (scanning_rich)
+				ret += thischar;
+			else {
+				// we need to escape '<' and '>'
+				if (thischar == '<')
+					ret += "&lt;";
+				else if (thischar == '>')
+					ret += "&gt;";
+				else 
+					ret += thischar;
+			}	
+		}	else if (!scanning_rich /* && !richtext */)
+			ret += thischar;
+		// else the character is discarded, which will happen only if
+		// richtext == false and we are scanning rich text
+		fmt = fmt.substr(1);
+	} // for loop
+	if (scanning_key) {
+		LYXERR0("Never found end of key in `" << format << "'!");
+		return _("ERROR!");
+	}
+	if (scanning_rich) {
+		LYXERR0("Never found end of rich text in `" << format << "'!");
+		return _("ERROR!");
+	}
+	return ret;
+}
+
+
+docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref,
+	Buffer const & buf, bool richtext) const
 {
 	if (!info_.empty())
 		return info_;
@@ -275,59 +532,15 @@ docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref) const
 		info_ = it->second;
 		return info_;
 	}
- 
-	// FIXME
-	// This could be made a lot better using the entry_type_
-	// field to customize the output based upon entry type.
-	
-	// Search for all possible "required" fields
-	docstring author = getValueForKey("author", xref);
-	if (author.empty())
-		author = getValueForKey("editor", xref);
- 
-	docstring year   = getValueForKey("year", xref);
-	docstring title  = getValueForKey("title", xref);
-	docstring docLoc = getValueForKey("pages", xref);
-	if (docLoc.empty()) {
-		docLoc = getValueForKey("chapter", xref);
-		if (!docLoc.empty())
-			docLoc = _("Ch. ") + docLoc;
-	}	else {
-		docLoc = _("pp. ") + docLoc;
-	}
 
-	docstring media = getValueForKey("journal", xref);
-	if (media.empty()) {
-		media = getValueForKey("publisher", xref);
-		if (media.empty()) {
-			media = getValueForKey("school", xref);
-			if (media.empty())
-				media = getValueForKey("institution");
-		}
-	}
-	docstring volume = getValueForKey("volume", xref);
-
-	odocstringstream result;
-	if (!author.empty())
-		result << author << ", ";
-	if (!title.empty())
-		result << title;
-	if (!media.empty())
-		result << ", " << media;
-	if (!year.empty())
-		result << " (" << year << ")";
-	if (!docLoc.empty())
-		result << ", " << docLoc;
-
-	docstring const result_str = rtrim(result.str());
-	if (!result_str.empty()) {
-		info_ = convertLaTeXCommands(result_str);
-		return info_;
-	}
+	DocumentClass const & dc = buf.params().documentClass();
+	string const & format = dc.getCiteFormat(to_utf8(entry_type_));
+	int counter = 0;
+	info_ = expandFormat(format, xref, counter, buf, richtext);
 
-	// This should never happen (or at least be very unusual!)
-	static docstring e = docstring();
-	return e;
+	if (!info_.empty())
+		info_ = convertLaTeXCommands(info_);
+	return info_;
 }
 
 
@@ -420,29 +633,42 @@ docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key) const
 }
 
 
-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;
 }
 
 
-docstring const BiblioInfo::getInfo(docstring const & key) const
+docstring const BiblioInfo::getInfo(docstring const & key, 
+	Buffer const & buf, bool richtext) const
 {
 	BiblioInfo::const_iterator it = find(key);
 	if (it == end())
@@ -455,10 +681,20 @@ docstring const BiblioInfo::getInfo(docstring const & key) const
 		if (xrefit != end())
 			xrefptr = &(xrefit->second);
 	}
-	return data.getInfo(xrefptr);
+	return data.getInfo(xrefptr, buf, richtext);
 }
 
 
+bool BiblioInfo::isBibtex(docstring const & key) const
+{
+	BiblioInfo::const_iterator it = find(key);
+	if (it == end())
+		return false;
+	return it->second.isBibTeX();
+}
+
+
+
 vector<docstring> const BiblioInfo::getCiteStrings(
 	docstring const & key, Buffer const & buf) const
 {
@@ -597,6 +833,117 @@ void BiblioInfo::mergeBiblioInfo(BiblioInfo const & info)
 }
 
 
+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
@@ -669,13 +1016,13 @@ string citationStyleToString(const CitationStyle & s)
 	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';
 	}