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 "Language.h"
  22 #include "Paragraph.h"
  23 #include "TextClass.h"
  24 #include "TocBackend.h"
  25
  26 #include "support/convert.h"
  27 #include "support/debug.h"
  28 #include "support/docstream.h"
  29 #include "support/gettext.h"
  30 #include "support/lassert.h"
  31 #include "support/lstrings.h"
  32 #include "support/regex.h"
  33 #include "support/textutils.h"
  34
  35 #include <set>
  36
  37 using namespace std;
  38 using namespace lyx::support;
  39
  40
  41 namespace lyx {
  42
  43 namespace {
  44
  45 // gets the "family name" from an author-type string
  46 docstring familyName(docstring const & name)
  47 {
  48         if (name.empty())
  49                 return docstring();
  50
  51         // first we look for a comma, and take the last name to be everything
  52         // preceding the right-most one, so that we also get the "jr" part.
  53         docstring::size_type idx = name.rfind(',');
  54         if (idx != docstring::npos)
  55                 return ltrim(name.substr(0, idx));
  56
  57         // OK, so now we want to look for the last name. We're going to
  58         // include the "von" part. This isn't perfect.
  59         // Split on spaces, to get various tokens.
  60         vector<docstring> pieces = getVectorFromString(name, from_ascii(" "));
  61         // If we only get two, assume the last one is the last name
  62         if (pieces.size() <= 2)
  63                 return pieces.back();
  64
  65         // Now we look for the first token that begins with a lower case letter.
  66         vector<docstring>::const_iterator it = pieces.begin();
  67         vector<docstring>::const_iterator en = pieces.end();
  68         for (; it != en; ++it) {
  69                 if ((*it).size() == 0)
  70                         continue;
  71                 char_type const c = (*it)[0];
  72                 if (isLower(c))
  73                         break;
  74         }
  75
  76         if (it == en) // we never found a "von"
  77                 return pieces.back();
  78
  79         // reconstruct what we need to return
  80         docstring retval;
  81         bool first = true;
  82         for (; it != en; ++it) {
  83                 if (!first)
  84                         retval += " ";
  85                 else
  86                         first = false;
  87                 retval += *it;
  88         }
  89         return retval;
  90 }
  91
  92 // converts a string containing LaTeX commands into unicode
  93 // for display.
  94 docstring convertLaTeXCommands(docstring const & str)
  95 {
  96         docstring val = str;
  97         docstring ret;
  98
  99         bool scanning_cmd = false;
 100         bool scanning_math = false;
 101         bool escaped = false; // used to catch \$, etc.
 102         while (val.size()) {
 103                 char_type const ch = val[0];
 104
 105                 // if we're scanning math, we output everything until we
 106                 // find an unescaped $, at which point we break out.
 107                 if (scanning_math) {
 108                         if (escaped)
 109                                 escaped = false;
 110                         else if (ch == '\\')
 111                                 escaped = true;
 112                         else if (ch == '$')
 113                                 scanning_math = false;
 114                         ret += ch;
 115                         val = val.substr(1);
 116                         continue;
 117                 }
 118
 119                 // if we're scanning a command name, then we just
 120                 // discard characters until we hit something that
 121                 // isn't alpha.
 122                 if (scanning_cmd) {
 123                         if (isAlphaASCII(ch)) {
 124                                 val = val.substr(1);
 125                                 escaped = false;
 126                                 continue;
 127                         }
 128                         // so we're done with this command.
 129                         // now we fall through and check this character.
 130                         scanning_cmd = false;
 131                 }
 132
 133                 // was the last character a \? If so, then this is something like:
 134                 // \\ or \$, so we'll just output it. That's probably not always right...
 135                 if (escaped) {
 136                         // exception: output \, as THIN SPACE
 137                         if (ch == ',')
 138                                 ret.push_back(0x2009);
 139                         else
 140                                 ret += ch;
 141                         val = val.substr(1);
 142                         escaped = false;
 143                         continue;
 144                 }
 145
 146                 if (ch == '$') {
 147                         ret += ch;
 148                         val = val.substr(1);
 149                         scanning_math = true;
 150                         continue;
 151                 }
 152
 153                 // we just ignore braces
 154                 if (ch == '{' || ch == '}') {
 155                         val = val.substr(1);
 156                         continue;
 157                 }
 158
 159                 // we're going to check things that look like commands, so if
 160                 // this doesn't, just output it.
 161                 if (ch != '\\') {
 162                         ret += ch;
 163                         val = val.substr(1);
 164                         continue;
 165                 }
 166
 167                 // ok, could be a command of some sort
 168                 // let's see if it corresponds to some unicode
 169                 // unicodesymbols has things in the form: \"{u},
 170                 // whereas we may see things like: \"u. So we'll
 171                 // look for that and change it, if necessary.
 172                 static lyx::regex const reg("^\\\\\\W\\w");
 173                 if (lyx::regex_search(to_utf8(val), reg)) {
 174                         val.insert(3, from_ascii("}"));
 175                         val.insert(2, from_ascii("{"));
 176                 }
 177                 docstring rem;
 178                 docstring const cnvtd = Encodings::fromLaTeXCommand(val, rem,
 179                                                         Encodings::TEXT_CMD);
 180                 if (!cnvtd.empty()) {
 181                         // it did, so we'll take that bit and proceed with what's left
 182                         ret += cnvtd;
 183                         val = rem;
 184                         continue;
 185                 }
 186                 // it's a command of some sort
 187                 scanning_cmd = true;
 188                 escaped = true;
 189                 val = val.substr(1);
 190         }
 191         return ret;
 192 }
 193
 194 } // anon namespace
 195
 196
 197 //////////////////////////////////////////////////////////////////////
 198 //
 199 // BibTeXInfo
 200 //
 201 //////////////////////////////////////////////////////////////////////
 202
 203 BibTeXInfo::BibTeXInfo(docstring const & key, docstring const & type)
 204         : is_bibtex_(true), bib_key_(key), entry_type_(type), info_(),
 205           modifier_(0)
 206 {}
 207
 208
 209 docstring const BibTeXInfo::getAbbreviatedAuthor() const
 210 {
 211         if (!is_bibtex_) {
 212                 docstring const opt = label();
 213                 if (opt.empty())
 214                         return docstring();
 215
 216                 docstring authors;
 217                 docstring const remainder = trim(split(opt, authors, '('));
 218                 if (remainder.empty())
 219                         // in this case, we didn't find a "(",
 220                         // so we don't have author (year)
 221                         return docstring();
 222                 return authors;
 223         }
 224
 225         docstring author = convertLaTeXCommands(operator[]("author"));
 226         if (author.empty()) {
 227                 author = convertLaTeXCommands(operator[]("editor"));
 228                 if (author.empty())
 229                         return bib_key_;
 230         }
 231
 232         // FIXME Move this to a separate routine that can
 233         // be called from elsewhere.
 234         //
 235         // OK, we've got some names. Let's format them.
 236         // Try to split the author list on " and "
 237         vector<docstring> const authors =
 238                 getVectorFromString(author, from_ascii(" and "));
 239
 240         if (authors.size() == 2)
 241                 return bformat(_("%1$s and %2$s"),
 242                         familyName(authors[0]), familyName(authors[1]));
 243
 244         if (authors.size() > 2)
 245                 return bformat(_("%1$s et al."), familyName(authors[0]));
 246
 247         return familyName(authors[0]);
 248 }
 249
 250
 251 docstring const BibTeXInfo::getYear() const
 252 {
 253         if (is_bibtex_)
 254                 return operator[]("year");
 255
 256         docstring const opt = label();
 257         if (opt.empty())
 258                 return docstring();
 259
 260         docstring authors;
 261         docstring tmp = split(opt, authors, '(');
 262         if (tmp.empty())
 263                 // we don't have author (year)
 264                 return docstring();
 265         docstring year;
 266         tmp = split(tmp, year, ')');
 267         return year;
 268 }
 269
 270
 271 docstring const BibTeXInfo::getXRef() const
 272 {
 273         if (!is_bibtex_)
 274                 return docstring();
 275         return operator[]("crossref");
 276 }
 277
 278
 279 namespace {
 280         string parseOptions(string const & format, string & optkey,
 281                         string & ifpart, string & elsepart);
 282
 283         // Calls parseOptions to deal with an embedded option, such as:
 284         //   {%number%[[, no.~%number%]]}
 285         // which must appear at the start of format. ifelsepart gets the
 286         // whole of the option, and we return what's left after the option.
 287         // we return format if there is an error.
 288         string parseEmbeddedOption(string const & format, string & ifelsepart)
 289         {
 290                 LASSERT(format[0] == '{' && format[1] == '%', return format);
 291                 string optkey;
 292                 string ifpart;
 293                 string elsepart;
 294                 string const rest = parseOptions(format, optkey, ifpart, elsepart);
 295                 if (format == rest) { // parse error
 296                         LYXERR0("ERROR! Couldn't parse `" << format <<"'.");
 297                         return format;
 298                 }
 299                 LASSERT(rest.size() <= format.size(), /* */);
 300                 ifelsepart = format.substr(0, format.size() - rest.size());
 301                 return rest;
 302         }
 303
 304
 305         // Gets a "clause" from a format string, where the clause is
 306         // delimited by '[[' and ']]'. Returns what is left after the
 307         // clause is removed, and returns format if there is an error.
 308         string getClause(string const & format, string & clause)
 309         {
 310                 string fmt = format;
 311                 // remove '[['
 312                 fmt = fmt.substr(2);
 313                 // we'll remove characters from the front of fmt as we
 314                 // deal with them
 315                 while (fmt.size()) {
 316                         if (fmt[0] == ']' && fmt.size() > 1 && fmt[1] == ']') {
 317                                 // that's the end
 318                                 fmt = fmt.substr(2);
 319                                 break;
 320                         }
 321                         // check for an embedded option
 322                         if (fmt[0] == '{' && fmt.size() > 1 && fmt[1] == '%') {
 323                                 string part;
 324                                 string const rest = parseEmbeddedOption(fmt, part);
 325                                 if (fmt == rest) {
 326                                         LYXERR0("ERROR! Couldn't parse embedded option in `" << format <<"'.");
 327                                         return format;
 328                                 }
 329                                 clause += part;
 330                                 fmt = rest;
 331                         } else { // it's just a normal character
 332                                 clause += fmt[0];
 333                                 fmt = fmt.substr(1);
 334                         }
 335                 }
 336                 return fmt;
 337         }
 338
 339
 340         // parse an options string, which must appear at the start of the
 341         // format parameter. puts the parsed bits in optkey, ifpart, and
 342         // elsepart and returns what's left after the option is removed.
 343         // if there's an error, it returns format itself.
 344         string parseOptions(string const & format, string & optkey,
 345                         string & ifpart, string & elsepart)
 346         {
 347                 LASSERT(format[0] == '{' && format[1] == '%', return format);
 348                 // strip '{%'
 349                 string fmt = format.substr(2);
 350                 size_t pos = fmt.find('%'); // end of key
 351                 if (pos == string::npos) {
 352                         LYXERR0("Error parsing  `" << format <<"'. Can't find end of key.");
 353                         return format;
 354                 }
 355                 optkey = fmt.substr(0,pos);
 356                 fmt = fmt.substr(pos + 1);
 357                 // [[format]] should be next
 358                 if (fmt[0] != '[' || fmt[1] != '[') {
 359                         LYXERR0("Error parsing  `" << format <<"'. Can't find '[[' after key.");
 360                         return format;
 361                 }
 362
 363                 string curfmt = fmt;
 364                 fmt = getClause(curfmt, ifpart);
 365                 if (fmt == curfmt) {
 366                         LYXERR0("Error parsing  `" << format <<"'. Couldn't get if clause.");
 367                         return format;
 368                 }
 369
 370                 if (fmt[0] == '}') // we're done, no else clause
 371                         return fmt.substr(1);
 372
 373                 // else part should follow
 374                 if (fmt[0] != '[' || fmt[1] != '[') {
 375                         LYXERR0("Error parsing  `" << format <<"'. Can't find else clause.");
 376                         return format;
 377                 }
 378
 379                 curfmt = fmt;
 380                 fmt = getClause(curfmt, elsepart);
 381                 // we should be done
 382                 if (fmt == curfmt || fmt[0] != '}') {
 383                         LYXERR0("Error parsing  `" << format <<"'. Can't find end of option.");
 384                         return format;
 385                 }
 386                 return fmt.substr(1);
 387 }
 388
 389 } // anon namespace
 390
 391
 392 docstring BibTeXInfo::expandFormat(string const & format,
 393                 BibTeXInfo const * const xref, int & counter, Buffer const & buf,
 394                 bool richtext) const
 395 {
 396         // incorrect use of macros could put us in an infinite loop
 397         static int max_passes = 5000;
 398         docstring ret; // return value
 399         string key;
 400         bool scanning_key = false;
 401         bool scanning_rich = false;
 402
 403         string fmt = format;
 404         // we'll remove characters from the front of fmt as we
 405         // deal with them
 406         while (fmt.size()) {
 407                 if (counter++ > max_passes) {
 408                         LYXERR0("Recursion limit reached while parsing `"
 409                                 << format << "'.");
 410                         return _("ERROR!");
 411                 }
 412
 413                 char_type thischar = fmt[0];
 414                 if (thischar == '%') {
 415                         // beginning or end of key
 416                         if (scanning_key) {
 417                                 // end of key
 418                                 scanning_key = false;
 419                                 // so we replace the key with its value, which may be empty
 420                                 if (key[0] == '!') {
 421                                         // macro
 422                                         string const val =
 423                                                 buf.params().documentClass().getCiteMacro(key);
 424                                         fmt = val + fmt.substr(1);
 425                                         continue;
 426                                 } else if (key[0] == '_') {
 427                                         // a translatable bit
 428                                         string const val =
 429                                                 buf.params().documentClass().getCiteMacro(key);
 430                                         docstring const trans =
 431                                                 translateIfPossible(from_utf8(val), buf.params().language->code());
 432                                         ret += trans;
 433                                 } else {
 434                                         docstring const val = getValueForKey(key, xref);
 435                                         ret += val;
 436                                 }
 437                         } else {
 438                                 // beginning of key
 439                                 key.clear();
 440                                 scanning_key = true;
 441                         }
 442                 }
 443                 else if (thischar == '{') {
 444                         // beginning of option?
 445                         if (scanning_key) {
 446                                 LYXERR0("ERROR: Found `{' when scanning key in `" << format << "'.");
 447                                 return _("ERROR!");
 448                         }
 449                         if (fmt.size() > 1) {
 450                                 if (fmt[1] == '%') {
 451                                         // it is the beginning of an optional format
 452                                         string optkey;
 453                                         string ifpart;
 454                                         string elsepart;
 455                                         string const newfmt =
 456                                                 parseOptions(fmt, optkey, ifpart, elsepart);
 457                                         if (newfmt == fmt) // parse error
 458                                                 return _("ERROR!");
 459                                         fmt = newfmt;
 460                                         docstring const val = getValueForKey(optkey, xref);
 461                                         if (!val.empty())
 462                                                 ret += expandFormat(ifpart, xref, counter, buf, richtext);
 463                                         else if (!elsepart.empty())
 464                                                 ret += expandFormat(elsepart, xref, counter, buf, richtext);
 465                                         // fmt will have been shortened for us already
 466                                         continue;
 467                                 }
 468                                 if (fmt[1] == '!') {
 469                                         // beginning of rich text
 470                                         scanning_rich = true;
 471                                         fmt = fmt.substr(2);
 472                                         continue;
 473                                 }
 474                         }
 475                         // we are here if '{' was not followed by % or !.
 476                         // So it's just a character.
 477                         ret += thischar;
 478                 }
 479                 else if (scanning_rich && thischar == '!'
 480                          && fmt.size() > 1 && fmt[1] == '}') {
 481                         // end of rich text
 482                         scanning_rich = false;
 483                         fmt = fmt.substr(2);
 484                         continue;
 485                 }
 486                 else if (scanning_key)
 487                         key += char(thischar);
 488                 else if (richtext) {
 489                         if (scanning_rich)
 490                                 ret += thischar;
 491                         else {
 492                                 // we need to escape '<' and '>'
 493                                 if (thischar == '<')
 494                                         ret += "&lt;";
 495                                 else if (thischar == '>')
 496                                         ret += "&gt;";
 497                                 else
 498                                         ret += thischar;
 499                         }
 500                 }       else if (!scanning_rich /* && !richtext */)
 501                         ret += thischar;
 502                 // else the character is discarded, which will happen only if
 503                 // richtext == false and we are scanning rich text
 504                 fmt = fmt.substr(1);
 505         } // for loop
 506         if (scanning_key) {
 507                 LYXERR0("Never found end of key in `" << format << "'!");
 508                 return _("ERROR!");
 509         }
 510         if (scanning_rich) {
 511                 LYXERR0("Never found end of rich text in `" << format << "'!");
 512                 return _("ERROR!");
 513         }
 514         return ret;
 515 }
 516
 517
 518 docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref,
 519         Buffer const & buf, bool richtext) const
 520 {
 521         if (!info_.empty())
 522                 return info_;
 523
 524         if (!is_bibtex_) {
 525                 BibTeXInfo::const_iterator it = find(from_ascii("ref"));
 526                 info_ = it->second;
 527                 return info_;
 528         }
 529
 530         DocumentClass const & dc = buf.params().documentClass();
 531         string const & format = dc.getCiteFormat(to_utf8(entry_type_));
 532         int counter = 0;
 533         info_ = expandFormat(format, xref, counter, buf, richtext);
 534
 535         if (!info_.empty())
 536                 info_ = convertLaTeXCommands(info_);
 537         return info_;
 538 }
 539
 540
 541 docstring const & BibTeXInfo::operator[](docstring const & field) const
 542 {
 543         BibTeXInfo::const_iterator it = find(field);
 544         if (it != end())
 545                 return it->second;
 546         static docstring const empty_value = docstring();
 547         return empty_value;
 548 }
 549
 550
 551 docstring const & BibTeXInfo::operator[](string const & field) const
 552 {
 553         return operator[](from_ascii(field));
 554 }
 555
 556
 557 docstring BibTeXInfo::getValueForKey(string const & key,
 558                 BibTeXInfo const * const xref) const
 559 {
 560         docstring const ret = operator[](key);
 561         if (!ret.empty() || !xref)
 562                 return ret;
 563         return (*xref)[key];
 564 }
 565
 566
 567 //////////////////////////////////////////////////////////////////////
 568 //
 569 // BiblioInfo
 570 //
 571 //////////////////////////////////////////////////////////////////////
 572
 573 namespace {
 574 // A functor for use with sort, leading to case insensitive sorting
 575         class compareNoCase: public binary_function<docstring, docstring, bool>
 576         {
 577                 public:
 578                         bool operator()(docstring const & s1, docstring const & s2) const {
 579                                 return compare_no_case(s1, s2) < 0;
 580                         }
 581         };
 582 } // namespace anon
 583
 584
 585 vector<docstring> const BiblioInfo::getKeys() const
 586 {
 587         vector<docstring> bibkeys;
 588         BiblioInfo::const_iterator it  = begin();
 589         for (; it != end(); ++it)
 590                 bibkeys.push_back(it->first);
 591         sort(bibkeys.begin(), bibkeys.end(), compareNoCase());
 592         return bibkeys;
 593 }
 594
 595
 596 vector<docstring> const BiblioInfo::getFields() const
 597 {
 598         vector<docstring> bibfields;
 599         set<docstring>::const_iterator it = field_names_.begin();
 600         set<docstring>::const_iterator end = field_names_.end();
 601         for (; it != end; ++it)
 602                 bibfields.push_back(*it);
 603         sort(bibfields.begin(), bibfields.end());
 604         return bibfields;
 605 }
 606
 607
 608 vector<docstring> const BiblioInfo::getEntries() const
 609 {
 610         vector<docstring> bibentries;
 611         set<docstring>::const_iterator it = entry_types_.begin();
 612         set<docstring>::const_iterator end = entry_types_.end();
 613         for (; it != end; ++it)
 614                 bibentries.push_back(*it);
 615         sort(bibentries.begin(), bibentries.end());
 616         return bibentries;
 617 }
 618
 619
 620 docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key) const
 621 {
 622         BiblioInfo::const_iterator it = find(key);
 623         if (it == end())
 624                 return docstring();
 625         BibTeXInfo const & data = it->second;
 626         return data.getAbbreviatedAuthor();
 627 }
 628
 629
 630 docstring const BiblioInfo::getCiteNumber(docstring const & key) const
 631 {
 632         BiblioInfo::const_iterator it = find(key);
 633         if (it == end())
 634                 return docstring();
 635         BibTeXInfo const & data = it->second;
 636         return data.citeNumber();
 637 }
 638
 639
 640 docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier) const
 641 {
 642         BiblioInfo::const_iterator it = find(key);
 643         if (it == end())
 644                 return docstring();
 645         BibTeXInfo const & data = it->second;
 646         docstring year = data.getYear();
 647         if (year.empty()) {
 648                 // let's try the crossref
 649                 docstring const xref = data.getXRef();
 650                 if (xref.empty())
 651                         return _("No year"); // no luck
 652                 BiblioInfo::const_iterator const xrefit = find(xref);
 653                 if (xrefit == end())
 654                         return _("No year"); // no luck again
 655                 BibTeXInfo const & xref_data = xrefit->second;
 656                 year = xref_data.getYear();
 657         }
 658         if (use_modifier && data.modifier() != 0)
 659                 year += data.modifier();
 660         return year;
 661 }
 662
 663
 664 docstring const BiblioInfo::getInfo(docstring const & key,
 665         Buffer const & buf, bool richtext) const
 666 {
 667         BiblioInfo::const_iterator it = find(key);
 668         if (it == end())
 669                 return docstring(_("BibTeX entry not found!"));
 670         BibTeXInfo const & data = it->second;
 671         BibTeXInfo const * xrefptr = 0;
 672         docstring const xref = data.getXRef();
 673         if (!xref.empty()) {
 674                 BiblioInfo::const_iterator const xrefit = find(xref);
 675                 if (xrefit != end())
 676                         xrefptr = &(xrefit->second);
 677         }
 678         return data.getInfo(xrefptr, buf, richtext);
 679 }
 680
 681
 682 bool BiblioInfo::isBibtex(docstring const & key) const
 683 {
 684         BiblioInfo::const_iterator it = find(key);
 685         if (it == end())
 686                 return false;
 687         return it->second.isBibTeX();
 688 }
 689
 690
 691
 692 vector<docstring> const BiblioInfo::getCiteStrings(
 693         docstring const & key, Buffer const & buf) const
 694 {
 695         CiteEngine const engine = buf.params().citeEngine();
 696         if (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL)
 697                 return getNumericalStrings(key, buf);
 698         else
 699                 return getAuthorYearStrings(key, buf);
 700 }
 701
 702
 703 vector<docstring> const BiblioInfo::getNumericalStrings(
 704         docstring const & key, Buffer const & buf) const
 705 {
 706         if (empty())
 707                 return vector<docstring>();
 708
 709         docstring const author = getAbbreviatedAuthor(key);
 710         docstring const year   = getYear(key);
 711         if (author.empty() || year.empty())
 712                 return vector<docstring>();
 713
 714         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 715
 716         vector<docstring> vec(styles.size());
 717         for (size_t i = 0; i != vec.size(); ++i) {
 718                 docstring str;
 719
 720                 switch (styles[i]) {
 721                         case CITE:
 722                         case CITEP:
 723                                 str = from_ascii("[#ID]");
 724                                 break;
 725
 726                         case NOCITE:
 727                                 str = _("Add to bibliography only.");
 728                                 break;
 729
 730                         case CITET:
 731                                 str = author + " [#ID]";
 732                                 break;
 733
 734                         case CITEALT:
 735                                 str = author + " #ID";
 736                                 break;
 737
 738                         case CITEALP:
 739                                 str = from_ascii("#ID");
 740                                 break;
 741
 742                         case CITEAUTHOR:
 743                                 str = author;
 744                                 break;
 745
 746                         case CITEYEAR:
 747                                 str = year;
 748                                 break;
 749
 750                         case CITEYEARPAR:
 751                                 str = '(' + year + ')';
 752                                 break;
 753                 }
 754
 755                 vec[i] = str;
 756         }
 757
 758         return vec;
 759 }
 760
 761
 762 vector<docstring> const BiblioInfo::getAuthorYearStrings(
 763         docstring const & key, Buffer const & buf) const
 764 {
 765         if (empty())
 766                 return vector<docstring>();
 767
 768         docstring const author = getAbbreviatedAuthor(key);
 769         docstring const year   = getYear(key);
 770         if (author.empty() || year.empty())
 771                 return vector<docstring>();
 772
 773         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 774
 775         vector<docstring> vec(styles.size());
 776         for (size_t i = 0; i != vec.size(); ++i) {
 777                 docstring str;
 778
 779                 switch (styles[i]) {
 780                         case CITE:
 781                 // jurabib only: Author/Annotator
 782                 // (i.e. the "before" field, 2nd opt arg)
 783                                 str = author + "/<" + _("before") + '>';
 784                                 break;
 785
 786                         case NOCITE:
 787                                 str = _("Add to bibliography only.");
 788                                 break;
 789
 790                         case CITET:
 791                                 str = author + " (" + year + ')';
 792                                 break;
 793
 794                         case CITEP:
 795                                 str = '(' + author + ", " + year + ')';
 796                                 break;
 797
 798                         case CITEALT:
 799                                 str = author + ' ' + year ;
 800                                 break;
 801
 802                         case CITEALP:
 803                                 str = author + ", " + year ;
 804                                 break;
 805
 806                         case CITEAUTHOR:
 807                                 str = author;
 808                                 break;
 809
 810                         case CITEYEAR:
 811                                 str = year;
 812                                 break;
 813
 814                         case CITEYEARPAR:
 815                                 str = '(' + year + ')';
 816                                 break;
 817                 }
 818                 vec[i] = str;
 819         }
 820         return vec;
 821 }
 822
 823
 824 void BiblioInfo::mergeBiblioInfo(BiblioInfo const & info)
 825 {
 826         bimap_.insert(info.begin(), info.end());
 827         field_names_.insert(info.field_names_.begin(), info.field_names_.end());
 828         entry_types_.insert(info.entry_types_.begin(), info.entry_types_.end());
 829 }
 830
 831
 832 namespace {
 833         // used in xhtml to sort a list of BibTeXInfo objects
 834         bool lSorter(BibTeXInfo const * lhs, BibTeXInfo const * rhs)
 835         {
 836                 docstring const lauth = lhs->getAbbreviatedAuthor();
 837                 docstring const rauth = rhs->getAbbreviatedAuthor();
 838                 docstring const lyear = lhs->getYear();
 839                 docstring const ryear = rhs->getYear();
 840                 docstring const ltitl = lhs->operator[]("title");
 841                 docstring const rtitl = rhs->operator[]("title");
 842                 return  (lauth < rauth)
 843                                 || (lauth == rauth && lyear < ryear)
 844                                 || (lauth == rauth && lyear == ryear && ltitl < rtitl);
 845         }
 846 }
 847
 848
 849 void BiblioInfo::collectCitedEntries(Buffer const & buf)
 850 {
 851         cited_entries_.clear();
 852         // We are going to collect all the citation keys used in the document,
 853         // getting them from the TOC.
 854         // FIXME We may want to collect these differently, in the first case,
 855         // so that we might have them in order of appearance.
 856         set<docstring> citekeys;
 857         Toc const & toc = buf.tocBackend().toc("citation");
 858         Toc::const_iterator it = toc.begin();
 859         Toc::const_iterator const en = toc.end();
 860         for (; it != en; ++it) {
 861                 if (it->str().empty())
 862                         continue;
 863                 vector<docstring> const keys = getVectorFromString(it->str());
 864                 citekeys.insert(keys.begin(), keys.end());
 865         }
 866         if (citekeys.empty())
 867                 return;
 868
 869         // We have a set of the keys used in this document.
 870         // We will now convert it to a list of the BibTeXInfo objects used in
 871         // this document...
 872         vector<BibTeXInfo const *> bi;
 873         set<docstring>::const_iterator cit = citekeys.begin();
 874         set<docstring>::const_iterator const cen = citekeys.end();
 875         for (; cit != cen; ++cit) {
 876                 BiblioInfo::const_iterator const bt = find(*cit);
 877                 if (bt == end() || !bt->second.isBibTeX())
 878                         continue;
 879                 bi.push_back(&(bt->second));
 880         }
 881         // ...and sort it.
 882         sort(bi.begin(), bi.end(), lSorter);
 883
 884         // Now we can write the sorted keys
 885         vector<BibTeXInfo const *>::const_iterator bit = bi.begin();
 886         vector<BibTeXInfo const *>::const_iterator ben = bi.end();
 887         for (; bit != ben; ++bit)
 888                 cited_entries_.push_back((*bit)->key());
 889 }
 890
 891
 892 void BiblioInfo::makeCitationLabels(Buffer const & buf)
 893 {
 894         collectCitedEntries(buf);
 895         CiteEngine const engine = buf.params().citeEngine();
 896         bool const numbers =
 897                 (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL);
 898
 899         int keynumber = 0;
 900         char modifier = 0;
 901         // used to remember the last one we saw
 902         // we'll be comparing entries to see if we need to add
 903         // modifiers, like "1984a"
 904         map<docstring, BibTeXInfo>::iterator last;
 905
 906         vector<docstring>::const_iterator it = cited_entries_.begin();
 907         vector<docstring>::const_iterator const en = cited_entries_.end();
 908         for (; it != en; ++it) {
 909                 map<docstring, BibTeXInfo>::iterator const biit = bimap_.find(*it);
 910                 // this shouldn't happen, but...
 911                 if (biit == bimap_.end())
 912                         // ...fail gracefully, anyway.
 913                         continue;
 914                 BibTeXInfo & entry = biit->second;
 915                 if (numbers) {
 916                         docstring const num = convert<docstring>(++keynumber);
 917                         entry.setCiteNumber(num);
 918                 } else {
 919                         if (it != cited_entries_.begin()
 920                             && entry.getAbbreviatedAuthor() == last->second.getAbbreviatedAuthor()
 921                             // we access the year via getYear() so as to get it from the xref,
 922                             // if we need to do so
 923                             && getYear(entry.key()) == getYear(last->second.key())) {
 924                                 if (modifier == 0) {
 925                                         // so the last one should have been 'a'
 926                                         last->second.setModifier('a');
 927                                         modifier = 'b';
 928                                 } else if (modifier == 'z')
 929                                         modifier = 'A';
 930                                 else
 931                                         modifier++;
 932                         } else {
 933                                 modifier = 0;
 934                         }
 935                         entry.setModifier(modifier);
 936                         // remember the last one
 937                         last = biit;
 938                 }
 939         }
 940 }
 941
 942
 943 //////////////////////////////////////////////////////////////////////
 944 //
 945 // CitationStyle
 946 //
 947 //////////////////////////////////////////////////////////////////////
 948
 949 namespace {
 950
 951
 952 char const * const citeCommands[] = {
 953         "cite", "citet", "citep", "citealt", "citealp",
 954         "citeauthor", "citeyear", "citeyearpar", "nocite" };
 955
 956 unsigned int const nCiteCommands =
 957                 sizeof(citeCommands) / sizeof(char *);
 958
 959 CiteStyle const citeStylesArray[] = {
 960         CITE, CITET, CITEP, CITEALT, CITEALP,
 961         CITEAUTHOR, CITEYEAR, CITEYEARPAR, NOCITE };
 962
 963 unsigned int const nCiteStyles =
 964                 sizeof(citeStylesArray) / sizeof(CiteStyle);
 965
 966 CiteStyle const citeStylesFull[] = {
 967         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 968
 969 unsigned int const nCiteStylesFull =
 970                 sizeof(citeStylesFull) / sizeof(CiteStyle);
 971
 972 CiteStyle const citeStylesUCase[] = {
 973         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 974
 975 unsigned int const nCiteStylesUCase =
 976         sizeof(citeStylesUCase) / sizeof(CiteStyle);
 977
 978 } // namespace anon
 979
 980
 981 CitationStyle citationStyleFromString(string const & command)
 982 {
 983         CitationStyle s;
 984         if (command.empty())
 985                 return s;
 986
 987         string cmd = command;
 988         if (cmd[0] == 'C') {
 989                 s.forceUpperCase = true;
 990                 cmd[0] = 'c';
 991         }
 992
 993         size_t const n = cmd.size() - 1;
 994         if (cmd != "cite" && cmd[n] == '*') {
 995                 s.full = true;
 996                 cmd = cmd.substr(0, n);
 997         }
 998
 999         char const * const * const last = citeCommands + nCiteCommands;
1000         char const * const * const ptr = find(citeCommands, last, cmd);
1001
1002         if (ptr != last) {
1003                 size_t idx = ptr - citeCommands;
1004                 s.style = citeStylesArray[idx];
1005         }
1006         return s;
1007 }
1008
1009
1010 string citationStyleToString(const CitationStyle & s)
1011 {
1012         string cite = citeCommands[s.style];
1013         if (s.full) {
1014                 CiteStyle const * last = citeStylesFull + nCiteStylesFull;
1015                 if (std::find(citeStylesFull, last, s.style) != last)
1016                         cite += '*';
1017         }
1018
1019         if (s.forceUpperCase) {
1020                 CiteStyle const * last = citeStylesUCase + nCiteStylesUCase;
1021                 if (std::find(citeStylesUCase, last, s.style) != last)
1022                         cite[0] = 'C';
1023         }
1024
1025         return cite;
1026 }
1027
1028 vector<CiteStyle> citeStyles(CiteEngine engine)
1029 {
1030         vector<CiteStyle> styles(0);
1031
1032         switch (engine) {
1033                 case ENGINE_BASIC:
1034                         styles.push_back(CITE);
1035                         break;
1036                 case ENGINE_JURABIB:
1037                         styles.push_back(CITE);
1038                 case ENGINE_NATBIB_AUTHORYEAR:
1039                         styles.push_back(CITET);
1040                         styles.push_back(CITEP);
1041                         styles.push_back(CITEALT);
1042                         styles.push_back(CITEALP);
1043                         styles.push_back(CITEAUTHOR);
1044                         styles.push_back(CITEYEAR);
1045                         styles.push_back(CITEYEARPAR);
1046                         break;
1047                 case ENGINE_NATBIB_NUMERICAL:
1048                         styles.push_back(CITET);
1049                         styles.push_back(CITEALT);
1050                         styles.push_back(CITEAUTHOR);
1051                         styles.push_back(CITEP);
1052                         styles.push_back(CITEALP);
1053                         styles.push_back(CITEYEAR);
1054                         styles.push_back(CITEYEARPAR);
1055                         break;
1056         }
1057
1058         styles.push_back(NOCITE);
1059
1060         return styles;
1061 }
1062
1063 } // namespace lyx
1064