3 * This file is part of LyX, the document processor.
4 * Licence details can be found in the file COPYING.
6 * \author Angus Leeming
9 * Full author contact details are available in file CREDITS.
17 #include "bufferparams.h"
20 #include "support/lstrings.h"
22 #include <boost/regex.hpp>
28 using std::ostringstream;
34 using support::ascii_lowercase;
35 using support::bformat;
36 using support::compare_ascii_no_case;
37 using support::contains;
38 using support::getVectorFromString;
40 using support::prefixIs;
51 vector<string> const init_possible_cite_commands()
53 char const * const pos[] = {
55 "citet", "citep", "citealt", "citealp",
56 "citeauthor", "citeyear", "citeyearpar",
57 "citet*", "citep*", "citealt*", "citealp*", "citeauthor*",
58 "Citet", "Citep", "Citealt", "Citealp", "Citeauthor",
59 "Citet*", "Citep*", "Citealt*", "Citealp*", "Citeauthor*",
61 "footcite", "footcitet", "footcitep", "footcitealt",
62 "footcitealp", "footciteauthor", "footciteyear",
68 size_t const size_pos = sizeof(pos) / sizeof(pos[0]);
70 return vector<string>(pos, pos + size_pos);
74 vector<string> const & possible_cite_commands()
76 static vector<string> const pos = init_possible_cite_commands();
81 bool is_possible_cite_command(string const & input)
83 vector<string> const & possibles = possible_cite_commands();
84 vector<string>::const_iterator const end = possibles.end();
85 return std::find(possibles.begin(), end, input) != end;
89 string const default_cite_command(CiteEngine engine)
96 case ENGINE_NATBIB_AUTHORYEAR:
99 case ENGINE_NATBIB_NUMERICAL:
112 string const asValidLatexCommand(string const & input,
113 CiteEngine_enum const & engine)
115 string const default_str = default_cite_command(engine);
116 if (!is_possible_cite_command(input))
122 output = default_str;
125 case ENGINE_NATBIB_AUTHORYEAR:
126 case ENGINE_NATBIB_NUMERICAL:
127 if (input == "cite" || input == "citefield" ||
128 input == "citetitle" || input == "cite*")
129 output = default_str;
130 else if (prefixIs(input, "foot"))
131 output = input.substr(4);
136 case ENGINE_JURABIB: {
137 // Jurabib does not support the 'uppercase' natbib style.
139 output = string(1, 'c') + input.substr(1);
143 // Jurabib does not support the 'full' natbib style.
144 string::size_type const n = output.size() - 1;
145 if (output != "cite*" && output[n] == '*')
146 output = output.substr(0, n);
156 string const familyName(string const & name)
158 // Very simple parser
161 // possible authorname combinations are:
162 // "Surname, FirstName"
164 // "FirstName Surname"
166 string::size_type idx = fname.find(',');
167 if (idx != string::npos)
168 return ltrim(fname.substr(0, idx));
169 idx = fname.rfind('.');
170 if (idx != string::npos)
171 fname = ltrim(fname.substr(idx + 1));
172 // test if we have a LaTeX Space in front
173 if (fname[0] == '\\')
174 return fname.substr(2);
180 string const getAbbreviatedAuthor(InfoMap const & map, string const & key)
182 BOOST_ASSERT(!map.empty());
184 InfoMap::const_iterator it = map.find(key);
187 string const & data = it->second;
189 // Is the entry a BibTeX one or one from lyx-layout "bibliography"?
190 string::size_type const pos = data.find("TheBibliographyRef");
191 if (pos != string::npos) {
196 string const opt = trim(data.substr(0, pos - 1));
201 split(opt, authors, '(');
205 string author = parseBibTeX(data, "author");
208 author = parseBibTeX(data, "editor");
210 if (author.empty()) {
211 author = parseBibTeX(data, "key");
217 vector<string> const authors = getVectorFromString(author, " and ");
221 if (authors.size() == 2)
222 return bformat(_("%1$s and %2$s"),
223 familyName(authors[0]), familyName(authors[1]));
225 if (authors.size() > 2)
226 return bformat(_("%1$s et al."), familyName(authors[0]));
228 return familyName(authors[0]);
232 string const getYear(InfoMap const & map, string const & key)
234 BOOST_ASSERT(!map.empty());
236 InfoMap::const_iterator it = map.find(key);
239 string const & data = it->second;
241 // Is the entry a BibTeX one or one from lyx-layout "bibliography"?
242 string::size_type const pos = data.find("TheBibliographyRef");
243 if (pos != string::npos) {
249 trim(data.substr(0, pos - 1));
254 string const tmp = split(opt, authors, '(');
256 split(tmp, year, ')');
261 string year = parseBibTeX(data, "year");
271 // A functor for use with std::sort, leading to case insensitive sorting
272 struct compareNoCase: public std::binary_function<string, string, bool>
274 bool operator()(string const & s1, string const & s2) const {
275 return compare_ascii_no_case(s1, s2) < 0;
282 vector<string> const getKeys(InfoMap const & map)
284 vector<string> bibkeys;
285 InfoMap::const_iterator it = map.begin();
286 InfoMap::const_iterator end = map.end();
287 for (; it != end; ++it) {
288 bibkeys.push_back(it->first);
291 std::sort(bibkeys.begin(), bibkeys.end(), compareNoCase());
296 string const getInfo(InfoMap const & map, string const & key)
298 BOOST_ASSERT(!map.empty());
300 InfoMap::const_iterator it = map.find(key);
303 string const & data = it->second;
305 // is the entry a BibTeX one or one from lyx-layout "bibliography"?
306 string const separator("TheBibliographyRef");
307 string::size_type const pos = data.find(separator);
308 if (pos != string::npos) {
309 string::size_type const pos2 = pos + separator.size();
310 string const info = trim(data.substr(pos2));
314 // Search for all possible "required" keys
315 string author = parseBibTeX(data, "author");
317 author = parseBibTeX(data, "editor");
319 string year = parseBibTeX(data, "year");
320 string title = parseBibTeX(data, "title");
321 string booktitle = parseBibTeX(data, "booktitle");
322 string chapter = parseBibTeX(data, "chapter");
323 string number = parseBibTeX(data, "number");
324 string volume = parseBibTeX(data, "volume");
325 string pages = parseBibTeX(data, "pages");
327 string media = parseBibTeX(data, "journal");
329 media = parseBibTeX(data, "publisher");
331 media = parseBibTeX(data, "school");
333 media = parseBibTeX(data, "institution");
335 ostringstream result;
337 result << author << ", ";
340 if (!booktitle.empty())
341 result << ", in " << booktitle;
342 if (!chapter.empty())
343 result << ", Ch. " << chapter;
345 result << ", " << media;
347 result << ", vol. " << volume;
349 result << ", no. " << number;
351 result << ", pp. " << pages;
353 result << ", " << year;
355 string const result_str = rtrim(result.str());
356 if (!result_str.empty())
359 // This should never happen (or at least be very unusual!)
366 // Escape special chars.
367 // All characters are literals except: '.|*?+(){}[]^$\'
368 // These characters are literals when preceded by a "\", which is done here
369 string const escape_special_chars(string const & expr)
371 // Search for all chars '.|*?+(){}[^$]\'
372 // Note that '[' and '\' must be escaped.
373 // This is a limitation of boost::regex, but all other chars in BREs
374 // are assumed literal.
375 boost::RegEx reg("[].|*?+(){}^$\\[\\\\]");
377 // $& is a perl-like expression that expands to all of the current match
378 // The '$' must be prefixed with the escape character '\' for
379 // boost to treat it as a literal.
380 // Thus, to prefix a matched expression with '\', we use:
381 return reg.Merge(expr, "\\\\$&");
385 // A functor for use with std::find_if, used to ascertain whether a
386 // data entry matches the required regex_
387 struct RegexMatch : public std::unary_function<string, bool>
389 // re and icase are used to construct an instance of boost::RegEx.
390 // if icase is true, then matching is insensitive to case
391 RegexMatch(InfoMap const & m, string const & re, bool icase)
392 : map_(m), regex_(re, icase) {}
394 bool operator()(string const & key) const {
398 // the data searched is the key + its associated BibTeX/biblio
401 InfoMap::const_iterator info = map_.find(key);
402 if (info != map_.end())
403 data += ' ' + info->second;
405 // Attempts to find a match for the current RE
406 // somewhere in data.
407 return regex_.Search(data);
410 bool validRE() const { return regex_.error_code() == 0; }
414 mutable boost::RegEx regex_;
420 vector<string>::const_iterator
421 searchKeys(InfoMap const & theMap,
422 vector<string> const & keys,
423 string const & search_expr,
424 vector<string>::const_iterator start,
429 // Preliminary checks
430 if (start < keys.begin() || start >= keys.end())
433 string expr = trim(search_expr);
438 // We must escape special chars in the search_expr so that
439 // it is treated as a simple string by boost::regex.
440 expr = escape_special_chars(expr);
442 // Build the functor that will be passed to find_if.
443 RegexMatch const match(theMap, expr, !caseSensitive);
444 if (!match.validRE())
447 // Search the vector of 'keys' from 'start' for one that matches the
448 // predicate 'match'. Searching can be forward or backward from start.
450 return std::find_if(start, keys.end(), match);
452 vector<string>::const_reverse_iterator rit(start);
453 vector<string>::const_reverse_iterator rend = keys.rend();
454 rit = std::find_if(rit, rend, match);
458 // This is correct and always safe.
459 // (See Meyer's Effective STL, Item 28.)
460 return (++rit).base();
464 string const parseBibTeX(string data, string const & findkey)
467 // at first we delete all characters right of '%' and
468 // replace tabs through a space and remove leading spaces
469 // we read the data line by line so that the \n are
473 string dummy = token(data,'\n', Entries);
474 while (!dummy.empty()) {
475 dummy = subst(dummy, '\t', ' '); // no tabs
476 dummy = ltrim(dummy); // no leading spaces
477 // ignore lines with a beginning '%' or ignore all right of %
478 string::size_type const idx =
479 dummy.empty() ? string::npos : dummy.find('%');
480 if (idx != string::npos)
481 dummy.erase(idx, string::npos);
482 // do we have a new token or a new line of
483 // the same one? In the first case we ignore
484 // the \n and in the second we replace it
486 if (!dummy.empty()) {
487 if (!contains(dummy, '='))
488 data_ += ' ' + dummy;
492 dummy = token(data, '\n', ++Entries);
495 // replace double commas with "" for easy scanning
496 data = subst(data_, ",,", "\"\"");
502 // now get only the important line of the bibtex entry.
503 // all entries are devided by ',' except the last one.
504 data += ','; // now we have same behaviour for all entries
505 // because the last one is "blah ... }"
508 // parsing of title and booktitle is different from the
509 // others, because booktitle contains title
511 dummy = token(data, ',', Entries++);
512 if (!dummy.empty()) {
513 found = contains(ascii_lowercase(dummy), findkey);
514 if (findkey == "title" &&
515 contains(ascii_lowercase(dummy), "booktitle"))
518 } while (!found && !dummy.empty());
523 // we are not sure, if we get all, because "key= "blah, blah" is
525 // Therefore we read all until the next "=" character, which follows a
528 dummy = token(data, ',', Entries++);
529 while (!contains(dummy, '=') && !dummy.empty()) {
530 keyvalue += ',' + dummy;
531 dummy = token(data, ',', Entries++);
534 // replace double "" with originals ,, (two commas)
535 // leaving us with the all-important line
536 data = subst(keyvalue, "\"\"", ",,");
541 // 2. if there is no opening '{' then a closing '{' is probably cruft.
542 if (!contains(data, '{'))
543 data = rtrim(data, "}");
544 // happens, when last keyword
545 string::size_type const idx =
546 !data.empty() ? data.find('=') : string::npos;
548 if (idx == string::npos)
551 data = trim(data.substr(idx));
553 if (data.length() < 2 || data[0] != '=') { // a valid entry?
556 // delete '=' and the following spaces
557 data = ltrim(data, " =");
558 if (data.length() < 2) {
559 return data; // not long enough to find delimiters
561 string::size_type keypos = 1;
563 if (data[0] == '{') {
565 } else if (data[0] == '"') {
568 // no {} and no "", pure data but with a
569 // possible ',' at the end
570 return rtrim(data, ",");
572 string tmp = data.substr(keypos);
573 while (tmp.find('{') != string::npos &&
574 tmp.find('}') != string::npos &&
575 tmp.find('{') < tmp.find('}') &&
576 tmp.find('{') < tmp.find(enclosing)) {
578 keypos += tmp.find('{') + 1;
579 tmp = data.substr(keypos);
580 keypos += tmp.find('}') + 1;
581 tmp = data.substr(keypos);
583 if (tmp.find(enclosing) == string::npos)
586 keypos += tmp.find(enclosing);
587 return data.substr(1, keypos - 1);
597 char const * const citeCommands[] = {
598 "cite", "citet", "citep", "citealt", "citealp", "citeauthor",
599 "citeyear", "citeyearpar" };
601 unsigned int const nCiteCommands =
602 sizeof(citeCommands) / sizeof(char *);
604 CiteStyle const citeStyles[] = {
605 CITE, CITET, CITEP, CITEALT, CITEALP,
606 CITEAUTHOR, CITEYEAR, CITEYEARPAR };
608 unsigned int const nCiteStyles =
609 sizeof(citeStyles) / sizeof(CiteStyle);
611 CiteStyle const citeStylesFull[] = {
612 CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
614 unsigned int const nCiteStylesFull =
615 sizeof(citeStylesFull) / sizeof(CiteStyle);
617 CiteStyle const citeStylesUCase[] = {
618 CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
620 unsigned int const nCiteStylesUCase =
621 sizeof(citeStylesUCase) / sizeof(CiteStyle);
626 CitationStyle::CitationStyle(string const & command)
627 : style(CITE), full(false), forceUCase(false)
632 string cmd = command;
638 string::size_type const n = cmd.size() - 1;
639 if (cmd != "cite" && cmd[n] == '*') {
641 cmd = cmd.substr(0,n);
644 char const * const * const last = citeCommands + nCiteCommands;
645 char const * const * const ptr = std::find(citeCommands, last, cmd);
648 size_t idx = ptr - citeCommands;
649 style = citeStyles[idx];
654 string const CitationStyle::asLatexStr() const
656 string cite = citeCommands[style];
658 CiteStyle const * last = citeStylesFull + nCiteStylesFull;
659 if (std::find(citeStylesFull, last, style) != last)
664 CiteStyle const * last = citeStylesUCase + nCiteStylesUCase;
665 if (std::find(citeStylesUCase, last, style) != last)
673 CiteEngine_enum getEngine(Buffer const & buffer)
675 return buffer.params().cite_engine;
679 vector<CiteStyle> const getCiteStyles(CiteEngine_enum const & engine)
681 unsigned int nStyles = 0;
682 unsigned int start = 0;
689 case ENGINE_NATBIB_AUTHORYEAR:
690 case ENGINE_NATBIB_NUMERICAL:
691 nStyles = nCiteStyles - 1;
695 nStyles = nCiteStyles;
700 typedef vector<CiteStyle> cite_vec;
702 cite_vec styles(nStyles);
703 cite_vec::size_type i = 0;
705 for (; i != styles.size(); ++i, ++j)
706 styles[i] = citeStyles[j];
713 getNumericalStrings(string const & key,
714 InfoMap const & map, vector<CiteStyle> const & styles)
717 return vector<string>();
720 string const author = getAbbreviatedAuthor(map, key);
721 string const year = getYear(map, key);
722 if (author.empty() || year.empty())
723 return vector<string>();
725 vector<string> vec(styles.size());
726 for (vector<string>::size_type i = 0; i != vec.size(); ++i) {
736 str = author + " [#ID]";
740 str = author + " #ID";
756 str = '(' + year + ')';
768 getAuthorYearStrings(string const & key,
769 InfoMap const & map, vector<CiteStyle> const & styles)
772 return vector<string>();
775 string const author = getAbbreviatedAuthor(map, key);
776 string const year = getYear(map, key);
777 if (author.empty() || year.empty())
778 return vector<string>();
780 vector<string> vec(styles.size());
781 for (vector<string>::size_type i = 0; i != vec.size(); ++i) {
786 // jurabib only: Author/Annotator
787 // (i.e. the "before" field, 2nd opt arg)
788 str = author + "/<" + _("before") + '>';
792 str = author + " (" + year + ')';
796 str = '(' + author + ", " + year + ')';
800 str = author + ' ' + year ;
804 str = author + ", " + year ;
816 str = '(' + year + ')';
826 } // namespace biblio