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         docstring parseOptions(docstring const & format, docstring & optkey,
 285                         docstring & ifpart, docstring & 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         docstring parseEmbeddedOption(docstring const & format,
 293                         docstring & ifelsepart)
 294         {
 295                 LASSERT(format[0] == '{' && format[1] == '%', return format);
 296                 docstring optkey;
 297                 docstring ifpart;
 298                 docstring elsepart;
 299                 docstring const rest = parseOptions(format, optkey, ifpart, elsepart);
 300                 if (format == rest) { // parse error
 301                         LYXERR0("ERROR! Couldn't parse `" << format <<"'.");
 302                         return format;
 303                 }
 304                 LASSERT(rest.size() <= format.size(), /* */);
 305                 ifelsepart = format.substr(0, format.size() - rest.size());
 306                 return rest;
 307         }
 308
 309
 310         // Gets a "clause" from a format string, where the clause is
 311         // delimited by '[[' and ']]'. Returns what is left after the
 312         // clause is removed, and returns format if there is an error.
 313         docstring getClause(docstring const & format, docstring & clause)
 314         {
 315                 docstring fmt = format;
 316                 // remove '[['
 317                 fmt = fmt.substr(2);
 318                 // we'll remove characters from the front of fmt as we
 319                 // deal with them
 320                 while (fmt.size()) {
 321                         if (fmt[0] == ']' && fmt.size() > 1 && fmt[1] == ']') {
 322                           // that's the end
 323                                 fmt = fmt.substr(2);
 324                                 break;
 325                         }
 326                         // check for an embedded option
 327                         if (fmt[0] == '{' && fmt.size() > 1 && fmt[1] == '%') {
 328                                 docstring part;
 329                                 docstring const rest = parseEmbeddedOption(fmt, part);
 330                                 if (fmt == rest) {
 331                                         LYXERR0("ERROR! Couldn't parse `" << format <<"'.");
 332                                         return format;
 333                                 }
 334                                 clause += part;
 335                                 fmt = rest;
 336                         } else { // it's just a normal character
 337                                 clause += fmt[0];
 338                                 fmt = fmt.substr(1);
 339                         }
 340                 }
 341                 return fmt;
 342         }
 343
 344
 345         /// parse an options string, which must appear at the start of the
 346         /// format parameter. puts the parsed bits in optkey, ifpart, and
 347         /// elsepart and returns what's left after the option is removed.
 348         /// if there's an error, it returns format itself.
 349         docstring parseOptions(docstring const & format, docstring & optkey,
 350                         docstring & ifpart, docstring & elsepart)
 351         {
 352                 LASSERT(format[0] == '{' && format[1] == '%', return format);
 353                 // strip '{%'
 354                 docstring fmt = format.substr(2);
 355                 size_t pos = fmt.find('%'); // end of key
 356                 if (pos == string::npos) {
 357                         LYXERR0("Error parsing  `" << format <<"'. Can't find end of key.");
 358                         return format;
 359                 }
 360                 optkey = fmt.substr(0,pos);
 361                 fmt = fmt.substr(pos + 1);
 362                 // [[format]] should be next
 363                 if (fmt[0] != '[' || fmt[1] != '[') {
 364                         LYXERR0("Error parsing  `" << format <<"'. Can't find '[[' after key.");
 365                         return format;
 366                 }
 367
 368                 docstring curfmt = fmt;
 369                 fmt = getClause(curfmt, ifpart);
 370                 if (fmt == curfmt) {
 371                         LYXERR0("Error parsing  `" << format <<"'. Couldn't get if clause.");
 372                         return format;
 373                 }
 374
 375                 if (fmt[0] == '}') // we're done, no else clause
 376                         return fmt.substr(1);
 377
 378                 // else part should follow
 379                 if (fmt[0] != '[' || fmt[1] != '[') {
 380                         LYXERR0("Error parsing  `" << format <<"'. Can't find else clause.");
 381                         return format;
 382                 }
 383
 384                 curfmt = fmt;
 385                 fmt = getClause(curfmt, elsepart);
 386                 // we should be done
 387                 if (fmt == curfmt || fmt[0] != '}') {
 388                         LYXERR0("Error parsing  `" << format <<"'. Can't find end of option.");
 389                         return format;
 390                 }
 391                 return fmt.substr(1);
 392 }
 393
 394 } // anon namespace
 395
 396
 397 docstring BibTeXInfo::expandFormat(docstring const & format,
 398                 BibTeXInfo const * const xref, bool richtext) const
 399 {
 400         // return value
 401         docstring ret;
 402         docstring key;
 403         bool scanning_key = false;
 404         bool scanning_rich = false;
 405
 406         docstring fmt = format;
 407         // we'll remove characters from the front of fmt as we
 408         // deal with them
 409         while (fmt.size()) {
 410                 char_type thischar = fmt[0];
 411                 if (thischar == '%') {
 412                         // beginning or end of key
 413                         if (scanning_key) {
 414                                 // end of key
 415                                 scanning_key = false;
 416                                 // so we replace the key with its value, which may be empty
 417                                 docstring const val = getValueForKey(to_utf8(key), xref);
 418                                 key.clear();
 419                                 ret += val;
 420                         } else {
 421                                 // beginning of key
 422                                 scanning_key = true;
 423                         }
 424                 }
 425                 else if (thischar == '{') {
 426                         // beginning of option?
 427                         if (scanning_key) {
 428                                 LYXERR0("ERROR: Found `{' when scanning key in `" << format << "'.");
 429                                 return _("ERROR!");
 430                         }
 431                         if (fmt.size() > 1) {
 432                                 if (fmt[1] == '%') {
 433                                         // it is the beginning of an optional format
 434                                         docstring optkey;
 435                                         docstring ifpart;
 436                                         docstring elsepart;
 437                                         docstring const newfmt =
 438                                                 parseOptions(fmt, optkey, ifpart, elsepart);
 439                                         if (newfmt == fmt) // parse error
 440                                                 return _("ERROR!");
 441                                         fmt = newfmt;
 442                                         docstring const val = getValueForKey(to_utf8(optkey), xref);
 443                                         if (!val.empty())
 444                                                 ret += expandFormat(ifpart, xref, richtext);
 445                                         else if (!elsepart.empty())
 446                                                 ret += expandFormat(elsepart, xref, richtext);
 447                                         // fmt will have been shortened for us already
 448                                         continue;
 449                                 }
 450                                 if (fmt[1] == '!') {
 451                                         // beginning of rich text
 452                                         scanning_rich = true;
 453                                         fmt = fmt.substr(2);
 454                                         continue;
 455                                 }
 456                         }
 457                         // we are here if the '{' was at the end of the format. hmm.
 458                         ret += thischar;
 459                 }
 460                 else if (scanning_rich && thischar == '!'
 461                          && fmt.size() > 1 && fmt[1] == '}') {
 462                         // end of rich text
 463                         scanning_rich = false;
 464                         fmt = fmt.substr(2);
 465                         continue;
 466                 }
 467                 else if (scanning_key)
 468                         key += thischar;
 469                 else if (richtext || !scanning_rich)
 470                         ret += thischar;
 471                 // else the character is discarded, which will happen only if
 472                 // richtext == false and we are scanning rich text
 473                 fmt = fmt.substr(1);
 474         } // for loop
 475         if (scanning_key) {
 476                 LYXERR0("Never found end of key in `" << format << "'!");
 477                 return _("ERROR!");
 478         }
 479         if (scanning_rich) {
 480                 LYXERR0("Never found end of rich text in `" << format << "'!");
 481                 return _("ERROR!");
 482         }
 483         return ret;
 484 }
 485
 486
 487 namespace {
 488
 489 // FIXME These would be better read from a file, so that they
 490 // could be customized.
 491
 492         static docstring articleFormat = from_ascii("%author%, \"%title%\", {!<i>!}%journal%{!</i>!} {%volume%[[ %volume%{%number%[[, %number%]]}]]} (%year%){%pages%[[, pp. %pages%]]}.{%note%[[ %note%]]}");
 493
 494         static docstring bookFormat = from_ascii("{%author%[[%author%]][[%editor%, ed.]]}, {!<i>!}%title%{!</i>!}{%volume%[[ vol. %volume%]][[{%number%[[no. %number%]]}]]}{%edition%[[%edition%]]} ({%address%[[%address%: ]]}%publisher%, %year%).{%note%[[ %note%]]}");
 495
 496 static docstring inSomething = from_ascii("%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%]]}");
 497
 498 static docstring thesis = from_ascii("%author%, %title% ({%address%[[%address%: ]]}%school%, %year%).{%note%[[ %note%]]}");
 499
 500 static docstring defaultFormat = from_ascii("{%author%[[%author%, ]][[{%editor%[[%editor%, ed., ]]}]]}\"%title%\"{%journal%[[, {!<i>!}%journal%{!</i>!}]][[{%publisher%[[, %publisher%]][[{%institution%[[, %institution%]]}]]}]]}{%year%[[ (%year%)]]}{%pages%[[, %pages%]]}.");
 501
 502 }
 503
 504 docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref,
 505         bool richtext) const
 506 {
 507         if (!info_.empty())
 508                 return info_;
 509
 510         if (!is_bibtex_) {
 511                 BibTeXInfo::const_iterator it = find(from_ascii("ref"));
 512                 info_ = it->second;
 513                 return info_;
 514         }
 515
 516         if (entry_type_ == "article")
 517                 info_ = expandFormat(articleFormat, xref, richtext);
 518         else if (entry_type_ == "book")
 519                 info_ = expandFormat(bookFormat, xref, richtext);
 520         else if (entry_type_.substr(0,2) == "in")
 521                 info_ = expandFormat(inSomething, xref, richtext);
 522         else if (entry_type_ == "phdthesis" || entry_type_ == "mastersthesis")
 523                 info_ = expandFormat(thesis, xref, richtext);
 524         else
 525                 info_ = expandFormat(defaultFormat, xref, richtext);
 526
 527         if (!info_.empty())
 528                 info_ = convertLaTeXCommands(info_);
 529         return info_;
 530 }
 531
 532
 533 docstring const & BibTeXInfo::operator[](docstring const & field) const
 534 {
 535         BibTeXInfo::const_iterator it = find(field);
 536         if (it != end())
 537                 return it->second;
 538         static docstring const empty_value = docstring();
 539         return empty_value;
 540 }
 541
 542
 543 docstring const & BibTeXInfo::operator[](string const & field) const
 544 {
 545         return operator[](from_ascii(field));
 546 }
 547
 548
 549 docstring BibTeXInfo::getValueForKey(string const & key,
 550                 BibTeXInfo const * const xref) const
 551 {
 552         docstring const ret = operator[](key);
 553         if (!ret.empty() || !xref)
 554                 return ret;
 555         return (*xref)[key];
 556 }
 557
 558
 559 //////////////////////////////////////////////////////////////////////
 560 //
 561 // BiblioInfo
 562 //
 563 //////////////////////////////////////////////////////////////////////
 564
 565 namespace {
 566 // A functor for use with sort, leading to case insensitive sorting
 567         class compareNoCase: public binary_function<docstring, docstring, bool>
 568         {
 569                 public:
 570                         bool operator()(docstring const & s1, docstring const & s2) const {
 571                                 return compare_no_case(s1, s2) < 0;
 572                         }
 573         };
 574 } // namespace anon
 575
 576
 577 vector<docstring> const BiblioInfo::getKeys() const
 578 {
 579         vector<docstring> bibkeys;
 580         BiblioInfo::const_iterator it  = begin();
 581         for (; it != end(); ++it)
 582                 bibkeys.push_back(it->first);
 583         sort(bibkeys.begin(), bibkeys.end(), compareNoCase());
 584         return bibkeys;
 585 }
 586
 587
 588 vector<docstring> const BiblioInfo::getFields() const
 589 {
 590         vector<docstring> bibfields;
 591         set<docstring>::const_iterator it = field_names_.begin();
 592         set<docstring>::const_iterator end = field_names_.end();
 593         for (; it != end; ++it)
 594                 bibfields.push_back(*it);
 595         sort(bibfields.begin(), bibfields.end());
 596         return bibfields;
 597 }
 598
 599
 600 vector<docstring> const BiblioInfo::getEntries() const
 601 {
 602         vector<docstring> bibentries;
 603         set<docstring>::const_iterator it = entry_types_.begin();
 604         set<docstring>::const_iterator end = entry_types_.end();
 605         for (; it != end; ++it)
 606                 bibentries.push_back(*it);
 607         sort(bibentries.begin(), bibentries.end());
 608         return bibentries;
 609 }
 610
 611
 612 docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key) const
 613 {
 614         BiblioInfo::const_iterator it = find(key);
 615         if (it == end())
 616                 return docstring();
 617         BibTeXInfo const & data = it->second;
 618         return data.getAbbreviatedAuthor();
 619 }
 620
 621
 622 docstring const BiblioInfo::getCiteNumber(docstring const & key) const
 623 {
 624         BiblioInfo::const_iterator it = find(key);
 625         if (it == end())
 626                 return docstring();
 627         BibTeXInfo const & data = it->second;
 628         return data.citeNumber();
 629 }
 630
 631
 632 docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier) const
 633 {
 634         BiblioInfo::const_iterator it = find(key);
 635         if (it == end())
 636                 return docstring();
 637         BibTeXInfo const & data = it->second;
 638         docstring year = data.getYear();
 639         if (year.empty()) {
 640                 // let's try the crossref
 641                 docstring const xref = data.getXRef();
 642                 if (xref.empty())
 643                         return _("No year"); // no luck
 644                 BiblioInfo::const_iterator const xrefit = find(xref);
 645                 if (xrefit == end())
 646                         return _("No year"); // no luck again
 647                 BibTeXInfo const & xref_data = xrefit->second;
 648                 year = xref_data.getYear();
 649         }
 650         if (use_modifier && data.modifier() != 0)
 651                 year += data.modifier();
 652         return year;
 653 }
 654
 655
 656 docstring const BiblioInfo::getInfo(docstring const & key, bool richtext) const
 657 {
 658         BiblioInfo::const_iterator it = find(key);
 659         if (it == end())
 660                 return docstring();
 661         BibTeXInfo const & data = it->second;
 662         BibTeXInfo const * xrefptr = 0;
 663         docstring const xref = data.getXRef();
 664         if (!xref.empty()) {
 665                 BiblioInfo::const_iterator const xrefit = find(xref);
 666                 if (xrefit != end())
 667                         xrefptr = &(xrefit->second);
 668         }
 669         return data.getInfo(xrefptr, richtext);
 670 }
 671
 672
 673 bool BiblioInfo::isBibtex(docstring const & key) const
 674 {
 675         BiblioInfo::const_iterator it = find(key);
 676         if (it == end())
 677                 return false;
 678         return it->second.isBibTeX();
 679 }
 680
 681
 682
 683 vector<docstring> const BiblioInfo::getCiteStrings(
 684         docstring const & key, Buffer const & buf) const
 685 {
 686         CiteEngine const engine = buf.params().citeEngine();
 687         if (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL)
 688                 return getNumericalStrings(key, buf);
 689         else
 690                 return getAuthorYearStrings(key, buf);
 691 }
 692
 693
 694 vector<docstring> const BiblioInfo::getNumericalStrings(
 695         docstring const & key, Buffer const & buf) const
 696 {
 697         if (empty())
 698                 return vector<docstring>();
 699
 700         docstring const author = getAbbreviatedAuthor(key);
 701         docstring const year   = getYear(key);
 702         if (author.empty() || year.empty())
 703                 return vector<docstring>();
 704
 705         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 706
 707         vector<docstring> vec(styles.size());
 708         for (size_t i = 0; i != vec.size(); ++i) {
 709                 docstring str;
 710
 711                 switch (styles[i]) {
 712                         case CITE:
 713                         case CITEP:
 714                                 str = from_ascii("[#ID]");
 715                                 break;
 716
 717                         case NOCITE:
 718                                 str = _("Add to bibliography only.");
 719                                 break;
 720
 721                         case CITET:
 722                                 str = author + " [#ID]";
 723                                 break;
 724
 725                         case CITEALT:
 726                                 str = author + " #ID";
 727                                 break;
 728
 729                         case CITEALP:
 730                                 str = from_ascii("#ID");
 731                                 break;
 732
 733                         case CITEAUTHOR:
 734                                 str = author;
 735                                 break;
 736
 737                         case CITEYEAR:
 738                                 str = year;
 739                                 break;
 740
 741                         case CITEYEARPAR:
 742                                 str = '(' + year + ')';
 743                                 break;
 744                 }
 745
 746                 vec[i] = str;
 747         }
 748
 749         return vec;
 750 }
 751
 752
 753 vector<docstring> const BiblioInfo::getAuthorYearStrings(
 754         docstring const & key, Buffer const & buf) const
 755 {
 756         if (empty())
 757                 return vector<docstring>();
 758
 759         docstring const author = getAbbreviatedAuthor(key);
 760         docstring const year   = getYear(key);
 761         if (author.empty() || year.empty())
 762                 return vector<docstring>();
 763
 764         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 765
 766         vector<docstring> vec(styles.size());
 767         for (size_t i = 0; i != vec.size(); ++i) {
 768                 docstring str;
 769
 770                 switch (styles[i]) {
 771                         case CITE:
 772                 // jurabib only: Author/Annotator
 773                 // (i.e. the "before" field, 2nd opt arg)
 774                                 str = author + "/<" + _("before") + '>';
 775                                 break;
 776
 777                         case NOCITE:
 778                                 str = _("Add to bibliography only.");
 779                                 break;
 780
 781                         case CITET:
 782                                 str = author + " (" + year + ')';
 783                                 break;
 784
 785                         case CITEP:
 786                                 str = '(' + author + ", " + year + ')';
 787                                 break;
 788
 789                         case CITEALT:
 790                                 str = author + ' ' + year ;
 791                                 break;
 792
 793                         case CITEALP:
 794                                 str = author + ", " + year ;
 795                                 break;
 796
 797                         case CITEAUTHOR:
 798                                 str = author;
 799                                 break;
 800
 801                         case CITEYEAR:
 802                                 str = year;
 803                                 break;
 804
 805                         case CITEYEARPAR:
 806                                 str = '(' + year + ')';
 807                                 break;
 808                 }
 809                 vec[i] = str;
 810         }
 811         return vec;
 812 }
 813
 814
 815 void BiblioInfo::mergeBiblioInfo(BiblioInfo const & info)
 816 {
 817         bimap_.insert(info.begin(), info.end());
 818 }
 819
 820
 821 namespace {
 822         // used in xhtml to sort a list of BibTeXInfo objects
 823         bool lSorter(BibTeXInfo const * lhs, BibTeXInfo const * rhs)
 824         {
 825                 docstring const lauth = lhs->getAbbreviatedAuthor();
 826                 docstring const rauth = rhs->getAbbreviatedAuthor();
 827                 docstring const lyear = lhs->getYear();
 828                 docstring const ryear = rhs->getYear();
 829                 docstring const ltitl = lhs->operator[]("title");
 830                 docstring const rtitl = rhs->operator[]("title");
 831                 return  (lauth < rauth)
 832                                 || (lauth == rauth && lyear < ryear)
 833                                 || (lauth == rauth && lyear == ryear && ltitl < rtitl);
 834         }
 835 }
 836
 837
 838 void BiblioInfo::collectCitedEntries(Buffer const & buf)
 839 {
 840         cited_entries_.clear();
 841         // We are going to collect all the citation keys used in the document,
 842         // getting them from the TOC.
 843         // FIXME We may want to collect these differently, in the first case,
 844         // so that we might have them in order of appearance.
 845         set<docstring> citekeys;
 846         Toc const & toc = buf.tocBackend().toc("citation");
 847         Toc::const_iterator it = toc.begin();
 848         Toc::const_iterator const en = toc.end();
 849         for (; it != en; ++it) {
 850                 if (it->str().empty())
 851                         continue;
 852                 vector<docstring> const keys = getVectorFromString(it->str());
 853                 citekeys.insert(keys.begin(), keys.end());
 854         }
 855         if (citekeys.empty())
 856                 return;
 857
 858         // We have a set of the keys used in this document.
 859         // We will now convert it to a list of the BibTeXInfo objects used in
 860         // this document...
 861         vector<BibTeXInfo const *> bi;
 862         set<docstring>::const_iterator cit = citekeys.begin();
 863         set<docstring>::const_iterator const cen = citekeys.end();
 864         for (; cit != cen; ++cit) {
 865                 BiblioInfo::const_iterator const bt = find(*cit);
 866                 if (bt == end() || !bt->second.isBibTeX())
 867                         continue;
 868                 bi.push_back(&(bt->second));
 869         }
 870         // ...and sort it.
 871         sort(bi.begin(), bi.end(), lSorter);
 872
 873         // Now we can write the sorted keys
 874         vector<BibTeXInfo const *>::const_iterator bit = bi.begin();
 875         vector<BibTeXInfo const *>::const_iterator ben = bi.end();
 876         for (; bit != ben; ++bit)
 877                 cited_entries_.push_back((*bit)->key());
 878 }
 879
 880
 881 void BiblioInfo::makeCitationLabels(Buffer const & buf)
 882 {
 883         collectCitedEntries(buf);
 884         CiteEngine const engine = buf.params().citeEngine();
 885         bool const numbers =
 886                 (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL);
 887
 888         int keynumber = 0;
 889         char modifier = 0;
 890         // used to remember the last one we saw
 891         // we'll be comparing entries to see if we need to add
 892         // modifiers, like "1984a"
 893         map<docstring, BibTeXInfo>::iterator last;
 894
 895         vector<docstring>::const_iterator it = cited_entries_.begin();
 896         vector<docstring>::const_iterator const en = cited_entries_.end();
 897         for (; it != en; ++it) {
 898                 map<docstring, BibTeXInfo>::iterator const biit = bimap_.find(*it);
 899                 // this shouldn't happen, but...
 900                 if (biit == bimap_.end())
 901                         // ...fail gracefully, anyway.
 902                         continue;
 903                 BibTeXInfo & entry = biit->second;
 904                 if (numbers) {
 905                         docstring const num = convert<docstring>(++keynumber);
 906                         entry.setCiteNumber(num);
 907                 } else {
 908                         if (it != cited_entries_.begin()
 909                             && entry.getAbbreviatedAuthor() == last->second.getAbbreviatedAuthor()
 910                             // we access the year via getYear() so as to get it from the xref,
 911                             // if we need to do so
 912                             && getYear(entry.key()) == getYear(last->second.key())) {
 913                                 if (modifier == 0) {
 914                                         // so the last one should have been 'a'
 915                                         last->second.setModifier('a');
 916                                         modifier = 'b';
 917                                 } else if (modifier == 'z')
 918                                         modifier = 'A';
 919                                 else
 920                                         modifier++;
 921                         } else {
 922                                 modifier = 0;
 923                         }
 924                         entry.setModifier(modifier);
 925                         // remember the last one
 926                         last = biit;
 927                 }
 928         }
 929 }
 930
 931
 932 //////////////////////////////////////////////////////////////////////
 933 //
 934 // CitationStyle
 935 //
 936 //////////////////////////////////////////////////////////////////////
 937
 938 namespace {
 939
 940
 941 char const * const citeCommands[] = {
 942         "cite", "citet", "citep", "citealt", "citealp",
 943         "citeauthor", "citeyear", "citeyearpar", "nocite" };
 944
 945 unsigned int const nCiteCommands =
 946                 sizeof(citeCommands) / sizeof(char *);
 947
 948 CiteStyle const citeStylesArray[] = {
 949         CITE, CITET, CITEP, CITEALT, CITEALP,
 950         CITEAUTHOR, CITEYEAR, CITEYEARPAR, NOCITE };
 951
 952 unsigned int const nCiteStyles =
 953                 sizeof(citeStylesArray) / sizeof(CiteStyle);
 954
 955 CiteStyle const citeStylesFull[] = {
 956         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 957
 958 unsigned int const nCiteStylesFull =
 959                 sizeof(citeStylesFull) / sizeof(CiteStyle);
 960
 961 CiteStyle const citeStylesUCase[] = {
 962         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 963
 964 unsigned int const nCiteStylesUCase =
 965         sizeof(citeStylesUCase) / sizeof(CiteStyle);
 966
 967 } // namespace anon
 968
 969
 970 CitationStyle citationStyleFromString(string const & command)
 971 {
 972         CitationStyle s;
 973         if (command.empty())
 974                 return s;
 975
 976         string cmd = command;
 977         if (cmd[0] == 'C') {
 978                 s.forceUpperCase = true;
 979                 cmd[0] = 'c';
 980         }
 981
 982         size_t const n = cmd.size() - 1;
 983         if (cmd != "cite" && cmd[n] == '*') {
 984                 s.full = true;
 985                 cmd = cmd.substr(0, n);
 986         }
 987
 988         char const * const * const last = citeCommands + nCiteCommands;
 989         char const * const * const ptr = find(citeCommands, last, cmd);
 990
 991         if (ptr != last) {
 992                 size_t idx = ptr - citeCommands;
 993                 s.style = citeStylesArray[idx];
 994         }
 995         return s;
 996 }
 997
 998
 999 string citationStyleToString(const CitationStyle & s)
1000 {
1001         string cite = citeCommands[s.style];
1002         if (s.full) {
1003                 CiteStyle const * last = citeStylesFull + nCiteStylesFull;
1004                 if (std::find(citeStylesFull, last, s.style) != last)
1005                         cite += '*';
1006         }
1007
1008         if (s.forceUpperCase) {
1009                 CiteStyle const * last = citeStylesUCase + nCiteStylesUCase;
1010                 if (std::find(citeStylesUCase, last, s.style) != last)
1011                         cite[0] = 'C';
1012         }
1013
1014         return cite;
1015 }
1016
1017 vector<CiteStyle> citeStyles(CiteEngine engine)
1018 {
1019         unsigned int nStyles = 0;
1020         unsigned int start = 0;
1021
1022         switch (engine) {
1023                 case ENGINE_BASIC:
1024                         nStyles = 2;
1025                         start = 0;
1026                         break;
1027                 case ENGINE_NATBIB_AUTHORYEAR:
1028                 case ENGINE_NATBIB_NUMERICAL:
1029                         nStyles = nCiteStyles - 1;
1030                         start = 1;
1031                         break;
1032                 case ENGINE_JURABIB:
1033                         nStyles = nCiteStyles;
1034                         start = 0;
1035                         break;
1036         }
1037
1038         vector<CiteStyle> styles(nStyles);
1039         size_t i = 0;
1040         int j = start;
1041         for (; i != styles.size(); ++i, ++j)
1042                 styles[i] = citeStylesArray[j];
1043
1044         return styles;
1045 }
1046
1047 } // namespace lyx
1048