src/BiblioInfo.cpp

   1 /**
   2  * \file BiblioInfo.cpp
   3  * This file is part of LyX, the document processor.
   4  * Licence details can be found in the file COPYING.
   5  *
   6  * \author Angus Leeming
   7  * \author Herbert Voß
   8  * \author Richard Heck
   9  *
  10  * Full author contact details are available in file CREDITS.
  11  */
  12
  13 #include <config.h>
  14
  15 #include "BiblioInfo.h"
  16 #include "Buffer.h"
  17 #include "BufferParams.h"
  18 #include "buffer_funcs.h"
  19 #include "Encoding.h"
  20 #include "InsetIterator.h"
  21 #include "Paragraph.h"
  22 #include "TocBackend.h"
  23
  24 #include "insets/Inset.h"
  25 #include "insets/InsetBibitem.h"
  26 #include "insets/InsetBibtex.h"
  27 #include "insets/InsetInclude.h"
  28
  29 #include "support/convert.h"
  30 #include "support/debug.h"
  31 #include "support/docstream.h"
  32 #include "support/gettext.h"
  33 #include "support/lassert.h"
  34 #include "support/lstrings.h"
  35 #include "support/textutils.h"
  36
  37 #include "boost/regex.hpp"
  38
  39 #include <set>
  40
  41 using namespace std;
  42 using namespace lyx::support;
  43
  44
  45 namespace lyx {
  46
  47 namespace {
  48
  49 // gets the "family name" from an author-type string
  50 docstring familyName(docstring const & name)
  51 {
  52         if (name.empty())
  53                 return docstring();
  54
  55         // first we look for a comma, and take the last name to be everything
  56         // preceding the right-most one, so that we also get the "jr" part.
  57         docstring::size_type idx = name.rfind(',');
  58         if (idx != docstring::npos)
  59                 return ltrim(name.substr(0, idx));
  60
  61         // OK, so now we want to look for the last name. We're going to
  62         // include the "von" part. This isn't perfect.
  63         // Split on spaces, to get various tokens.
  64         vector<docstring> pieces = getVectorFromString(name, from_ascii(" "));
  65         // If we only get two, assume the last one is the last name
  66         if (pieces.size() <= 2)
  67                 return pieces.back();
  68
  69         // Now we look for the first token that begins with a lower case letter.
  70         vector<docstring>::const_iterator it = pieces.begin();
  71         vector<docstring>::const_iterator en = pieces.end();
  72         for (; it != en; ++it) {
  73                 if ((*it).size() == 0)
  74                         continue;
  75                 char_type const c = (*it)[0];
  76                 if (isLower(c))
  77                         break;
  78         }
  79
  80         if (it == en) // we never found a "von"
  81                 return pieces.back();
  82
  83         // reconstruct what we need to return
  84         docstring retval;
  85         bool first = true;
  86         for (; it != en; ++it) {
  87                 if (!first)
  88                         retval += " ";
  89                 else
  90                         first = false;
  91                 retval += *it;
  92         }
  93         return retval;
  94 }
  95
  96 // converts a string containing LaTeX commands into unicode
  97 // for display.
  98 docstring convertLaTeXCommands(docstring const & str)
  99 {
 100         docstring val = str;
 101         docstring ret;
 102
 103         bool scanning_cmd = false;
 104         bool scanning_math = false;
 105         bool escaped = false; // used to catch \$, etc.
 106         while (val.size()) {
 107                 char_type const ch = val[0];
 108
 109                 // if we're scanning math, we output everything until we
 110                 // find an unescaped $, at which point we break out.
 111                 if (scanning_math) {
 112                         if (escaped)
 113                                 escaped = false;
 114                         else if (ch == '\\')
 115                                 escaped = true;
 116                         else if (ch == '$')
 117                                 scanning_math = false;
 118                         ret += ch;
 119                         val = val.substr(1);
 120                         continue;
 121                 }
 122
 123                 // if we're scanning a command name, then we just
 124                 // discard characters until we hit something that
 125                 // isn't alpha.
 126                 if (scanning_cmd) {
 127                         if (isAlphaASCII(ch)) {
 128                                 val = val.substr(1);
 129                                 escaped = false;
 130                                 continue;
 131                         }
 132                         // so we're done with this command.
 133                         // now we fall through and check this character.
 134                         scanning_cmd = false;
 135                 }
 136
 137                 // was the last character a \? If so, then this is something like:
 138                 // \\ or \$, so we'll just output it. That's probably not always right...
 139                 if (escaped) {
 140                         // exception: output \, as THIN SPACE
 141                         if (ch == ',')
 142                                 ret.push_back(0x2009);
 143                         else
 144                                 ret += ch;
 145                         val = val.substr(1);
 146                         escaped = false;
 147                         continue;
 148                 }
 149
 150                 if (ch == '$') {
 151                         ret += ch;
 152                         val = val.substr(1);
 153                         scanning_math = true;
 154                         continue;
 155                 }
 156
 157                 // we just ignore braces
 158                 if (ch == '{' || ch == '}') {
 159                         val = val.substr(1);
 160                         continue;
 161                 }
 162
 163                 // we're going to check things that look like commands, so if
 164                 // this doesn't, just output it.
 165                 if (ch != '\\') {
 166                         ret += ch;
 167                         val = val.substr(1);
 168                         continue;
 169                 }
 170
 171                 // ok, could be a command of some sort
 172                 // let's see if it corresponds to some unicode
 173                 // unicodesymbols has things in the form: \"{u},
 174                 // whereas we may see things like: \"u. So we'll
 175                 // look for that and change it, if necessary.
 176                 static boost::regex const reg("^\\\\\\W\\w");
 177                 if (boost::regex_search(to_utf8(val), reg)) {
 178                         val.insert(3, from_ascii("}"));
 179                         val.insert(2, from_ascii("{"));
 180                 }
 181                 docstring rem;
 182                 docstring const cnvtd = Encodings::fromLaTeXCommand(val, rem,
 183                                                         Encodings::TEXT_CMD);
 184                 if (!cnvtd.empty()) {
 185                         // it did, so we'll take that bit and proceed with what's left
 186                         ret += cnvtd;
 187                         val = rem;
 188                         continue;
 189                 }
 190                 // it's a command of some sort
 191                 scanning_cmd = true;
 192                 escaped = true;
 193                 val = val.substr(1);
 194         }
 195         return ret;
 196 }
 197
 198 } // anon namespace
 199
 200
 201 //////////////////////////////////////////////////////////////////////
 202 //
 203 // BibTeXInfo
 204 //
 205 //////////////////////////////////////////////////////////////////////
 206
 207 BibTeXInfo::BibTeXInfo(docstring const & key, docstring const & type)
 208         : is_bibtex_(true), bib_key_(key), entry_type_(type), info_(),
 209           modifier_(0)
 210 {}
 211
 212
 213 docstring const BibTeXInfo::getAbbreviatedAuthor() const
 214 {
 215         if (!is_bibtex_) {
 216                 docstring const opt = label();
 217                 if (opt.empty())
 218                         return docstring();
 219
 220                 docstring authors;
 221                 docstring const remainder = trim(split(opt, authors, '('));
 222                 if (remainder.empty())
 223                         // in this case, we didn't find a "(",
 224                         // so we don't have author (year)
 225                         return docstring();
 226                 return authors;
 227         }
 228
 229         docstring author = convertLaTeXCommands(operator[]("author"));
 230         if (author.empty()) {
 231                 author = convertLaTeXCommands(operator[]("editor"));
 232                 if (author.empty())
 233                         return bib_key_;
 234         }
 235
 236         // FIXME Move this to a separate routine that can
 237         // be called from elsewhere.
 238         //
 239         // OK, we've got some names. Let's format them.
 240         // Try to split the author list on " and "
 241         vector<docstring> const authors =
 242                 getVectorFromString(author, from_ascii(" and "));
 243
 244         if (authors.size() == 2)
 245                 return bformat(_("%1$s and %2$s"),
 246                         familyName(authors[0]), familyName(authors[1]));
 247
 248         if (authors.size() > 2)
 249                 return bformat(_("%1$s et al."), familyName(authors[0]));
 250
 251         return familyName(authors[0]);
 252 }
 253
 254
 255 docstring const BibTeXInfo::getYear() const
 256 {
 257         if (is_bibtex_)
 258                 return operator[]("year");
 259
 260         docstring const opt = label();
 261         if (opt.empty())
 262                 return docstring();
 263
 264         docstring authors;
 265         docstring tmp = split(opt, authors, '(');
 266         if (tmp.empty())
 267                 // we don't have author (year)
 268                 return docstring();
 269         docstring year;
 270         tmp = split(tmp, year, ')');
 271         return year;
 272 }
 273
 274
 275 docstring const BibTeXInfo::getXRef() const
 276 {
 277         if (!is_bibtex_)
 278                 return docstring();
 279         return operator[]("crossref");
 280 }
 281
 282
 283 namespace {
 284         string parseOptions(string const & format, string & optkey,
 285                         string & ifpart, string & elsepart);
 286
 287         /// Calls parseOptions to deal with an embedded option, such as:
 288         ///   {%number%[[, no.~%number%]]}
 289         /// which must appear at the start of format. ifelsepart gets the
 290         /// whole of the option, and we return what's left after the option.
 291         /// we return format if there is an error.
 292         string parseEmbeddedOption(string const & format, string & ifelsepart)
 293         {
 294                 LASSERT(format[0] == '{' && format[1] == '%', return format);
 295                 string optkey;
 296                 string ifpart;
 297                 string elsepart;
 298                 string const rest = parseOptions(format, optkey, ifpart, elsepart);
 299                 if (format == rest) { // parse error
 300                         LYXERR0("ERROR! Couldn't parse `" << format <<"'.");
 301                         return format;
 302                 }
 303                 LASSERT(rest.size() <= format.size(), /* */);
 304                 ifelsepart = format.substr(0, format.size() - rest.size());
 305                 return rest;
 306         }
 307
 308
 309         // Gets a "clause" from a format string, where the clause is
 310         // delimited by '[[' and ']]'. Returns what is left after the
 311         // clause is removed, and returns format if there is an error.
 312         string getClause(string const & format, string & clause)
 313         {
 314                 string fmt = format;
 315                 // remove '[['
 316                 fmt = fmt.substr(2);
 317                 // we'll remove characters from the front of fmt as we
 318                 // deal with them
 319                 while (fmt.size()) {
 320                         if (fmt[0] == ']' && fmt.size() > 1 && fmt[1] == ']') {
 321                           // that's the end
 322                                 fmt = fmt.substr(2);
 323                                 break;
 324                         }
 325                         // check for an embedded option
 326                         if (fmt[0] == '{' && fmt.size() > 1 && fmt[1] == '%') {
 327                                 string part;
 328                                 string const rest = parseEmbeddedOption(fmt, part);
 329                                 if (fmt == rest) {
 330                                         LYXERR0("ERROR! Couldn't parse `" << format <<"'.");
 331                                         return format;
 332                                 }
 333                                 clause += part;
 334                                 fmt = rest;
 335                         } else { // it's just a normal character
 336                                 clause += fmt[0];
 337                                 fmt = fmt.substr(1);
 338                         }
 339                 }
 340                 return fmt;
 341         }
 342
 343
 344         /// parse an options string, which must appear at the start of the
 345         /// format parameter. puts the parsed bits in optkey, ifpart, and
 346         /// elsepart and returns what's left after the option is removed.
 347         /// if there's an error, it returns format itself.
 348         string parseOptions(string const & format, string & optkey,
 349                         string & ifpart, string & elsepart)
 350         {
 351                 LASSERT(format[0] == '{' && format[1] == '%', return format);
 352                 // strip '{%'
 353                 string fmt = format.substr(2);
 354                 size_t pos = fmt.find('%'); // end of key
 355                 if (pos == string::npos) {
 356                         LYXERR0("Error parsing  `" << format <<"'. Can't find end of key.");
 357                         return format;
 358                 }
 359                 optkey = fmt.substr(0,pos);
 360                 fmt = fmt.substr(pos + 1);
 361                 // [[format]] should be next
 362                 if (fmt[0] != '[' || fmt[1] != '[') {
 363                         LYXERR0("Error parsing  `" << format <<"'. Can't find '[[' after key.");
 364                         return format;
 365                 }
 366
 367                 string curfmt = fmt;
 368                 fmt = getClause(curfmt, ifpart);
 369                 if (fmt == curfmt) {
 370                         LYXERR0("Error parsing  `" << format <<"'. Couldn't get if clause.");
 371                         return format;
 372                 }
 373
 374                 if (fmt[0] == '}') // we're done, no else clause
 375                         return fmt.substr(1);
 376
 377                 // else part should follow
 378                 if (fmt[0] != '[' || fmt[1] != '[') {
 379                         LYXERR0("Error parsing  `" << format <<"'. Can't find else clause.");
 380                         return format;
 381                 }
 382
 383                 curfmt = fmt;
 384                 fmt = getClause(curfmt, elsepart);
 385                 // we should be done
 386                 if (fmt == curfmt || fmt[0] != '}') {
 387                         LYXERR0("Error parsing  `" << format <<"'. Can't find end of option.");
 388                         return format;
 389                 }
 390                 return fmt.substr(1);
 391 }
 392
 393 } // anon namespace
 394
 395
 396 docstring BibTeXInfo::expandFormat(string const & format,
 397                 BibTeXInfo const * const xref, bool richtext) const
 398 {
 399         // return value
 400         docstring ret;
 401         string key;
 402         bool scanning_key = false;
 403         bool scanning_rich = false;
 404
 405         string fmt = format;
 406         // we'll remove characters from the front of fmt as we
 407         // deal with them
 408         while (fmt.size()) {
 409                 char_type thischar = fmt[0];
 410                 if (thischar == '%') {
 411                         // beginning or end of key
 412                         if (scanning_key) {
 413                                 // end of key
 414                                 scanning_key = false;
 415                                 // so we replace the key with its value, which may be empty
 416                                 docstring const val = getValueForKey(key, xref);
 417                                 key.clear();
 418                                 ret += val;
 419                         } else {
 420                                 // beginning of key
 421                                 scanning_key = true;
 422                         }
 423                 }
 424                 else if (thischar == '{') {
 425                         // beginning of option?
 426                         if (scanning_key) {
 427                                 LYXERR0("ERROR: Found `{' when scanning key in `" << format << "'.");
 428                                 return _("ERROR!");
 429                         }
 430                         if (fmt.size() > 1) {
 431                                 if (fmt[1] == '%') {
 432                                         // it is the beginning of an optional format
 433                                         string optkey;
 434                                         string ifpart;
 435                                         string elsepart;
 436                                         string const newfmt =
 437                                                 parseOptions(fmt, optkey, ifpart, elsepart);
 438                                         if (newfmt == fmt) // parse error
 439                                                 return _("ERROR!");
 440                                         fmt = newfmt;
 441                                         docstring const val = getValueForKey(optkey, xref);
 442                                         if (!val.empty())
 443                                                 ret += expandFormat(ifpart, xref, richtext);
 444                                         else if (!elsepart.empty())
 445                                                 ret += expandFormat(elsepart, xref, richtext);
 446                                         // fmt will have been shortened for us already
 447                                         continue;
 448                                 }
 449                                 if (fmt[1] == '!') {
 450                                         // beginning of rich text
 451                                         scanning_rich = true;
 452                                         fmt = fmt.substr(2);
 453                                         continue;
 454                                 }
 455                         }
 456                         // we are here if the '{' was at the end of the format. hmm.
 457                         ret += thischar;
 458                 }
 459                 else if (scanning_rich && thischar == '!'
 460                          && fmt.size() > 1 && fmt[1] == '}') {
 461                         // end of rich text
 462                         scanning_rich = false;
 463                         fmt = fmt.substr(2);
 464                         continue;
 465                 }
 466                 else if (scanning_key)
 467                         key += thischar;
 468                 else if (richtext || !scanning_rich)
 469                         ret += thischar;
 470                 // else the character is discarded, which will happen only if
 471                 // richtext == false and we are scanning rich text
 472                 fmt = fmt.substr(1);
 473         } // for loop
 474         if (scanning_key) {
 475                 LYXERR0("Never found end of key in `" << format << "'!");
 476                 return _("ERROR!");
 477         }
 478         if (scanning_rich) {
 479                 LYXERR0("Never found end of rich text in `" << format << "'!");
 480                 return _("ERROR!");
 481         }
 482         return ret;
 483 }
 484
 485
 486 namespace {
 487
 488 // FIXME These would be better read from a file, so that they
 489 // could be customized.
 490
 491         static string articleFormat = "%author%, \"%title%\", {!<i>!}%journal%{!</i>!} {%volume%[[ %volume%{%number%[[, %number%]]}]]} (%year%){%pages%[[, pp. %pages%]]}.{%note%[[ %note%]]}";
 492
 493         static string bookFormat = "{%author%[[%author%]][[%editor%, ed.]]}, {!<i>!}%title%{!</i>!}{%volume%[[ vol. %volume%]][[{%number%[[no. %number%]]}]]}{%edition%[[%edition%]]} ({%address%[[%address%: ]]}%publisher%, %year%).{%note%[[ %note%]]}";
 494
 495         static string inSomething = "%author%, \"%title%\", in{%editor%[[ %editor%, ed.,]]} {!<i>!}%booktitle%{!</i>!}{%volume%[[ vol. %volume%]][[{%number%[[no. %number%]]}]]}{%edition%[[%edition%]]} ({%address%[[%address%: ]]}%publisher%, %year%){%pages%[[, pp. %pages%]]}.{%note%[[ %note%]]}";
 496
 497         static string thesis = "%author%, %title% ({%address%[[%address%: ]]}%school%, %year%).{%note%[[ %note%]]}";
 498
 499         static string defaultFormat = "{%author%[[%author%, ]][[{%editor%[[%editor%, ed., ]]}]]}\"%title%\"{%journal%[[, {!<i>!}%journal%{!</i>!}]][[{%publisher%[[, %publisher%]][[{%institution%[[, %institution%]]}]]}]]}{%year%[[ (%year%)]]}{%pages%[[, %pages%]]}.";
 500
 501 }
 502
 503 docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref,
 504         bool richtext) const
 505 {
 506         if (!info_.empty())
 507                 return info_;
 508
 509         if (!is_bibtex_) {
 510                 BibTeXInfo::const_iterator it = find(from_ascii("ref"));
 511                 info_ = it->second;
 512                 return info_;
 513         }
 514
 515         if (entry_type_ == "article")
 516                 info_ = expandFormat(articleFormat, xref, richtext);
 517         else if (entry_type_ == "book")
 518                 info_ = expandFormat(bookFormat, xref, richtext);
 519         else if (entry_type_.substr(0,2) == "in")
 520                 info_ = expandFormat(inSomething, xref, richtext);
 521         else if (entry_type_ == "phdthesis" || entry_type_ == "mastersthesis")
 522                 info_ = expandFormat(thesis, xref, richtext);
 523         else
 524                 info_ = expandFormat(defaultFormat, xref, richtext);
 525
 526         if (!info_.empty())
 527                 info_ = convertLaTeXCommands(info_);
 528         return info_;
 529 }
 530
 531
 532 docstring const & BibTeXInfo::operator[](docstring const & field) const
 533 {
 534         BibTeXInfo::const_iterator it = find(field);
 535         if (it != end())
 536                 return it->second;
 537         static docstring const empty_value = docstring();
 538         return empty_value;
 539 }
 540
 541
 542 docstring const & BibTeXInfo::operator[](string const & field) const
 543 {
 544         return operator[](from_ascii(field));
 545 }
 546
 547
 548 docstring BibTeXInfo::getValueForKey(string const & key,
 549                 BibTeXInfo const * const xref) const
 550 {
 551         docstring const ret = operator[](key);
 552         if (!ret.empty() || !xref)
 553                 return ret;
 554         return (*xref)[key];
 555 }
 556
 557
 558 //////////////////////////////////////////////////////////////////////
 559 //
 560 // BiblioInfo
 561 //
 562 //////////////////////////////////////////////////////////////////////
 563
 564 namespace {
 565 // A functor for use with sort, leading to case insensitive sorting
 566         class compareNoCase: public binary_function<docstring, docstring, bool>
 567         {
 568                 public:
 569                         bool operator()(docstring const & s1, docstring const & s2) const {
 570                                 return compare_no_case(s1, s2) < 0;
 571                         }
 572         };
 573 } // namespace anon
 574
 575
 576 vector<docstring> const BiblioInfo::getKeys() const
 577 {
 578         vector<docstring> bibkeys;
 579         BiblioInfo::const_iterator it  = begin();
 580         for (; it != end(); ++it)
 581                 bibkeys.push_back(it->first);
 582         sort(bibkeys.begin(), bibkeys.end(), compareNoCase());
 583         return bibkeys;
 584 }
 585
 586
 587 vector<docstring> const BiblioInfo::getFields() const
 588 {
 589         vector<docstring> bibfields;
 590         set<docstring>::const_iterator it = field_names_.begin();
 591         set<docstring>::const_iterator end = field_names_.end();
 592         for (; it != end; ++it)
 593                 bibfields.push_back(*it);
 594         sort(bibfields.begin(), bibfields.end());
 595         return bibfields;
 596 }
 597
 598
 599 vector<docstring> const BiblioInfo::getEntries() const
 600 {
 601         vector<docstring> bibentries;
 602         set<docstring>::const_iterator it = entry_types_.begin();
 603         set<docstring>::const_iterator end = entry_types_.end();
 604         for (; it != end; ++it)
 605                 bibentries.push_back(*it);
 606         sort(bibentries.begin(), bibentries.end());
 607         return bibentries;
 608 }
 609
 610
 611 docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key) const
 612 {
 613         BiblioInfo::const_iterator it = find(key);
 614         if (it == end())
 615                 return docstring();
 616         BibTeXInfo const & data = it->second;
 617         return data.getAbbreviatedAuthor();
 618 }
 619
 620
 621 docstring const BiblioInfo::getCiteNumber(docstring const & key) const
 622 {
 623         BiblioInfo::const_iterator it = find(key);
 624         if (it == end())
 625                 return docstring();
 626         BibTeXInfo const & data = it->second;
 627         return data.citeNumber();
 628 }
 629
 630
 631 docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier) const
 632 {
 633         BiblioInfo::const_iterator it = find(key);
 634         if (it == end())
 635                 return docstring();
 636         BibTeXInfo const & data = it->second;
 637         docstring year = data.getYear();
 638         if (year.empty()) {
 639                 // let's try the crossref
 640                 docstring const xref = data.getXRef();
 641                 if (xref.empty())
 642                         return _("No year"); // no luck
 643                 BiblioInfo::const_iterator const xrefit = find(xref);
 644                 if (xrefit == end())
 645                         return _("No year"); // no luck again
 646                 BibTeXInfo const & xref_data = xrefit->second;
 647                 year = xref_data.getYear();
 648         }
 649         if (use_modifier && data.modifier() != 0)
 650                 year += data.modifier();
 651         return year;
 652 }
 653
 654
 655 docstring const BiblioInfo::getInfo(docstring const & key, bool richtext) const
 656 {
 657         BiblioInfo::const_iterator it = find(key);
 658         if (it == end())
 659                 return docstring();
 660         BibTeXInfo const & data = it->second;
 661         BibTeXInfo const * xrefptr = 0;
 662         docstring const xref = data.getXRef();
 663         if (!xref.empty()) {
 664                 BiblioInfo::const_iterator const xrefit = find(xref);
 665                 if (xrefit != end())
 666                         xrefptr = &(xrefit->second);
 667         }
 668         return data.getInfo(xrefptr, richtext);
 669 }
 670
 671
 672 bool BiblioInfo::isBibtex(docstring const & key) const
 673 {
 674         BiblioInfo::const_iterator it = find(key);
 675         if (it == end())
 676                 return false;
 677         return it->second.isBibTeX();
 678 }
 679
 680
 681
 682 vector<docstring> const BiblioInfo::getCiteStrings(
 683         docstring const & key, Buffer const & buf) const
 684 {
 685         CiteEngine const engine = buf.params().citeEngine();
 686         if (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL)
 687                 return getNumericalStrings(key, buf);
 688         else
 689                 return getAuthorYearStrings(key, buf);
 690 }
 691
 692
 693 vector<docstring> const BiblioInfo::getNumericalStrings(
 694         docstring const & key, Buffer const & buf) const
 695 {
 696         if (empty())
 697                 return vector<docstring>();
 698
 699         docstring const author = getAbbreviatedAuthor(key);
 700         docstring const year   = getYear(key);
 701         if (author.empty() || year.empty())
 702                 return vector<docstring>();
 703
 704         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 705
 706         vector<docstring> vec(styles.size());
 707         for (size_t i = 0; i != vec.size(); ++i) {
 708                 docstring str;
 709
 710                 switch (styles[i]) {
 711                         case CITE:
 712                         case CITEP:
 713                                 str = from_ascii("[#ID]");
 714                                 break;
 715
 716                         case NOCITE:
 717                                 str = _("Add to bibliography only.");
 718                                 break;
 719
 720                         case CITET:
 721                                 str = author + " [#ID]";
 722                                 break;
 723
 724                         case CITEALT:
 725                                 str = author + " #ID";
 726                                 break;
 727
 728                         case CITEALP:
 729                                 str = from_ascii("#ID");
 730                                 break;
 731
 732                         case CITEAUTHOR:
 733                                 str = author;
 734                                 break;
 735
 736                         case CITEYEAR:
 737                                 str = year;
 738                                 break;
 739
 740                         case CITEYEARPAR:
 741                                 str = '(' + year + ')';
 742                                 break;
 743                 }
 744
 745                 vec[i] = str;
 746         }
 747
 748         return vec;
 749 }
 750
 751
 752 vector<docstring> const BiblioInfo::getAuthorYearStrings(
 753         docstring const & key, Buffer const & buf) const
 754 {
 755         if (empty())
 756                 return vector<docstring>();
 757
 758         docstring const author = getAbbreviatedAuthor(key);
 759         docstring const year   = getYear(key);
 760         if (author.empty() || year.empty())
 761                 return vector<docstring>();
 762
 763         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 764
 765         vector<docstring> vec(styles.size());
 766         for (size_t i = 0; i != vec.size(); ++i) {
 767                 docstring str;
 768
 769                 switch (styles[i]) {
 770                         case CITE:
 771                 // jurabib only: Author/Annotator
 772                 // (i.e. the "before" field, 2nd opt arg)
 773                                 str = author + "/<" + _("before") + '>';
 774                                 break;
 775
 776                         case NOCITE:
 777                                 str = _("Add to bibliography only.");
 778                                 break;
 779
 780                         case CITET:
 781                                 str = author + " (" + year + ')';
 782                                 break;
 783
 784                         case CITEP:
 785                                 str = '(' + author + ", " + year + ')';
 786                                 break;
 787
 788                         case CITEALT:
 789                                 str = author + ' ' + year ;
 790                                 break;
 791
 792                         case CITEALP:
 793                                 str = author + ", " + year ;
 794                                 break;
 795
 796                         case CITEAUTHOR:
 797                                 str = author;
 798                                 break;
 799
 800                         case CITEYEAR:
 801                                 str = year;
 802                                 break;
 803
 804                         case CITEYEARPAR:
 805                                 str = '(' + year + ')';
 806                                 break;
 807                 }
 808                 vec[i] = str;
 809         }
 810         return vec;
 811 }
 812
 813
 814 void BiblioInfo::mergeBiblioInfo(BiblioInfo const & info)
 815 {
 816         bimap_.insert(info.begin(), info.end());
 817 }
 818
 819
 820 namespace {
 821         // used in xhtml to sort a list of BibTeXInfo objects
 822         bool lSorter(BibTeXInfo const * lhs, BibTeXInfo const * rhs)
 823         {
 824                 docstring const lauth = lhs->getAbbreviatedAuthor();
 825                 docstring const rauth = rhs->getAbbreviatedAuthor();
 826                 docstring const lyear = lhs->getYear();
 827                 docstring const ryear = rhs->getYear();
 828                 docstring const ltitl = lhs->operator[]("title");
 829                 docstring const rtitl = rhs->operator[]("title");
 830                 return  (lauth < rauth)
 831                                 || (lauth == rauth && lyear < ryear)
 832                                 || (lauth == rauth && lyear == ryear && ltitl < rtitl);
 833         }
 834 }
 835
 836
 837 void BiblioInfo::collectCitedEntries(Buffer const & buf)
 838 {
 839         cited_entries_.clear();
 840         // We are going to collect all the citation keys used in the document,
 841         // getting them from the TOC.
 842         // FIXME We may want to collect these differently, in the first case,
 843         // so that we might have them in order of appearance.
 844         set<docstring> citekeys;
 845         Toc const & toc = buf.tocBackend().toc("citation");
 846         Toc::const_iterator it = toc.begin();
 847         Toc::const_iterator const en = toc.end();
 848         for (; it != en; ++it) {
 849                 if (it->str().empty())
 850                         continue;
 851                 vector<docstring> const keys = getVectorFromString(it->str());
 852                 citekeys.insert(keys.begin(), keys.end());
 853         }
 854         if (citekeys.empty())
 855                 return;
 856
 857         // We have a set of the keys used in this document.
 858         // We will now convert it to a list of the BibTeXInfo objects used in
 859         // this document...
 860         vector<BibTeXInfo const *> bi;
 861         set<docstring>::const_iterator cit = citekeys.begin();
 862         set<docstring>::const_iterator const cen = citekeys.end();
 863         for (; cit != cen; ++cit) {
 864                 BiblioInfo::const_iterator const bt = find(*cit);
 865                 if (bt == end() || !bt->second.isBibTeX())
 866                         continue;
 867                 bi.push_back(&(bt->second));
 868         }
 869         // ...and sort it.
 870         sort(bi.begin(), bi.end(), lSorter);
 871
 872         // Now we can write the sorted keys
 873         vector<BibTeXInfo const *>::const_iterator bit = bi.begin();
 874         vector<BibTeXInfo const *>::const_iterator ben = bi.end();
 875         for (; bit != ben; ++bit)
 876                 cited_entries_.push_back((*bit)->key());
 877 }
 878
 879
 880 void BiblioInfo::makeCitationLabels(Buffer const & buf)
 881 {
 882         collectCitedEntries(buf);
 883         CiteEngine const engine = buf.params().citeEngine();
 884         bool const numbers =
 885                 (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL);
 886
 887         int keynumber = 0;
 888         char modifier = 0;
 889         // used to remember the last one we saw
 890         // we'll be comparing entries to see if we need to add
 891         // modifiers, like "1984a"
 892         map<docstring, BibTeXInfo>::iterator last;
 893
 894         vector<docstring>::const_iterator it = cited_entries_.begin();
 895         vector<docstring>::const_iterator const en = cited_entries_.end();
 896         for (; it != en; ++it) {
 897                 map<docstring, BibTeXInfo>::iterator const biit = bimap_.find(*it);
 898                 // this shouldn't happen, but...
 899                 if (biit == bimap_.end())
 900                         // ...fail gracefully, anyway.
 901                         continue;
 902                 BibTeXInfo & entry = biit->second;
 903                 if (numbers) {
 904                         docstring const num = convert<docstring>(++keynumber);
 905                         entry.setCiteNumber(num);
 906                 } else {
 907                         if (it != cited_entries_.begin()
 908                             && entry.getAbbreviatedAuthor() == last->second.getAbbreviatedAuthor()
 909                             // we access the year via getYear() so as to get it from the xref,
 910                             // if we need to do so
 911                             && getYear(entry.key()) == getYear(last->second.key())) {
 912                                 if (modifier == 0) {
 913                                         // so the last one should have been 'a'
 914                                         last->second.setModifier('a');
 915                                         modifier = 'b';
 916                                 } else if (modifier == 'z')
 917                                         modifier = 'A';
 918                                 else
 919                                         modifier++;
 920                         } else {
 921                                 modifier = 0;
 922                         }
 923                         entry.setModifier(modifier);
 924                         // remember the last one
 925                         last = biit;
 926                 }
 927         }
 928 }
 929
 930
 931 //////////////////////////////////////////////////////////////////////
 932 //
 933 // CitationStyle
 934 //
 935 //////////////////////////////////////////////////////////////////////
 936
 937 namespace {
 938
 939
 940 char const * const citeCommands[] = {
 941         "cite", "citet", "citep", "citealt", "citealp",
 942         "citeauthor", "citeyear", "citeyearpar", "nocite" };
 943
 944 unsigned int const nCiteCommands =
 945                 sizeof(citeCommands) / sizeof(char *);
 946
 947 CiteStyle const citeStylesArray[] = {
 948         CITE, CITET, CITEP, CITEALT, CITEALP,
 949         CITEAUTHOR, CITEYEAR, CITEYEARPAR, NOCITE };
 950
 951 unsigned int const nCiteStyles =
 952                 sizeof(citeStylesArray) / sizeof(CiteStyle);
 953
 954 CiteStyle const citeStylesFull[] = {
 955         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 956
 957 unsigned int const nCiteStylesFull =
 958                 sizeof(citeStylesFull) / sizeof(CiteStyle);
 959
 960 CiteStyle const citeStylesUCase[] = {
 961         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 962
 963 unsigned int const nCiteStylesUCase =
 964         sizeof(citeStylesUCase) / sizeof(CiteStyle);
 965
 966 } // namespace anon
 967
 968
 969 CitationStyle citationStyleFromString(string const & command)
 970 {
 971         CitationStyle s;
 972         if (command.empty())
 973                 return s;
 974
 975         string cmd = command;
 976         if (cmd[0] == 'C') {
 977                 s.forceUpperCase = true;
 978                 cmd[0] = 'c';
 979         }
 980
 981         size_t const n = cmd.size() - 1;
 982         if (cmd != "cite" && cmd[n] == '*') {
 983                 s.full = true;
 984                 cmd = cmd.substr(0, n);
 985         }
 986
 987         char const * const * const last = citeCommands + nCiteCommands;
 988         char const * const * const ptr = find(citeCommands, last, cmd);
 989
 990         if (ptr != last) {
 991                 size_t idx = ptr - citeCommands;
 992                 s.style = citeStylesArray[idx];
 993         }
 994         return s;
 995 }
 996
 997
 998 string citationStyleToString(const CitationStyle & s)
 999 {
1000         string cite = citeCommands[s.style];
1001         if (s.full) {
1002                 CiteStyle const * last = citeStylesFull + nCiteStylesFull;
1003                 if (std::find(citeStylesFull, last, s.style) != last)
1004                         cite += '*';
1005         }
1006
1007         if (s.forceUpperCase) {
1008                 CiteStyle const * last = citeStylesUCase + nCiteStylesUCase;
1009                 if (std::find(citeStylesUCase, last, s.style) != last)
1010                         cite[0] = 'C';
1011         }
1012
1013         return cite;
1014 }
1015
1016 vector<CiteStyle> citeStyles(CiteEngine engine)
1017 {
1018         unsigned int nStyles = 0;
1019         unsigned int start = 0;
1020
1021         switch (engine) {
1022                 case ENGINE_BASIC:
1023                         nStyles = 2;
1024                         start = 0;
1025                         break;
1026                 case ENGINE_NATBIB_AUTHORYEAR:
1027                 case ENGINE_NATBIB_NUMERICAL:
1028                         nStyles = nCiteStyles - 1;
1029                         start = 1;
1030                         break;
1031                 case ENGINE_JURABIB:
1032                         nStyles = nCiteStyles;
1033                         start = 0;
1034                         break;
1035         }
1036
1037         vector<CiteStyle> styles(nStyles);
1038         size_t i = 0;
1039         int j = start;
1040         for (; i != styles.size(); ++i, ++j)
1041                 styles[i] = citeStylesArray[j];
1042
1043         return styles;
1044 }
1045
1046 } // namespace lyx
1047