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 namespace {
  43
  44 // gets the "family name" from an author-type string
  45 docstring familyName(docstring const & name)
  46 {
  47         if (name.empty())
  48                 return docstring();
  49
  50         // first we look for a comma, and take the last name to be everything
  51         // preceding the right-most one, so that we also get the "jr" part.
  52         docstring::size_type idx = name.rfind(',');
  53         if (idx != docstring::npos)
  54                 return ltrim(name.substr(0, idx));
  55
  56         // OK, so now we want to look for the last name. We're going to
  57         // include the "von" part. This isn't perfect.
  58         // Split on spaces, to get various tokens.
  59         vector<docstring> pieces = getVectorFromString(name, from_ascii(" "));
  60         // If we only get two, assume the last one is the last name
  61         if (pieces.size() <= 2)
  62                 return pieces.back();
  63
  64         // Now we look for the first token that begins with a lower case letter.
  65         vector<docstring>::const_iterator it = pieces.begin();
  66         vector<docstring>::const_iterator en = pieces.end();
  67         for (; it != en; ++it) {
  68                 if ((*it).size() == 0)
  69                         continue;
  70                 char_type const c = (*it)[0];
  71                 if (isLower(c))
  72                         break;
  73         }
  74
  75         if (it == en) // we never found a "von"
  76                 return pieces.back();
  77
  78         // reconstruct what we need to return
  79         docstring retval;
  80         bool first = true;
  81         for (; it != en; ++it) {
  82                 if (!first)
  83                         retval += " ";
  84                 else
  85                         first = false;
  86                 retval += *it;
  87         }
  88         return retval;
  89 }
  90
  91 // converts a string containing LaTeX commands into unicode
  92 // for display.
  93 docstring convertLaTeXCommands(docstring const & str)
  94 {
  95         docstring val = str;
  96         docstring ret;
  97
  98         bool scanning_cmd = false;
  99         bool scanning_math = false;
 100         bool escaped = false; // used to catch \$, etc.
 101         while (val.size()) {
 102                 char_type const ch = val[0];
 103
 104                 // if we're scanning math, we output everything until we
 105                 // find an unescaped $, at which point we break out.
 106                 if (scanning_math) {
 107                         if (escaped)
 108                                 escaped = false;
 109                         else if (ch == '\\')
 110                                 escaped = true;
 111                         else if (ch == '$')
 112                                 scanning_math = false;
 113                         ret += ch;
 114                         val = val.substr(1);
 115                         continue;
 116                 }
 117
 118                 // if we're scanning a command name, then we just
 119                 // discard characters until we hit something that
 120                 // isn't alpha.
 121                 if (scanning_cmd) {
 122                         if (isAlphaASCII(ch)) {
 123                                 val = val.substr(1);
 124                                 escaped = false;
 125                                 continue;
 126                         }
 127                         // so we're done with this command.
 128                         // now we fall through and check this character.
 129                         scanning_cmd = false;
 130                 }
 131
 132                 // was the last character a \? If so, then this is something like: \\,
 133                 // or \$, so we'll just output it. That's probably not always right...
 134                 if (escaped) {
 135                         ret += ch;
 136                         val = val.substr(1);
 137                         escaped = false;
 138                         continue;
 139                 }
 140
 141                 if (ch == '$') {
 142                         ret += ch;
 143                         val = val.substr(1);
 144                         scanning_math = true;
 145                         continue;
 146                 }
 147
 148                 // we just ignore braces
 149                 if (ch == '{' || ch == '}') {
 150                         val = val.substr(1);
 151                         continue;
 152                 }
 153
 154                 // we're going to check things that look like commands, so if
 155                 // this doesn't, just output it.
 156                 if (ch != '\\') {
 157                         ret += ch;
 158                         val = val.substr(1);
 159                         continue;
 160                 }
 161
 162                 // ok, could be a command of some sort
 163                 // let's see if it corresponds to some unicode
 164                 // unicodesymbols has things in the form: \"{u},
 165                 // whereas we may see things like: \"u. So we'll
 166                 // look for that and change it, if necessary.
 167                 static boost::regex const reg("^\\\\\\W\\w");
 168                 if (boost::regex_search(to_utf8(val), reg)) {
 169                         val.insert(3, from_ascii("}"));
 170                         val.insert(2, from_ascii("{"));
 171                 }
 172                 docstring rem;
 173                 docstring const cnvtd = Encodings::fromLaTeXCommand(val, rem);
 174                 if (!cnvtd.empty()) {
 175                         // it did, so we'll take that bit and proceed with what's left
 176                         ret += cnvtd;
 177                         val = rem;
 178                         continue;
 179                 }
 180                 // it's a command of some sort
 181                 scanning_cmd = true;
 182                 escaped = true;
 183                 val = val.substr(1);
 184         }
 185         return ret;
 186 }
 187
 188 } // anon namespace
 189
 190
 191 //////////////////////////////////////////////////////////////////////
 192 //
 193 // BibTeXInfo
 194 //
 195 //////////////////////////////////////////////////////////////////////
 196
 197 BibTeXInfo::BibTeXInfo(docstring const & key, docstring const & type)
 198         : is_bibtex_(true), bib_key_(key), entry_type_(type), info_()
 199 {}
 200
 201
 202 bool BibTeXInfo::hasField(docstring const & field) const
 203 {
 204         return count(field) == 1;
 205 }
 206
 207
 208 docstring const BibTeXInfo::getAbbreviatedAuthor() const
 209 {
 210         if (!is_bibtex_) {
 211                 docstring const opt = trim(operator[]("label"));
 212                 if (opt.empty())
 213                         return docstring();
 214
 215                 docstring authors;
 216                 split(opt, authors, '(');
 217                 return authors;
 218         }
 219
 220         docstring author = operator[]("author");
 221         if (author.empty()) {
 222                 author = operator[]("editor");
 223                 if (author.empty())
 224                         return bib_key_;
 225         }
 226
 227         // OK, we've got some names. Let's format them.
 228         // Try to split the author list on " and "
 229         vector<docstring> const authors =
 230                 getVectorFromString(author, from_ascii(" and "));
 231
 232         if (authors.size() == 2)
 233                 return bformat(_("%1$s and %2$s"),
 234                         familyName(authors[0]), familyName(authors[1]));
 235
 236         if (authors.size() > 2)
 237                 return bformat(_("%1$s et al."), familyName(authors[0]));
 238
 239         return familyName(authors[0]);
 240 }
 241
 242
 243 docstring const BibTeXInfo::getYear() const
 244 {
 245         if (is_bibtex_)
 246                 return operator[]("year");
 247
 248         docstring const opt = trim(operator[]("label"));
 249         if (opt.empty())
 250                 return docstring();
 251
 252         docstring authors;
 253         docstring const tmp = split(opt, authors, '(');
 254         docstring year;
 255         split(tmp, year, ')');
 256         return year;
 257 }
 258
 259
 260 docstring const BibTeXInfo::getXRef() const
 261 {
 262         if (!is_bibtex_)
 263                 return docstring();
 264         return operator[]("crossref");
 265 }
 266
 267
 268 docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref) const
 269 {
 270         if (!info_.empty())
 271                 return info_;
 272
 273         if (!is_bibtex_) {
 274                 BibTeXInfo::const_iterator it = find(from_ascii("ref"));
 275                 info_ = it->second;
 276                 return info_;
 277         }
 278
 279         // FIXME
 280         // This could be made a lot better using the entry_type_
 281         // field to customize the output based upon entry type.
 282
 283         // Search for all possible "required" fields
 284         docstring author = getValueForKey("author", xref);
 285         if (author.empty())
 286                 author = getValueForKey("editor", xref);
 287
 288         docstring year   = getValueForKey("year", xref);
 289         docstring title  = getValueForKey("title", xref);
 290         docstring docLoc = getValueForKey("pages", xref);
 291         if (docLoc.empty()) {
 292                 docLoc = getValueForKey("chapter", xref);
 293                 if (!docLoc.empty())
 294                         docLoc = _("Ch. ") + docLoc;
 295         }       else {
 296                 docLoc = _("pp. ") + docLoc;
 297         }
 298
 299         docstring media = getValueForKey("journal", xref);
 300         if (media.empty()) {
 301                 media = getValueForKey("publisher", xref);
 302                 if (media.empty()) {
 303                         media = getValueForKey("school", xref);
 304                         if (media.empty())
 305                                 media = getValueForKey("institution");
 306                 }
 307         }
 308         docstring volume = getValueForKey("volume", xref);
 309
 310         odocstringstream result;
 311         if (!author.empty())
 312                 result << author << ", ";
 313         if (!title.empty())
 314                 result << title;
 315         if (!media.empty())
 316                 result << ", " << media;
 317         if (!year.empty())
 318                 result << " (" << year << ")";
 319         if (!docLoc.empty())
 320                 result << ", " << docLoc;
 321
 322         docstring const result_str = rtrim(result.str());
 323         if (!result_str.empty()) {
 324                 info_ = convertLaTeXCommands(result_str);
 325                 return info_;
 326         }
 327
 328         // This should never happen (or at least be very unusual!)
 329         static docstring e = docstring();
 330         return e;
 331 }
 332
 333
 334 docstring const & BibTeXInfo::operator[](docstring const & field) const
 335 {
 336         BibTeXInfo::const_iterator it = find(field);
 337         if (it != end())
 338                 return it->second;
 339         static docstring const empty_value = docstring();
 340         return empty_value;
 341 }
 342
 343
 344 docstring const & BibTeXInfo::operator[](string const & field) const
 345 {
 346         return operator[](from_ascii(field));
 347 }
 348
 349
 350 docstring BibTeXInfo::getValueForKey(string const & key,
 351                 BibTeXInfo const * const xref) const
 352 {
 353         docstring const ret = operator[](key);
 354         if (!ret.empty() || !xref)
 355                 return ret;
 356         return (*xref)[key];
 357 }
 358
 359
 360 //////////////////////////////////////////////////////////////////////
 361 //
 362 // BiblioInfo
 363 //
 364 //////////////////////////////////////////////////////////////////////
 365
 366 namespace {
 367 // A functor for use with sort, leading to case insensitive sorting
 368         class compareNoCase: public binary_function<docstring, docstring, bool>
 369         {
 370                 public:
 371                         bool operator()(docstring const & s1, docstring const & s2) const {
 372                                 return compare_no_case(s1, s2) < 0;
 373                         }
 374         };
 375 } // namespace anon
 376
 377
 378 vector<docstring> const BiblioInfo::getKeys() const
 379 {
 380         vector<docstring> bibkeys;
 381         BiblioInfo::const_iterator it  = begin();
 382         for (; it != end(); ++it)
 383                 bibkeys.push_back(it->first);
 384         sort(bibkeys.begin(), bibkeys.end(), compareNoCase());
 385         return bibkeys;
 386 }
 387
 388
 389 vector<docstring> const BiblioInfo::getFields() const
 390 {
 391         vector<docstring> bibfields;
 392         set<docstring>::const_iterator it = field_names_.begin();
 393         set<docstring>::const_iterator end = field_names_.end();
 394         for (; it != end; ++it)
 395                 bibfields.push_back(*it);
 396         sort(bibfields.begin(), bibfields.end());
 397         return bibfields;
 398 }
 399
 400
 401 vector<docstring> const BiblioInfo::getEntries() const
 402 {
 403         vector<docstring> bibentries;
 404         set<docstring>::const_iterator it = entry_types_.begin();
 405         set<docstring>::const_iterator end = entry_types_.end();
 406         for (; it != end; ++it)
 407                 bibentries.push_back(*it);
 408         sort(bibentries.begin(), bibentries.end());
 409         return bibentries;
 410 }
 411
 412
 413 docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key) const
 414 {
 415         BiblioInfo::const_iterator it = find(key);
 416         if (it == end())
 417                 return docstring();
 418         BibTeXInfo const & data = it->second;
 419         return data.getAbbreviatedAuthor();
 420 }
 421
 422
 423 docstring const BiblioInfo::getYear(docstring const & key) const
 424 {
 425         BiblioInfo::const_iterator it = find(key);
 426         if (it == end())
 427                 return docstring();
 428         BibTeXInfo const & data = it->second;
 429         docstring year = data.getYear();
 430         if (!year.empty())
 431                 return year;
 432         // let's try the crossref
 433         docstring const xref = data.getXRef();
 434         if (xref.empty())
 435                 return _("No year"); // no luck
 436         BiblioInfo::const_iterator const xrefit = find(xref);
 437         if (xrefit == end())
 438                 return _("No year"); // no luck again
 439         BibTeXInfo const & xref_data = xrefit->second;
 440         return xref_data.getYear();
 441         return data.getYear();
 442 }
 443
 444
 445 docstring const BiblioInfo::getInfo(docstring const & key) const
 446 {
 447         BiblioInfo::const_iterator it = find(key);
 448         if (it == end())
 449                 return docstring();
 450         BibTeXInfo const & data = it->second;
 451         BibTeXInfo const * xrefptr = 0;
 452         docstring const xref = data.getXRef();
 453         if (!xref.empty()) {
 454                 BiblioInfo::const_iterator const xrefit = find(xref);
 455                 if (xrefit != end())
 456                         xrefptr = &(xrefit->second);
 457         }
 458         return data.getInfo(xrefptr);
 459 }
 460
 461
 462 vector<docstring> const BiblioInfo::getCiteStrings(
 463         docstring const & key, Buffer const & buf) const
 464 {
 465         CiteEngine const engine = buf.params().citeEngine();
 466         if (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL)
 467                 return getNumericalStrings(key, buf);
 468         else
 469                 return getAuthorYearStrings(key, buf);
 470 }
 471
 472
 473 vector<docstring> const BiblioInfo::getNumericalStrings(
 474         docstring const & key, Buffer const & buf) const
 475 {
 476         if (empty())
 477                 return vector<docstring>();
 478
 479         docstring const author = getAbbreviatedAuthor(key);
 480         docstring const year   = getYear(key);
 481         if (author.empty() || year.empty())
 482                 return vector<docstring>();
 483
 484         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 485
 486         vector<docstring> vec(styles.size());
 487         for (size_t i = 0; i != vec.size(); ++i) {
 488                 docstring str;
 489
 490                 switch (styles[i]) {
 491                         case CITE:
 492                         case CITEP:
 493                                 str = from_ascii("[#ID]");
 494                                 break;
 495
 496                         case NOCITE:
 497                                 str = _("Add to bibliography only.");
 498                                 break;
 499
 500                         case CITET:
 501                                 str = author + " [#ID]";
 502                                 break;
 503
 504                         case CITEALT:
 505                                 str = author + " #ID";
 506                                 break;
 507
 508                         case CITEALP:
 509                                 str = from_ascii("#ID");
 510                                 break;
 511
 512                         case CITEAUTHOR:
 513                                 str = author;
 514                                 break;
 515
 516                         case CITEYEAR:
 517                                 str = year;
 518                                 break;
 519
 520                         case CITEYEARPAR:
 521                                 str = '(' + year + ')';
 522                                 break;
 523                 }
 524
 525                 vec[i] = str;
 526         }
 527
 528         return vec;
 529 }
 530
 531
 532 vector<docstring> const BiblioInfo::getAuthorYearStrings(
 533         docstring const & key, Buffer const & buf) const
 534 {
 535         if (empty())
 536                 return vector<docstring>();
 537
 538         docstring const author = getAbbreviatedAuthor(key);
 539         docstring const year   = getYear(key);
 540         if (author.empty() || year.empty())
 541                 return vector<docstring>();
 542
 543         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 544
 545         vector<docstring> vec(styles.size());
 546         for (size_t i = 0; i != vec.size(); ++i) {
 547                 docstring str;
 548
 549                 switch (styles[i]) {
 550                         case CITE:
 551                 // jurabib only: Author/Annotator
 552                 // (i.e. the "before" field, 2nd opt arg)
 553                                 str = author + "/<" + _("before") + '>';
 554                                 break;
 555
 556                         case NOCITE:
 557                                 str = _("Add to bibliography only.");
 558                                 break;
 559
 560                         case CITET:
 561                                 str = author + " (" + year + ')';
 562                                 break;
 563
 564                         case CITEP:
 565                                 str = '(' + author + ", " + year + ')';
 566                                 break;
 567
 568                         case CITEALT:
 569                                 str = author + ' ' + year ;
 570                                 break;
 571
 572                         case CITEALP:
 573                                 str = author + ", " + year ;
 574                                 break;
 575
 576                         case CITEAUTHOR:
 577                                 str = author;
 578                                 break;
 579
 580                         case CITEYEAR:
 581                                 str = year;
 582                                 break;
 583
 584                         case CITEYEARPAR:
 585                                 str = '(' + year + ')';
 586                                 break;
 587                 }
 588                 vec[i] = str;
 589         }
 590         return vec;
 591 }
 592
 593
 594 void BiblioInfo::mergeBiblioInfo(BiblioInfo const & info)
 595 {
 596         bimap_.insert(info.begin(), info.end());
 597 }
 598
 599
 600 //////////////////////////////////////////////////////////////////////
 601 //
 602 // CitationStyle
 603 //
 604 //////////////////////////////////////////////////////////////////////
 605
 606 namespace {
 607
 608
 609 char const * const citeCommands[] = {
 610         "cite", "citet", "citep", "citealt", "citealp",
 611         "citeauthor", "citeyear", "citeyearpar", "nocite" };
 612
 613 unsigned int const nCiteCommands =
 614                 sizeof(citeCommands) / sizeof(char *);
 615
 616 CiteStyle const citeStylesArray[] = {
 617         CITE, CITET, CITEP, CITEALT, CITEALP,
 618         CITEAUTHOR, CITEYEAR, CITEYEARPAR, NOCITE };
 619
 620 unsigned int const nCiteStyles =
 621                 sizeof(citeStylesArray) / sizeof(CiteStyle);
 622
 623 CiteStyle const citeStylesFull[] = {
 624         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 625
 626 unsigned int const nCiteStylesFull =
 627                 sizeof(citeStylesFull) / sizeof(CiteStyle);
 628
 629 CiteStyle const citeStylesUCase[] = {
 630         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 631
 632 unsigned int const nCiteStylesUCase =
 633         sizeof(citeStylesUCase) / sizeof(CiteStyle);
 634
 635 } // namespace anon
 636
 637
 638 CitationStyle citationStyleFromString(string const & command)
 639 {
 640         CitationStyle s;
 641         if (command.empty())
 642                 return s;
 643
 644         string cmd = command;
 645         if (cmd[0] == 'C') {
 646                 s.forceUpperCase = true;
 647                 cmd[0] = 'c';
 648         }
 649
 650         size_t const n = cmd.size() - 1;
 651         if (cmd != "cite" && cmd[n] == '*') {
 652                 s.full = true;
 653                 cmd = cmd.substr(0, n);
 654         }
 655
 656         char const * const * const last = citeCommands + nCiteCommands;
 657         char const * const * const ptr = find(citeCommands, last, cmd);
 658
 659         if (ptr != last) {
 660                 size_t idx = ptr - citeCommands;
 661                 s.style = citeStylesArray[idx];
 662         }
 663         return s;
 664 }
 665
 666
 667 string citationStyleToString(const CitationStyle & s)
 668 {
 669         string cite = citeCommands[s.style];
 670         if (s.full) {
 671                 CiteStyle const * last = citeStylesFull + nCiteStylesFull;
 672                 if (find(citeStylesFull, last, s.style) != last)
 673                         cite += '*';
 674         }
 675
 676         if (s.forceUpperCase) {
 677                 CiteStyle const * last = citeStylesUCase + nCiteStylesUCase;
 678                 if (find(citeStylesUCase, last, s.style) != last)
 679                         cite[0] = 'C';
 680         }
 681
 682         return cite;
 683 }
 684
 685 vector<CiteStyle> citeStyles(CiteEngine engine)
 686 {
 687         unsigned int nStyles = 0;
 688         unsigned int start = 0;
 689
 690         switch (engine) {
 691                 case ENGINE_BASIC:
 692                         nStyles = 2;
 693                         start = 0;
 694                         break;
 695                 case ENGINE_NATBIB_AUTHORYEAR:
 696                 case ENGINE_NATBIB_NUMERICAL:
 697                         nStyles = nCiteStyles - 1;
 698                         start = 1;
 699                         break;
 700                 case ENGINE_JURABIB:
 701                         nStyles = nCiteStyles;
 702                         start = 0;
 703                         break;
 704         }
 705
 706         vector<CiteStyle> styles(nStyles);
 707         size_t i = 0;
 708         int j = start;
 709         for (; i != styles.size(); ++i, ++j)
 710                 styles[i] = citeStylesArray[j];
 711
 712         return styles;
 713 }
 714
 715 } // namespace lyx
 716