src/BiblioInfo.cpp

   1 /**
   2  * \file BiblioInfo.cpp
   3  * This file is part of LyX, the document processor.
   4  * Licence details can be found in the file COPYING.
   5  *
   6  * \author Angus Leeming
   7  * \author Herbert Voß
   8  * \author Richard Heck
   9  *
  10  * Full author contact details are available in file CREDITS.
  11  */
  12
  13 #include <config.h>
  14
  15 #include "BiblioInfo.h"
  16 #include "Buffer.h"
  17 #include "BufferParams.h"
  18 #include "buffer_funcs.h"
  19 #include "Encoding.h"
  20 #include "InsetIterator.h"
  21 #include "Paragraph.h"
  22 #include "TocBackend.h"
  23
  24 #include "insets/Inset.h"
  25 #include "insets/InsetBibitem.h"
  26 #include "insets/InsetBibtex.h"
  27 #include "insets/InsetInclude.h"
  28
  29 #include "support/convert.h"
  30 #include "support/debug.h"
  31 #include "support/docstream.h"
  32 #include "support/gettext.h"
  33 #include "support/lassert.h"
  34 #include "support/lstrings.h"
  35 #include "support/textutils.h"
  36
  37 #include "boost/regex.hpp"
  38
  39 #include <set>
  40
  41 using namespace std;
  42 using namespace lyx::support;
  43
  44
  45 namespace lyx {
  46
  47 namespace {
  48
  49 // gets the "family name" from an author-type string
  50 docstring familyName(docstring const & name)
  51 {
  52         if (name.empty())
  53                 return docstring();
  54
  55         // first we look for a comma, and take the last name to be everything
  56         // preceding the right-most one, so that we also get the "jr" part.
  57         docstring::size_type idx = name.rfind(',');
  58         if (idx != docstring::npos)
  59                 return ltrim(name.substr(0, idx));
  60
  61         // OK, so now we want to look for the last name. We're going to
  62         // include the "von" part. This isn't perfect.
  63         // Split on spaces, to get various tokens.
  64         vector<docstring> pieces = getVectorFromString(name, from_ascii(" "));
  65         // If we only get two, assume the last one is the last name
  66         if (pieces.size() <= 2)
  67                 return pieces.back();
  68
  69         // Now we look for the first token that begins with a lower case letter.
  70         vector<docstring>::const_iterator it = pieces.begin();
  71         vector<docstring>::const_iterator en = pieces.end();
  72         for (; it != en; ++it) {
  73                 if ((*it).size() == 0)
  74                         continue;
  75                 char_type const c = (*it)[0];
  76                 if (isLower(c))
  77                         break;
  78         }
  79
  80         if (it == en) // we never found a "von"
  81                 return pieces.back();
  82
  83         // reconstruct what we need to return
  84         docstring retval;
  85         bool first = true;
  86         for (; it != en; ++it) {
  87                 if (!first)
  88                         retval += " ";
  89                 else
  90                         first = false;
  91                 retval += *it;
  92         }
  93         return retval;
  94 }
  95
  96 // converts a string containing LaTeX commands into unicode
  97 // for display.
  98 docstring convertLaTeXCommands(docstring const & str)
  99 {
 100         docstring val = str;
 101         docstring ret;
 102
 103         bool scanning_cmd = false;
 104         bool scanning_math = false;
 105         bool escaped = false; // used to catch \$, etc.
 106         while (val.size()) {
 107                 char_type const ch = val[0];
 108
 109                 // if we're scanning math, we output everything until we
 110                 // find an unescaped $, at which point we break out.
 111                 if (scanning_math) {
 112                         if (escaped)
 113                                 escaped = false;
 114                         else if (ch == '\\')
 115                                 escaped = true;
 116                         else if (ch == '$')
 117                                 scanning_math = false;
 118                         ret += ch;
 119                         val = val.substr(1);
 120                         continue;
 121                 }
 122
 123                 // if we're scanning a command name, then we just
 124                 // discard characters until we hit something that
 125                 // isn't alpha.
 126                 if (scanning_cmd) {
 127                         if (isAlphaASCII(ch)) {
 128                                 val = val.substr(1);
 129                                 escaped = false;
 130                                 continue;
 131                         }
 132                         // so we're done with this command.
 133                         // now we fall through and check this character.
 134                         scanning_cmd = false;
 135                 }
 136
 137                 // was the last character a \? If so, then this is something like:
 138                 // \\ or \$, so we'll just output it. That's probably not always right...
 139                 if (escaped) {
 140                         // exception: output \, as THIN SPACE
 141                         if (ch == ',')
 142                                 ret.push_back(0x2009);
 143                         else
 144                                 ret += ch;
 145                         val = val.substr(1);
 146                         escaped = false;
 147                         continue;
 148                 }
 149
 150                 if (ch == '$') {
 151                         ret += ch;
 152                         val = val.substr(1);
 153                         scanning_math = true;
 154                         continue;
 155                 }
 156
 157                 // we just ignore braces
 158                 if (ch == '{' || ch == '}') {
 159                         val = val.substr(1);
 160                         continue;
 161                 }
 162
 163                 // we're going to check things that look like commands, so if
 164                 // this doesn't, just output it.
 165                 if (ch != '\\') {
 166                         ret += ch;
 167                         val = val.substr(1);
 168                         continue;
 169                 }
 170
 171                 // ok, could be a command of some sort
 172                 // let's see if it corresponds to some unicode
 173                 // unicodesymbols has things in the form: \"{u},
 174                 // whereas we may see things like: \"u. So we'll
 175                 // look for that and change it, if necessary.
 176                 static boost::regex const reg("^\\\\\\W\\w");
 177                 if (boost::regex_search(to_utf8(val), reg)) {
 178                         val.insert(3, from_ascii("}"));
 179                         val.insert(2, from_ascii("{"));
 180                 }
 181                 docstring rem;
 182                 docstring const cnvtd = Encodings::fromLaTeXCommand(val, rem,
 183                                                         Encodings::TEXT_CMD);
 184                 if (!cnvtd.empty()) {
 185                         // it did, so we'll take that bit and proceed with what's left
 186                         ret += cnvtd;
 187                         val = rem;
 188                         continue;
 189                 }
 190                 // it's a command of some sort
 191                 scanning_cmd = true;
 192                 escaped = true;
 193                 val = val.substr(1);
 194         }
 195         return ret;
 196 }
 197
 198 // these are used in the expandFormat() routine, etc.
 199
 200 static string const pp_text   = N_("pp.");
 201 static string const ed_text   = N_("ed.");
 202 static string const edby_text = N_("ed. by");
 203
 204 } // anon namespace
 205
 206
 207 //////////////////////////////////////////////////////////////////////
 208 //
 209 // BibTeXInfo
 210 //
 211 //////////////////////////////////////////////////////////////////////
 212
 213 BibTeXInfo::BibTeXInfo(docstring const & key, docstring const & type)
 214         : is_bibtex_(true), bib_key_(key), entry_type_(type), info_(),
 215           modifier_(0)
 216 {}
 217
 218
 219 docstring const BibTeXInfo::getAbbreviatedAuthor() const
 220 {
 221         if (!is_bibtex_) {
 222                 docstring const opt = label();
 223                 if (opt.empty())
 224                         return docstring();
 225
 226                 docstring authors;
 227                 docstring const remainder = trim(split(opt, authors, '('));
 228                 if (remainder.empty())
 229                         // in this case, we didn't find a "(",
 230                         // so we don't have author (year)
 231                         return docstring();
 232                 return authors;
 233         }
 234
 235         docstring author = convertLaTeXCommands(operator[]("author"));
 236         if (author.empty()) {
 237                 author = convertLaTeXCommands(operator[]("editor"));
 238                 if (author.empty())
 239                         return bib_key_;
 240         }
 241
 242         // FIXME Move this to a separate routine that can
 243         // be called from elsewhere.
 244         //
 245         // OK, we've got some names. Let's format them.
 246         // Try to split the author list on " and "
 247         vector<docstring> const authors =
 248                 getVectorFromString(author, from_ascii(" and "));
 249
 250         if (authors.size() == 2)
 251                 return bformat(_("%1$s and %2$s"),
 252                         familyName(authors[0]), familyName(authors[1]));
 253
 254         if (authors.size() > 2)
 255                 return bformat(_("%1$s et al."), familyName(authors[0]));
 256
 257         return familyName(authors[0]);
 258 }
 259
 260
 261 docstring const BibTeXInfo::getYear() const
 262 {
 263         if (is_bibtex_)
 264                 return operator[]("year");
 265
 266         docstring const opt = label();
 267         if (opt.empty())
 268                 return docstring();
 269
 270         docstring authors;
 271         docstring tmp = split(opt, authors, '(');
 272         if (tmp.empty())
 273                 // we don't have author (year)
 274                 return docstring();
 275         docstring year;
 276         tmp = split(tmp, year, ')');
 277         return year;
 278 }
 279
 280
 281 docstring const BibTeXInfo::getXRef() const
 282 {
 283         if (!is_bibtex_)
 284                 return docstring();
 285         return operator[]("crossref");
 286 }
 287
 288
 289 namespace {
 290         string parseOptions(string const & format, string & optkey,
 291                         string & ifpart, string & elsepart);
 292
 293         /// Calls parseOptions to deal with an embedded option, such as:
 294         ///   {%number%[[, no.~%number%]]}
 295         /// which must appear at the start of format. ifelsepart gets the
 296         /// whole of the option, and we return what's left after the option.
 297         /// we return format if there is an error.
 298         string parseEmbeddedOption(string const & format, string & ifelsepart)
 299         {
 300                 LASSERT(format[0] == '{' && format[1] == '%', return format);
 301                 string optkey;
 302                 string ifpart;
 303                 string elsepart;
 304                 string const rest = parseOptions(format, optkey, ifpart, elsepart);
 305                 if (format == rest) { // parse error
 306                         LYXERR0("ERROR! Couldn't parse `" << format <<"'.");
 307                         return format;
 308                 }
 309                 LASSERT(rest.size() <= format.size(), /* */);
 310                 ifelsepart = format.substr(0, format.size() - rest.size());
 311                 return rest;
 312         }
 313
 314
 315         // Gets a "clause" from a format string, where the clause is
 316         // delimited by '[[' and ']]'. Returns what is left after the
 317         // clause is removed, and returns format if there is an error.
 318         string getClause(string const & format, string & clause)
 319         {
 320                 string fmt = format;
 321                 // remove '[['
 322                 fmt = fmt.substr(2);
 323                 // we'll remove characters from the front of fmt as we
 324                 // deal with them
 325                 while (fmt.size()) {
 326                         if (fmt[0] == ']' && fmt.size() > 1 && fmt[1] == ']') {
 327                           // that's the end
 328                                 fmt = fmt.substr(2);
 329                                 break;
 330                         }
 331                         // check for an embedded option
 332                         if (fmt[0] == '{' && fmt.size() > 1 && fmt[1] == '%') {
 333                                 string part;
 334                                 string const rest = parseEmbeddedOption(fmt, part);
 335                                 if (fmt == rest) {
 336                                         LYXERR0("ERROR! Couldn't parse `" << format <<"'.");
 337                                         return format;
 338                                 }
 339                                 clause += part;
 340                                 fmt = rest;
 341                         } else { // it's just a normal character
 342                                 clause += fmt[0];
 343                                 fmt = fmt.substr(1);
 344                         }
 345                 }
 346                 return fmt;
 347         }
 348
 349
 350         /// parse an options string, which must appear at the start of the
 351         /// format parameter. puts the parsed bits in optkey, ifpart, and
 352         /// elsepart and returns what's left after the option is removed.
 353         /// if there's an error, it returns format itself.
 354         string parseOptions(string const & format, string & optkey,
 355                         string & ifpart, string & elsepart)
 356         {
 357                 LASSERT(format[0] == '{' && format[1] == '%', return format);
 358                 // strip '{%'
 359                 string fmt = format.substr(2);
 360                 size_t pos = fmt.find('%'); // end of key
 361                 if (pos == string::npos) {
 362                         LYXERR0("Error parsing  `" << format <<"'. Can't find end of key.");
 363                         return format;
 364                 }
 365                 optkey = fmt.substr(0,pos);
 366                 fmt = fmt.substr(pos + 1);
 367                 // [[format]] should be next
 368                 if (fmt[0] != '[' || fmt[1] != '[') {
 369                         LYXERR0("Error parsing  `" << format <<"'. Can't find '[[' after key.");
 370                         return format;
 371                 }
 372
 373                 string curfmt = fmt;
 374                 fmt = getClause(curfmt, ifpart);
 375                 if (fmt == curfmt) {
 376                         LYXERR0("Error parsing  `" << format <<"'. Couldn't get if clause.");
 377                         return format;
 378                 }
 379
 380                 if (fmt[0] == '}') // we're done, no else clause
 381                         return fmt.substr(1);
 382
 383                 // else part should follow
 384                 if (fmt[0] != '[' || fmt[1] != '[') {
 385                         LYXERR0("Error parsing  `" << format <<"'. Can't find else clause.");
 386                         return format;
 387                 }
 388
 389                 curfmt = fmt;
 390                 fmt = getClause(curfmt, elsepart);
 391                 // we should be done
 392                 if (fmt == curfmt || fmt[0] != '}') {
 393                         LYXERR0("Error parsing  `" << format <<"'. Can't find end of option.");
 394                         return format;
 395                 }
 396                 return fmt.substr(1);
 397 }
 398
 399 } // anon namespace
 400
 401
 402 docstring BibTeXInfo::expandFormat(string const & format,
 403                 BibTeXInfo const * const xref, bool richtext) const
 404 {
 405         // return value
 406         docstring ret;
 407         string key;
 408         bool scanning_key = false;
 409         bool scanning_rich = false;
 410
 411         string fmt = format;
 412         // we'll remove characters from the front of fmt as we
 413         // deal with them
 414         while (fmt.size()) {
 415                 char_type thischar = fmt[0];
 416                 if (thischar == '%') {
 417                         // beginning or end of key
 418                         if (scanning_key) {
 419                                 // end of key
 420                                 scanning_key = false;
 421                                 // so we replace the key with its value, which may be empty
 422                                 if (key == "pp_text")
 423                                         ret += _(pp_text);
 424                                 else if (key == "ed_text")
 425                                         ret += _(ed_text);
 426                                 else if(key == "edby_text")
 427                                         ret += _(edby_text);
 428                                 else {
 429                                         docstring const val = getValueForKey(key, xref);
 430                                         ret += val;
 431                                 }
 432                                 key.clear();
 433                         } else {
 434                                 // beginning of key
 435                                 scanning_key = true;
 436                         }
 437                 }
 438                 else if (thischar == '{') {
 439                         // beginning of option?
 440                         if (scanning_key) {
 441                                 LYXERR0("ERROR: Found `{' when scanning key in `" << format << "'.");
 442                                 return _("ERROR!");
 443                         }
 444                         if (fmt.size() > 1) {
 445                                 if (fmt[1] == '%') {
 446                                         // it is the beginning of an optional format
 447                                         string optkey;
 448                                         string ifpart;
 449                                         string elsepart;
 450                                         string const newfmt =
 451                                                 parseOptions(fmt, optkey, ifpart, elsepart);
 452                                         if (newfmt == fmt) // parse error
 453                                                 return _("ERROR!");
 454                                         fmt = newfmt;
 455                                         docstring const val = getValueForKey(optkey, xref);
 456                                         if (!val.empty())
 457                                                 ret += expandFormat(ifpart, xref, richtext);
 458                                         else if (!elsepart.empty())
 459                                                 ret += expandFormat(elsepart, xref, richtext);
 460                                         // fmt will have been shortened for us already
 461                                         continue;
 462                                 }
 463                                 if (fmt[1] == '!') {
 464                                         // beginning of rich text
 465                                         scanning_rich = true;
 466                                         fmt = fmt.substr(2);
 467                                         continue;
 468                                 }
 469                         }
 470                         // we are here if the '{' was at the end of the format. hmm.
 471                         ret += thischar;
 472                 }
 473                 else if (scanning_rich && thischar == '!'
 474                          && fmt.size() > 1 && fmt[1] == '}') {
 475                         // end of rich text
 476                         scanning_rich = false;
 477                         fmt = fmt.substr(2);
 478                         continue;
 479                 }
 480                 else if (scanning_key)
 481                         key += thischar;
 482                 else if (richtext || !scanning_rich)
 483                         ret += thischar;
 484                 // else the character is discarded, which will happen only if
 485                 // richtext == false and we are scanning rich text
 486                 fmt = fmt.substr(1);
 487         } // for loop
 488         if (scanning_key) {
 489                 LYXERR0("Never found end of key in `" << format << "'!");
 490                 return _("ERROR!");
 491         }
 492         if (scanning_rich) {
 493                 LYXERR0("Never found end of rich text in `" << format << "'!");
 494                 return _("ERROR!");
 495         }
 496         return ret;
 497 }
 498
 499
 500 namespace {
 501
 502 // FIXME These would be better read from a file, so that they
 503 // could be customized.
 504
 505         static string articleFormat = "%author%, \"%title%\", {!<i>!}%journal%{!</i>!} {%volume%[[ %volume%{%number%[[, %number%]]}]]} (%year%){%pages%[[, %pp_text% %pages%]]}.{%note%[[ %note%]]}";
 506
 507         static string bookFormat = "{%author%[[%author%]][[%editor%, %ed_text%]]}, {!<i>!}%title%{!</i>!}{%volume%[[ vol. %volume%]][[{%number%[[no. %number%]]}]]}{%edition%[[%edition%]]} ({%address%[[%address%: ]]}%publisher%, %year%).{%note%[[ %note%]]}";
 508
 509         static string inSomething = "%author%, \"%title%\", in{%editor%[[ %editor%, %ed_text%,]]} {!<i>!}%booktitle%{!</i>!}{%volume%[[ vol. %volume%]][[{%number%[[no. %number%]]}]]}{%edition%[[%edition%]]} ({%address%[[%address%: ]]}%publisher%, %year%){%pages%[[, %pp_text% %pages%]]}.{%note%[[ %note%]]}";
 510
 511         static string thesis = "%author%, %title% ({%address%[[%address%: ]]}%school%, %year%).{%note%[[ %note%]]}";
 512
 513         static string defaultFormat = "{%author%[[%author%, ]][[{%editor%[[%editor%, %ed_text%, ]]}]]}\"%title%\"{%journal%[[, {!<i>!}%journal%{!</i>!}]][[{%publisher%[[, %publisher%]][[{%institution%[[, %institution%]]}]]}]]}{%year%[[ (%year%)]]}{%pages%[[, %pages%]]}.";
 514
 515 }
 516
 517 docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref,
 518         bool richtext) const
 519 {
 520         if (!info_.empty())
 521                 return info_;
 522
 523         if (!is_bibtex_) {
 524                 BibTeXInfo::const_iterator it = find(from_ascii("ref"));
 525                 info_ = it->second;
 526                 return info_;
 527         }
 528
 529         if (entry_type_ == "article")
 530                 info_ = expandFormat(articleFormat, xref, richtext);
 531         else if (entry_type_ == "book")
 532                 info_ = expandFormat(bookFormat, xref, richtext);
 533         else if (entry_type_.substr(0,2) == "in")
 534                 info_ = expandFormat(inSomething, xref, richtext);
 535         else if (entry_type_ == "phdthesis" || entry_type_ == "mastersthesis")
 536                 info_ = expandFormat(thesis, xref, richtext);
 537         else
 538                 info_ = expandFormat(defaultFormat, xref, richtext);
 539
 540         if (!info_.empty())
 541                 info_ = convertLaTeXCommands(info_);
 542         return info_;
 543 }
 544
 545
 546 docstring const & BibTeXInfo::operator[](docstring const & field) const
 547 {
 548         BibTeXInfo::const_iterator it = find(field);
 549         if (it != end())
 550                 return it->second;
 551         static docstring const empty_value = docstring();
 552         return empty_value;
 553 }
 554
 555
 556 docstring const & BibTeXInfo::operator[](string const & field) const
 557 {
 558         return operator[](from_ascii(field));
 559 }
 560
 561
 562 docstring BibTeXInfo::getValueForKey(string const & key,
 563                 BibTeXInfo const * const xref) const
 564 {
 565         docstring const ret = operator[](key);
 566         if (!ret.empty() || !xref)
 567                 return ret;
 568         return (*xref)[key];
 569 }
 570
 571
 572 //////////////////////////////////////////////////////////////////////
 573 //
 574 // BiblioInfo
 575 //
 576 //////////////////////////////////////////////////////////////////////
 577
 578 namespace {
 579 // A functor for use with sort, leading to case insensitive sorting
 580         class compareNoCase: public binary_function<docstring, docstring, bool>
 581         {
 582                 public:
 583                         bool operator()(docstring const & s1, docstring const & s2) const {
 584                                 return compare_no_case(s1, s2) < 0;
 585                         }
 586         };
 587 } // namespace anon
 588
 589
 590 vector<docstring> const BiblioInfo::getKeys() const
 591 {
 592         vector<docstring> bibkeys;
 593         BiblioInfo::const_iterator it  = begin();
 594         for (; it != end(); ++it)
 595                 bibkeys.push_back(it->first);
 596         sort(bibkeys.begin(), bibkeys.end(), compareNoCase());
 597         return bibkeys;
 598 }
 599
 600
 601 vector<docstring> const BiblioInfo::getFields() const
 602 {
 603         vector<docstring> bibfields;
 604         set<docstring>::const_iterator it = field_names_.begin();
 605         set<docstring>::const_iterator end = field_names_.end();
 606         for (; it != end; ++it)
 607                 bibfields.push_back(*it);
 608         sort(bibfields.begin(), bibfields.end());
 609         return bibfields;
 610 }
 611
 612
 613 vector<docstring> const BiblioInfo::getEntries() const
 614 {
 615         vector<docstring> bibentries;
 616         set<docstring>::const_iterator it = entry_types_.begin();
 617         set<docstring>::const_iterator end = entry_types_.end();
 618         for (; it != end; ++it)
 619                 bibentries.push_back(*it);
 620         sort(bibentries.begin(), bibentries.end());
 621         return bibentries;
 622 }
 623
 624
 625 docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key) const
 626 {
 627         BiblioInfo::const_iterator it = find(key);
 628         if (it == end())
 629                 return docstring();
 630         BibTeXInfo const & data = it->second;
 631         return data.getAbbreviatedAuthor();
 632 }
 633
 634
 635 docstring const BiblioInfo::getCiteNumber(docstring const & key) const
 636 {
 637         BiblioInfo::const_iterator it = find(key);
 638         if (it == end())
 639                 return docstring();
 640         BibTeXInfo const & data = it->second;
 641         return data.citeNumber();
 642 }
 643
 644
 645 docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier) const
 646 {
 647         BiblioInfo::const_iterator it = find(key);
 648         if (it == end())
 649                 return docstring();
 650         BibTeXInfo const & data = it->second;
 651         docstring year = data.getYear();
 652         if (year.empty()) {
 653                 // let's try the crossref
 654                 docstring const xref = data.getXRef();
 655                 if (xref.empty())
 656                         return _("No year"); // no luck
 657                 BiblioInfo::const_iterator const xrefit = find(xref);
 658                 if (xrefit == end())
 659                         return _("No year"); // no luck again
 660                 BibTeXInfo const & xref_data = xrefit->second;
 661                 year = xref_data.getYear();
 662         }
 663         if (use_modifier && data.modifier() != 0)
 664                 year += data.modifier();
 665         return year;
 666 }
 667
 668
 669 docstring const BiblioInfo::getInfo(docstring const & key, bool richtext) const
 670 {
 671         BiblioInfo::const_iterator it = find(key);
 672         if (it == end())
 673                 return docstring();
 674         BibTeXInfo const & data = it->second;
 675         BibTeXInfo const * xrefptr = 0;
 676         docstring const xref = data.getXRef();
 677         if (!xref.empty()) {
 678                 BiblioInfo::const_iterator const xrefit = find(xref);
 679                 if (xrefit != end())
 680                         xrefptr = &(xrefit->second);
 681         }
 682         return data.getInfo(xrefptr, richtext);
 683 }
 684
 685
 686 bool BiblioInfo::isBibtex(docstring const & key) const
 687 {
 688         BiblioInfo::const_iterator it = find(key);
 689         if (it == end())
 690                 return false;
 691         return it->second.isBibTeX();
 692 }
 693
 694
 695
 696 vector<docstring> const BiblioInfo::getCiteStrings(
 697         docstring const & key, Buffer const & buf) const
 698 {
 699         CiteEngine const engine = buf.params().citeEngine();
 700         if (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL)
 701                 return getNumericalStrings(key, buf);
 702         else
 703                 return getAuthorYearStrings(key, buf);
 704 }
 705
 706
 707 vector<docstring> const BiblioInfo::getNumericalStrings(
 708         docstring const & key, Buffer const & buf) const
 709 {
 710         if (empty())
 711                 return vector<docstring>();
 712
 713         docstring const author = getAbbreviatedAuthor(key);
 714         docstring const year   = getYear(key);
 715         if (author.empty() || year.empty())
 716                 return vector<docstring>();
 717
 718         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 719
 720         vector<docstring> vec(styles.size());
 721         for (size_t i = 0; i != vec.size(); ++i) {
 722                 docstring str;
 723
 724                 switch (styles[i]) {
 725                         case CITE:
 726                         case CITEP:
 727                                 str = from_ascii("[#ID]");
 728                                 break;
 729
 730                         case NOCITE:
 731                                 str = _("Add to bibliography only.");
 732                                 break;
 733
 734                         case CITET:
 735                                 str = author + " [#ID]";
 736                                 break;
 737
 738                         case CITEALT:
 739                                 str = author + " #ID";
 740                                 break;
 741
 742                         case CITEALP:
 743                                 str = from_ascii("#ID");
 744                                 break;
 745
 746                         case CITEAUTHOR:
 747                                 str = author;
 748                                 break;
 749
 750                         case CITEYEAR:
 751                                 str = year;
 752                                 break;
 753
 754                         case CITEYEARPAR:
 755                                 str = '(' + year + ')';
 756                                 break;
 757                 }
 758
 759                 vec[i] = str;
 760         }
 761
 762         return vec;
 763 }
 764
 765
 766 vector<docstring> const BiblioInfo::getAuthorYearStrings(
 767         docstring const & key, Buffer const & buf) const
 768 {
 769         if (empty())
 770                 return vector<docstring>();
 771
 772         docstring const author = getAbbreviatedAuthor(key);
 773         docstring const year   = getYear(key);
 774         if (author.empty() || year.empty())
 775                 return vector<docstring>();
 776
 777         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 778
 779         vector<docstring> vec(styles.size());
 780         for (size_t i = 0; i != vec.size(); ++i) {
 781                 docstring str;
 782
 783                 switch (styles[i]) {
 784                         case CITE:
 785                 // jurabib only: Author/Annotator
 786                 // (i.e. the "before" field, 2nd opt arg)
 787                                 str = author + "/<" + _("before") + '>';
 788                                 break;
 789
 790                         case NOCITE:
 791                                 str = _("Add to bibliography only.");
 792                                 break;
 793
 794                         case CITET:
 795                                 str = author + " (" + year + ')';
 796                                 break;
 797
 798                         case CITEP:
 799                                 str = '(' + author + ", " + year + ')';
 800                                 break;
 801
 802                         case CITEALT:
 803                                 str = author + ' ' + year ;
 804                                 break;
 805
 806                         case CITEALP:
 807                                 str = author + ", " + year ;
 808                                 break;
 809
 810                         case CITEAUTHOR:
 811                                 str = author;
 812                                 break;
 813
 814                         case CITEYEAR:
 815                                 str = year;
 816                                 break;
 817
 818                         case CITEYEARPAR:
 819                                 str = '(' + year + ')';
 820                                 break;
 821                 }
 822                 vec[i] = str;
 823         }
 824         return vec;
 825 }
 826
 827
 828 void BiblioInfo::mergeBiblioInfo(BiblioInfo const & info)
 829 {
 830         bimap_.insert(info.begin(), info.end());
 831 }
 832
 833
 834 namespace {
 835         // used in xhtml to sort a list of BibTeXInfo objects
 836         bool lSorter(BibTeXInfo const * lhs, BibTeXInfo const * rhs)
 837         {
 838                 docstring const lauth = lhs->getAbbreviatedAuthor();
 839                 docstring const rauth = rhs->getAbbreviatedAuthor();
 840                 docstring const lyear = lhs->getYear();
 841                 docstring const ryear = rhs->getYear();
 842                 docstring const ltitl = lhs->operator[]("title");
 843                 docstring const rtitl = rhs->operator[]("title");
 844                 return  (lauth < rauth)
 845                                 || (lauth == rauth && lyear < ryear)
 846                                 || (lauth == rauth && lyear == ryear && ltitl < rtitl);
 847         }
 848 }
 849
 850
 851 void BiblioInfo::collectCitedEntries(Buffer const & buf)
 852 {
 853         cited_entries_.clear();
 854         // We are going to collect all the citation keys used in the document,
 855         // getting them from the TOC.
 856         // FIXME We may want to collect these differently, in the first case,
 857         // so that we might have them in order of appearance.
 858         set<docstring> citekeys;
 859         Toc const & toc = buf.tocBackend().toc("citation");
 860         Toc::const_iterator it = toc.begin();
 861         Toc::const_iterator const en = toc.end();
 862         for (; it != en; ++it) {
 863                 if (it->str().empty())
 864                         continue;
 865                 vector<docstring> const keys = getVectorFromString(it->str());
 866                 citekeys.insert(keys.begin(), keys.end());
 867         }
 868         if (citekeys.empty())
 869                 return;
 870
 871         // We have a set of the keys used in this document.
 872         // We will now convert it to a list of the BibTeXInfo objects used in
 873         // this document...
 874         vector<BibTeXInfo const *> bi;
 875         set<docstring>::const_iterator cit = citekeys.begin();
 876         set<docstring>::const_iterator const cen = citekeys.end();
 877         for (; cit != cen; ++cit) {
 878                 BiblioInfo::const_iterator const bt = find(*cit);
 879                 if (bt == end() || !bt->second.isBibTeX())
 880                         continue;
 881                 bi.push_back(&(bt->second));
 882         }
 883         // ...and sort it.
 884         sort(bi.begin(), bi.end(), lSorter);
 885
 886         // Now we can write the sorted keys
 887         vector<BibTeXInfo const *>::const_iterator bit = bi.begin();
 888         vector<BibTeXInfo const *>::const_iterator ben = bi.end();
 889         for (; bit != ben; ++bit)
 890                 cited_entries_.push_back((*bit)->key());
 891 }
 892
 893
 894 void BiblioInfo::makeCitationLabels(Buffer const & buf)
 895 {
 896         collectCitedEntries(buf);
 897         CiteEngine const engine = buf.params().citeEngine();
 898         bool const numbers =
 899                 (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL);
 900
 901         int keynumber = 0;
 902         char modifier = 0;
 903         // used to remember the last one we saw
 904         // we'll be comparing entries to see if we need to add
 905         // modifiers, like "1984a"
 906         map<docstring, BibTeXInfo>::iterator last;
 907
 908         vector<docstring>::const_iterator it = cited_entries_.begin();
 909         vector<docstring>::const_iterator const en = cited_entries_.end();
 910         for (; it != en; ++it) {
 911                 map<docstring, BibTeXInfo>::iterator const biit = bimap_.find(*it);
 912                 // this shouldn't happen, but...
 913                 if (biit == bimap_.end())
 914                         // ...fail gracefully, anyway.
 915                         continue;
 916                 BibTeXInfo & entry = biit->second;
 917                 if (numbers) {
 918                         docstring const num = convert<docstring>(++keynumber);
 919                         entry.setCiteNumber(num);
 920                 } else {
 921                         if (it != cited_entries_.begin()
 922                             && entry.getAbbreviatedAuthor() == last->second.getAbbreviatedAuthor()
 923                             // we access the year via getYear() so as to get it from the xref,
 924                             // if we need to do so
 925                             && getYear(entry.key()) == getYear(last->second.key())) {
 926                                 if (modifier == 0) {
 927                                         // so the last one should have been 'a'
 928                                         last->second.setModifier('a');
 929                                         modifier = 'b';
 930                                 } else if (modifier == 'z')
 931                                         modifier = 'A';
 932                                 else
 933                                         modifier++;
 934                         } else {
 935                                 modifier = 0;
 936                         }
 937                         entry.setModifier(modifier);
 938                         // remember the last one
 939                         last = biit;
 940                 }
 941         }
 942 }
 943
 944
 945 //////////////////////////////////////////////////////////////////////
 946 //
 947 // CitationStyle
 948 //
 949 //////////////////////////////////////////////////////////////////////
 950
 951 namespace {
 952
 953
 954 char const * const citeCommands[] = {
 955         "cite", "citet", "citep", "citealt", "citealp",
 956         "citeauthor", "citeyear", "citeyearpar", "nocite" };
 957
 958 unsigned int const nCiteCommands =
 959                 sizeof(citeCommands) / sizeof(char *);
 960
 961 CiteStyle const citeStylesArray[] = {
 962         CITE, CITET, CITEP, CITEALT, CITEALP,
 963         CITEAUTHOR, CITEYEAR, CITEYEARPAR, NOCITE };
 964
 965 unsigned int const nCiteStyles =
 966                 sizeof(citeStylesArray) / sizeof(CiteStyle);
 967
 968 CiteStyle const citeStylesFull[] = {
 969         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 970
 971 unsigned int const nCiteStylesFull =
 972                 sizeof(citeStylesFull) / sizeof(CiteStyle);
 973
 974 CiteStyle const citeStylesUCase[] = {
 975         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 976
 977 unsigned int const nCiteStylesUCase =
 978         sizeof(citeStylesUCase) / sizeof(CiteStyle);
 979
 980 } // namespace anon
 981
 982
 983 CitationStyle citationStyleFromString(string const & command)
 984 {
 985         CitationStyle s;
 986         if (command.empty())
 987                 return s;
 988
 989         string cmd = command;
 990         if (cmd[0] == 'C') {
 991                 s.forceUpperCase = true;
 992                 cmd[0] = 'c';
 993         }
 994
 995         size_t const n = cmd.size() - 1;
 996         if (cmd != "cite" && cmd[n] == '*') {
 997                 s.full = true;
 998                 cmd = cmd.substr(0, n);
 999         }
1000
1001         char const * const * const last = citeCommands + nCiteCommands;
1002         char const * const * const ptr = find(citeCommands, last, cmd);
1003
1004         if (ptr != last) {
1005                 size_t idx = ptr - citeCommands;
1006                 s.style = citeStylesArray[idx];
1007         }
1008         return s;
1009 }
1010
1011
1012 string citationStyleToString(const CitationStyle & s)
1013 {
1014         string cite = citeCommands[s.style];
1015         if (s.full) {
1016                 CiteStyle const * last = citeStylesFull + nCiteStylesFull;
1017                 if (std::find(citeStylesFull, last, s.style) != last)
1018                         cite += '*';
1019         }
1020
1021         if (s.forceUpperCase) {
1022                 CiteStyle const * last = citeStylesUCase + nCiteStylesUCase;
1023                 if (std::find(citeStylesUCase, last, s.style) != last)
1024                         cite[0] = 'C';
1025         }
1026
1027         return cite;
1028 }
1029
1030 vector<CiteStyle> citeStyles(CiteEngine engine)
1031 {
1032         unsigned int nStyles = 0;
1033         unsigned int start = 0;
1034
1035         switch (engine) {
1036                 case ENGINE_BASIC:
1037                         nStyles = 2;
1038                         start = 0;
1039                         break;
1040                 case ENGINE_NATBIB_AUTHORYEAR:
1041                 case ENGINE_NATBIB_NUMERICAL:
1042                         nStyles = nCiteStyles - 1;
1043                         start = 1;
1044                         break;
1045                 case ENGINE_JURABIB:
1046                         nStyles = nCiteStyles;
1047                         start = 0;
1048                         break;
1049         }
1050
1051         vector<CiteStyle> styles(nStyles);
1052         size_t i = 0;
1053         int j = start;
1054         for (; i != styles.size(); ++i, ++j)
1055                 styles[i] = citeStylesArray[j];
1056
1057         return styles;
1058 }
1059
1060 } // namespace lyx
1061