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
  23 #include "insets/Inset.h"
  24 #include "insets/InsetBibitem.h"
  25 #include "insets/InsetBibtex.h"
  26 #include "insets/InsetInclude.h"
  27
  28 #include "support/docstream.h"
  29 #include "support/gettext.h"
  30 #include "support/lassert.h"
  31 #include "support/lstrings.h"
  32 #include "support/textutils.h"
  33
  34 #include "boost/regex.hpp"
  35
  36 using namespace std;
  37 using namespace lyx::support;
  38
  39
  40 namespace lyx {
  41
  42 //////////////////////////////////////////////////////////////////////
  43 //
  44 // BibTeXInfo
  45 //
  46 //////////////////////////////////////////////////////////////////////
  47
  48 BibTeXInfo::BibTeXInfo(docstring const & key, docstring const & type)
  49         : is_bibtex_(true), bib_key_(key), entry_type_(type), info_()
  50 {}
  51
  52
  53 bool BibTeXInfo::hasField(docstring const & field) const
  54 {
  55         return count(field) == 1;
  56 }
  57
  58
  59 docstring const & BibTeXInfo::operator[](docstring const & field) const
  60 {
  61         BibTeXInfo::const_iterator it = find(field);
  62         if (it != end())
  63                 return it->second;
  64         static docstring const empty_value = docstring();
  65         return empty_value;
  66 }
  67
  68
  69 docstring const & BibTeXInfo::operator[](string const & field) const
  70 {
  71         return operator[](from_ascii(field));
  72 }
  73
  74
  75 docstring familyName(docstring const & name)
  76 {
  77         if (name.empty())
  78                 return docstring();
  79
  80         // first we look for a comma, and take the last name to be everything
  81         // preceding the right-most one, so that we also get the "jr" part.
  82         docstring::size_type idx = name.rfind(',');
  83         if (idx != docstring::npos)
  84                 return ltrim(name.substr(0, idx));
  85
  86         // OK, so now we want to look for the last name. We're going to
  87         // include the "von" part. This isn't perfect.
  88         // Split on spaces, to get various tokens.
  89         vector<docstring> pieces = getVectorFromString(name, from_ascii(" "));
  90         // If we only get two, assume the last one is the last name
  91         if (pieces.size() <= 2)
  92                 return pieces.back();
  93
  94         // Now we look for the first token that begins with a lower case letter.
  95         vector<docstring>::const_iterator it = pieces.begin();
  96         vector<docstring>::const_iterator en = pieces.end();
  97         for (; it != en; ++it) {
  98                 if ((*it).size() == 0)
  99                         continue;
 100                 char_type const c = (*it)[0];
 101                 if (isLower(c))
 102                         break;
 103         }
 104
 105         if (it == en) // we never found a "von"
 106                 return pieces.back();
 107
 108         // reconstruct what we need to return
 109         docstring retval;
 110         bool first = true;
 111         for (; it != en; ++it) {
 112                 if (!first)
 113                         retval += " ";
 114                 else
 115                         first = false;
 116                 retval += *it;
 117         }
 118         return retval;
 119 }
 120
 121 docstring const BibTeXInfo::getAbbreviatedAuthor() const
 122 {
 123         if (!is_bibtex_) {
 124                 docstring const opt = trim(operator[]("label"));
 125                 if (opt.empty())
 126                         return docstring();
 127
 128                 docstring authors;
 129                 split(opt, authors, '(');
 130                 return authors;
 131         }
 132
 133         docstring author = operator[]("author");
 134         if (author.empty()) {
 135                 author = operator[]("editor");
 136                 if (author.empty())
 137                         return bib_key_;
 138         }
 139
 140         // OK, we've got some names. Let's format them.
 141         // Try to split the author list on " and "
 142         vector<docstring> const authors =
 143                 getVectorFromString(author, from_ascii(" and "));
 144
 145         if (authors.size() == 2)
 146                 return bformat(_("%1$s and %2$s"),
 147                         familyName(authors[0]), familyName(authors[1]));
 148
 149         if (authors.size() > 2)
 150                 return bformat(_("%1$s et al."), familyName(authors[0]));
 151
 152         return familyName(authors[0]);
 153 }
 154
 155
 156 docstring const BibTeXInfo::getYear() const
 157 {
 158         if (!is_bibtex_) {
 159                 docstring const opt = trim(operator[]("label"));
 160                 if (opt.empty())
 161                         return docstring();
 162
 163                 docstring authors;
 164                 docstring const tmp = split(opt, authors, '(');
 165                 docstring year;
 166                 split(tmp, year, ')');
 167                 return year;
 168         }
 169
 170         docstring year = operator[]("year");
 171         if (year.empty())
 172                 year = _("No year");
 173         return year;
 174 }
 175
 176
 177 namespace {
 178
 179         docstring convertLaTeXCommands(docstring const & str)
 180         {
 181                 docstring val = str;
 182                 docstring ret;
 183
 184                 bool scanning_cmd = false;
 185                 bool scanning_math = false;
 186                 bool escaped = false; // used to catch \$, etc.
 187                 while (val.size()) {
 188                         char_type const ch = val[0];
 189
 190                         // if we're scanning math, we output everything until we
 191                         // find an unescaped $, at which point we break out.
 192                         if (scanning_math) {
 193                                 if (escaped)
 194                                         escaped = false;
 195                                 else if (ch == '\\')
 196                                         escaped = true;
 197                                 else if (ch == '$')
 198                                         scanning_math = false;
 199                                 ret += ch;
 200                                 val = val.substr(1);
 201                                 continue;
 202                         }
 203
 204                         // if we're scanning a command name, then we just
 205                         // discard characters until we hit something that
 206                         // isn't alpha.
 207                         if (scanning_cmd) {
 208                                 if (isAlphaASCII(ch)) {
 209                                         val = val.substr(1);
 210                                         escaped = false;
 211                                         continue;
 212                                 }
 213                                 // so we're done with this command.
 214                                 // now we fall through and check this character.
 215                                 scanning_cmd = false;
 216                         }
 217
 218                         // was the last character a \? If so, then this is something like: \\,
 219                         // or \$, so we'll just output it. That's probably not always right...
 220                         if (escaped) {
 221                                 ret += ch;
 222                                 val = val.substr(1);
 223                                 escaped = false;
 224                                 continue;
 225                         }
 226
 227                         if (ch == '$') {
 228                                 ret += ch;
 229                                 val = val.substr(1);
 230                                 scanning_math = true;
 231                                 continue;
 232                         }
 233
 234                         // we just ignore braces
 235                         if (ch == '{' || ch == '}') {
 236                                 val = val.substr(1);
 237                                 continue;
 238                         }
 239
 240                         // we're going to check things that look like commands, so if
 241                         // this doesn't, just output it.
 242                         if (ch != '\\') {
 243                                 ret += ch;
 244                                 val = val.substr(1);
 245                                 continue;
 246                         }
 247
 248                         // ok, could be a command of some sort
 249                         // let's see if it corresponds to some unicode
 250                         // unicodesymbols has things in the form: \"{u},
 251                         // whereas we may see things like: \"u. So we'll
 252                         // look for that and change it, if necessary.
 253                         static boost::regex const reg("^\\\\\\W\\w");
 254                         if (boost::regex_search(to_utf8(val), reg)) {
 255                                 val.insert(3, from_ascii("}"));
 256                                 val.insert(2, from_ascii("{"));
 257                         }
 258                         docstring rem;
 259                         docstring const cnvtd = Encodings::fromLaTeXCommand(val, rem);
 260                         if (!cnvtd.empty()) {
 261                                 // it did, so we'll take that bit and proceed with what's left
 262                                 ret += cnvtd;
 263                                 val = rem;
 264                                 continue;
 265                         }
 266                         // it's a command of some sort
 267                         scanning_cmd = true;
 268                         escaped = true;
 269                         val = val.substr(1);
 270                 }
 271                 return ret;
 272         }
 273
 274 } // anon namespace
 275
 276
 277 docstring const BibTeXInfo::getInfo() const
 278 {
 279         if (!info_.empty())
 280                 return info_;
 281
 282         if (!is_bibtex_) {
 283                 BibTeXInfo::const_iterator it = find(from_ascii("ref"));
 284                 info_ = it->second;
 285                 return info_;
 286         }
 287
 288         // FIXME
 289         // This could be made a lot better using the entryType
 290         // field to customize the output based upon entry type.
 291
 292         // Search for all possible "required" fields
 293         docstring author = operator[]("author");
 294         if (author.empty())
 295                 author = operator[]("editor");
 296
 297         docstring year   = operator[]("year");
 298         docstring title  = operator[]("title");
 299         docstring docLoc = operator[]("pages");
 300         if (docLoc.empty()) {
 301                 docLoc = operator[]("chapter");
 302                 if (!docLoc.empty())
 303                         docLoc = from_ascii("Ch. ") + docLoc;
 304         }       else {
 305                 docLoc = from_ascii("pp. ") + docLoc;
 306         }
 307
 308         docstring media = operator[]("journal");
 309         if (media.empty()) {
 310                 media = operator[]("publisher");
 311                 if (media.empty()) {
 312                         media = operator[]("school");
 313                         if (media.empty())
 314                                 media = operator[]("institution");
 315                 }
 316         }
 317         docstring volume = operator[]("volume");
 318
 319         odocstringstream result;
 320         if (!author.empty())
 321                 result << author << ", ";
 322         if (!title.empty())
 323                 result << title;
 324         if (!media.empty())
 325                 result << ", " << media;
 326         if (!year.empty())
 327                 result << ", " << year;
 328         if (!docLoc.empty())
 329                 result << ", " << docLoc;
 330
 331         docstring const result_str = rtrim(result.str());
 332         if (!result_str.empty()) {
 333                 info_ = convertLaTeXCommands(result_str);
 334                 return info_;
 335         }
 336
 337         // This should never happen (or at least be very unusual!)
 338         return docstring();
 339 }
 340
 341
 342 //////////////////////////////////////////////////////////////////////
 343 //
 344 // BiblioInfo
 345 //
 346 //////////////////////////////////////////////////////////////////////
 347
 348 namespace {
 349 // A functor for use with sort, leading to case insensitive sorting
 350         class compareNoCase: public binary_function<docstring, docstring, bool>
 351         {
 352                 public:
 353                         bool operator()(docstring const & s1, docstring const & s2) const {
 354                                 return compare_no_case(s1, s2) < 0;
 355                         }
 356         };
 357 } // namespace anon
 358
 359
 360 vector<docstring> const BiblioInfo::getKeys() const
 361 {
 362         vector<docstring> bibkeys;
 363         BiblioInfo::const_iterator it  = begin();
 364         for (; it != end(); ++it)
 365                 bibkeys.push_back(it->first);
 366         sort(bibkeys.begin(), bibkeys.end(), compareNoCase());
 367         return bibkeys;
 368 }
 369
 370
 371 vector<docstring> const BiblioInfo::getFields() const
 372 {
 373         vector<docstring> bibfields;
 374         set<docstring>::const_iterator it = field_names_.begin();
 375         set<docstring>::const_iterator end = field_names_.end();
 376         for (; it != end; ++it)
 377                 bibfields.push_back(*it);
 378         sort(bibfields.begin(), bibfields.end());
 379         return bibfields;
 380 }
 381
 382
 383 vector<docstring> const BiblioInfo::getEntries() const
 384 {
 385         vector<docstring> bibentries;
 386         set<docstring>::const_iterator it = entry_types_.begin();
 387         set<docstring>::const_iterator end = entry_types_.end();
 388         for (; it != end; ++it)
 389                 bibentries.push_back(*it);
 390         sort(bibentries.begin(), bibentries.end());
 391         return bibentries;
 392 }
 393
 394
 395 docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key) const
 396 {
 397         BiblioInfo::const_iterator it = find(key);
 398         if (it == end())
 399                 return docstring();
 400         BibTeXInfo const & data = it->second;
 401         return data.getAbbreviatedAuthor();
 402 }
 403
 404
 405 docstring const BiblioInfo::getYear(docstring const & key) const
 406 {
 407         BiblioInfo::const_iterator it = find(key);
 408         if (it == end())
 409                 return docstring();
 410         BibTeXInfo const & data = it->second;
 411         return data.getYear();
 412 }
 413
 414
 415 docstring const BiblioInfo::getInfo(docstring const & key) const
 416 {
 417         BiblioInfo::const_iterator it = find(key);
 418         if (it == end())
 419                 return docstring();
 420         BibTeXInfo const & data = it->second;
 421         return data.getInfo();
 422 }
 423
 424
 425 vector<docstring> const BiblioInfo::getCiteStrings(
 426         docstring const & key, Buffer const & buf) const
 427 {
 428         CiteEngine const engine = buf.params().citeEngine();
 429         if (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL)
 430                 return getNumericalStrings(key, buf);
 431         else
 432                 return getAuthorYearStrings(key, buf);
 433 }
 434
 435
 436 vector<docstring> const BiblioInfo::getNumericalStrings(
 437         docstring const & key, Buffer const & buf) const
 438 {
 439         if (empty())
 440                 return vector<docstring>();
 441
 442         docstring const author = getAbbreviatedAuthor(key);
 443         docstring const year   = getYear(key);
 444         if (author.empty() || year.empty())
 445                 return vector<docstring>();
 446
 447         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 448
 449         vector<docstring> vec(styles.size());
 450         for (size_t i = 0; i != vec.size(); ++i) {
 451                 docstring str;
 452
 453                 switch (styles[i]) {
 454                         case CITE:
 455                         case CITEP:
 456                                 str = from_ascii("[#ID]");
 457                                 break;
 458
 459                         case NOCITE:
 460                                 str = _("Add to bibliography only.");
 461                                 break;
 462
 463                         case CITET:
 464                                 str = author + " [#ID]";
 465                                 break;
 466
 467                         case CITEALT:
 468                                 str = author + " #ID";
 469                                 break;
 470
 471                         case CITEALP:
 472                                 str = from_ascii("#ID");
 473                                 break;
 474
 475                         case CITEAUTHOR:
 476                                 str = author;
 477                                 break;
 478
 479                         case CITEYEAR:
 480                                 str = year;
 481                                 break;
 482
 483                         case CITEYEARPAR:
 484                                 str = '(' + year + ')';
 485                                 break;
 486                 }
 487
 488                 vec[i] = str;
 489         }
 490
 491         return vec;
 492 }
 493
 494
 495 vector<docstring> const BiblioInfo::getAuthorYearStrings(
 496         docstring const & key, Buffer const & buf) const
 497 {
 498         if (empty())
 499                 return vector<docstring>();
 500
 501         docstring const author = getAbbreviatedAuthor(key);
 502         docstring const year   = getYear(key);
 503         if (author.empty() || year.empty())
 504                 return vector<docstring>();
 505
 506         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 507
 508         vector<docstring> vec(styles.size());
 509         for (size_t i = 0; i != vec.size(); ++i) {
 510                 docstring str;
 511
 512                 switch (styles[i]) {
 513                         case CITE:
 514                 // jurabib only: Author/Annotator
 515                 // (i.e. the "before" field, 2nd opt arg)
 516                                 str = author + "/<" + _("before") + '>';
 517                                 break;
 518
 519                         case NOCITE:
 520                                 str = _("Add to bibliography only.");
 521                                 break;
 522
 523                         case CITET:
 524                                 str = author + " (" + year + ')';
 525                                 break;
 526
 527                         case CITEP:
 528                                 str = '(' + author + ", " + year + ')';
 529                                 break;
 530
 531                         case CITEALT:
 532                                 str = author + ' ' + year ;
 533                                 break;
 534
 535                         case CITEALP:
 536                                 str = author + ", " + year ;
 537                                 break;
 538
 539                         case CITEAUTHOR:
 540                                 str = author;
 541                                 break;
 542
 543                         case CITEYEAR:
 544                                 str = year;
 545                                 break;
 546
 547                         case CITEYEARPAR:
 548                                 str = '(' + year + ')';
 549                                 break;
 550                 }
 551                 vec[i] = str;
 552         }
 553         return vec;
 554 }
 555
 556
 557 void BiblioInfo::mergeBiblioInfo(BiblioInfo const & info)
 558 {
 559         bimap_.insert(info.begin(), info.end());
 560 }
 561
 562
 563 //////////////////////////////////////////////////////////////////////
 564 //
 565 // CitationStyle
 566 //
 567 //////////////////////////////////////////////////////////////////////
 568
 569 namespace {
 570
 571
 572 char const * const citeCommands[] = {
 573         "cite", "citet", "citep", "citealt", "citealp",
 574         "citeauthor", "citeyear", "citeyearpar", "nocite" };
 575
 576 unsigned int const nCiteCommands =
 577                 sizeof(citeCommands) / sizeof(char *);
 578
 579 CiteStyle const citeStylesArray[] = {
 580         CITE, CITET, CITEP, CITEALT, CITEALP,
 581         CITEAUTHOR, CITEYEAR, CITEYEARPAR, NOCITE };
 582
 583 unsigned int const nCiteStyles =
 584                 sizeof(citeStylesArray) / sizeof(CiteStyle);
 585
 586 CiteStyle const citeStylesFull[] = {
 587         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 588
 589 unsigned int const nCiteStylesFull =
 590                 sizeof(citeStylesFull) / sizeof(CiteStyle);
 591
 592 CiteStyle const citeStylesUCase[] = {
 593         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 594
 595 unsigned int const nCiteStylesUCase =
 596         sizeof(citeStylesUCase) / sizeof(CiteStyle);
 597
 598 } // namespace anon
 599
 600
 601 CitationStyle citationStyleFromString(string const & command)
 602 {
 603         CitationStyle s;
 604         if (command.empty())
 605                 return s;
 606
 607         string cmd = command;
 608         if (cmd[0] == 'C') {
 609                 s.forceUpperCase = true;
 610                 cmd[0] = 'c';
 611         }
 612
 613         size_t const n = cmd.size() - 1;
 614         if (cmd != "cite" && cmd[n] == '*') {
 615                 s.full = true;
 616                 cmd = cmd.substr(0, n);
 617         }
 618
 619         char const * const * const last = citeCommands + nCiteCommands;
 620         char const * const * const ptr = find(citeCommands, last, cmd);
 621
 622         if (ptr != last) {
 623                 size_t idx = ptr - citeCommands;
 624                 s.style = citeStylesArray[idx];
 625         }
 626         return s;
 627 }
 628
 629
 630 string citationStyleToString(const CitationStyle & s)
 631 {
 632         string cite = citeCommands[s.style];
 633         if (s.full) {
 634                 CiteStyle const * last = citeStylesFull + nCiteStylesFull;
 635                 if (find(citeStylesFull, last, s.style) != last)
 636                         cite += '*';
 637         }
 638
 639         if (s.forceUpperCase) {
 640                 CiteStyle const * last = citeStylesUCase + nCiteStylesUCase;
 641                 if (find(citeStylesUCase, last, s.style) != last)
 642                         cite[0] = 'C';
 643         }
 644
 645         return cite;
 646 }
 647
 648 vector<CiteStyle> citeStyles(CiteEngine engine)
 649 {
 650         unsigned int nStyles = 0;
 651         unsigned int start = 0;
 652
 653         switch (engine) {
 654                 case ENGINE_BASIC:
 655                         nStyles = 2;
 656                         start = 0;
 657                         break;
 658                 case ENGINE_NATBIB_AUTHORYEAR:
 659                 case ENGINE_NATBIB_NUMERICAL:
 660                         nStyles = nCiteStyles - 1;
 661                         start = 1;
 662                         break;
 663                 case ENGINE_JURABIB:
 664                         nStyles = nCiteStyles;
 665                         start = 0;
 666                         break;
 667         }
 668
 669         vector<CiteStyle> styles(nStyles);
 670         size_t i = 0;
 671         int j = start;
 672         for (; i != styles.size(); ++i, ++j)
 673                 styles[i] = citeStylesArray[j];
 674
 675         return styles;
 676 }
 677
 678 } // namespace lyx
 679