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 embedded option in `" << 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         docstring ret; // return value
 406         string key;
 407         bool scanning_key = false;
 408         bool scanning_rich = false;
 409
 410         string fmt = format;
 411         // we'll remove characters from the front of fmt as we
 412         // deal with them
 413         while (fmt.size()) {
 414                 char_type thischar = fmt[0];
 415                 if (thischar == '%') {
 416                         // beginning or end of key
 417                         if (scanning_key) {
 418                                 // end of key
 419                                 scanning_key = false;
 420                                 // so we replace the key with its value, which may be empty
 421                                 if (key == "pp_text")
 422                                         ret += _(pp_text);
 423                                 else if (key == "ed_text")
 424                                         ret += _(ed_text);
 425                                 else if(key == "edby_text")
 426                                         ret += _(edby_text);
 427                                 else {
 428                                         docstring const val = getValueForKey(key, xref);
 429                                         ret += val;
 430                                 }
 431                                 key.clear();
 432                         } else {
 433                                 // beginning of key
 434                                 scanning_key = true;
 435                         }
 436                 }
 437                 else if (thischar == '{') {
 438                         // beginning of option?
 439                         if (scanning_key) {
 440                                 LYXERR0("ERROR: Found `{' when scanning key in `" << format << "'.");
 441                                 return _("ERROR!");
 442                         }
 443                         if (fmt.size() > 1) {
 444                                 if (fmt[1] == '%') {
 445                                         // it is the beginning of an optional format
 446                                         string optkey;
 447                                         string ifpart;
 448                                         string elsepart;
 449                                         string const newfmt =
 450                                                 parseOptions(fmt, optkey, ifpart, elsepart);
 451                                         if (newfmt == fmt) // parse error
 452                                                 return _("ERROR!");
 453                                         fmt = newfmt;
 454                                         docstring const val = getValueForKey(optkey, xref);
 455                                         if (!val.empty())
 456                                                 ret += expandFormat(ifpart, xref, richtext);
 457                                         else if (!elsepart.empty())
 458                                                 ret += expandFormat(elsepart, xref, richtext);
 459                                         // fmt will have been shortened for us already
 460                                         continue;
 461                                 }
 462                                 if (fmt[1] == '!') {
 463                                         // beginning of rich text
 464                                         scanning_rich = true;
 465                                         fmt = fmt.substr(2);
 466                                         continue;
 467                                 }
 468                         }
 469                         // we are here if the '{' was at the end of the format. hmm.
 470                         ret += thischar;
 471                 }
 472                 else if (scanning_rich && thischar == '!'
 473                          && fmt.size() > 1 && fmt[1] == '}') {
 474                         // end of rich text
 475                         scanning_rich = false;
 476                         fmt = fmt.substr(2);
 477                         continue;
 478                 }
 479                 else if (scanning_key)
 480                         key += thischar;
 481                 else if (richtext || !scanning_rich)
 482                         ret += thischar;
 483                 // else the character is discarded, which will happen only if
 484                 // richtext == false and we are scanning rich text
 485                 fmt = fmt.substr(1);
 486         } // for loop
 487         if (scanning_key) {
 488                 LYXERR0("Never found end of key in `" << format << "'!");
 489                 return _("ERROR!");
 490         }
 491         if (scanning_rich) {
 492                 LYXERR0("Never found end of rich text in `" << format << "'!");
 493                 return _("ERROR!");
 494         }
 495         return ret;
 496 }
 497
 498
 499 namespace {
 500
 501 // FIXME These would be better read from a file, so that they
 502 // could be customized.
 503
 504         static string articleFormat = "%author%, \"%title%\", {!<i>!}%journal%{!</i>!} {%volume%[[ %volume%{%number%[[, %number%]]}]]} (%year%){%pages%[[, %pp_text% %pages%]]}.{%note%[[ %note%]]}";
 505
 506         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%]]}";
 507
 508         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%]]}";
 509
 510         static string thesis = "%author%, %title% ({%address%[[%address%: ]]}%school%, %year%).{%note%[[ %note%]]}";
 511
 512         static string defaultFormat = "{%author%[[%author%, ]][[{%editor%[[%editor%, %ed_text%, ]]}]]}\"%title%\"{%journal%[[, {!<i>!}%journal%{!</i>!}]][[{%publisher%[[, %publisher%]][[{%institution%[[, %institution%]]}]]}]]}{%year%[[ (%year%)]]}{%pages%[[, %pages%]]}.";
 513
 514 }
 515
 516 docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref,
 517         bool richtext) const
 518 {
 519         if (!info_.empty())
 520                 return info_;
 521
 522         if (!is_bibtex_) {
 523                 BibTeXInfo::const_iterator it = find(from_ascii("ref"));
 524                 info_ = it->second;
 525                 return info_;
 526         }
 527
 528         if (entry_type_ == "article")
 529                 info_ = expandFormat(articleFormat, xref, richtext);
 530         else if (entry_type_ == "book")
 531                 info_ = expandFormat(bookFormat, xref, richtext);
 532         else if (entry_type_.substr(0,2) == "in")
 533                 info_ = expandFormat(inSomething, xref, richtext);
 534         else if (entry_type_ == "phdthesis" || entry_type_ == "mastersthesis")
 535                 info_ = expandFormat(thesis, xref, richtext);
 536         else
 537                 info_ = expandFormat(defaultFormat, xref, richtext);
 538
 539         if (!info_.empty())
 540                 info_ = convertLaTeXCommands(info_);
 541         return info_;
 542 }
 543
 544
 545 docstring const & BibTeXInfo::operator[](docstring const & field) const
 546 {
 547         BibTeXInfo::const_iterator it = find(field);
 548         if (it != end())
 549                 return it->second;
 550         static docstring const empty_value = docstring();
 551         return empty_value;
 552 }
 553
 554
 555 docstring const & BibTeXInfo::operator[](string const & field) const
 556 {
 557         return operator[](from_ascii(field));
 558 }
 559
 560
 561 docstring BibTeXInfo::getValueForKey(string const & key,
 562                 BibTeXInfo const * const xref) const
 563 {
 564         docstring const ret = operator[](key);
 565         if (!ret.empty() || !xref)
 566                 return ret;
 567         return (*xref)[key];
 568 }
 569
 570
 571 //////////////////////////////////////////////////////////////////////
 572 //
 573 // BiblioInfo
 574 //
 575 //////////////////////////////////////////////////////////////////////
 576
 577 namespace {
 578 // A functor for use with sort, leading to case insensitive sorting
 579         class compareNoCase: public binary_function<docstring, docstring, bool>
 580         {
 581                 public:
 582                         bool operator()(docstring const & s1, docstring const & s2) const {
 583                                 return compare_no_case(s1, s2) < 0;
 584                         }
 585         };
 586 } // namespace anon
 587
 588
 589 vector<docstring> const BiblioInfo::getKeys() const
 590 {
 591         vector<docstring> bibkeys;
 592         BiblioInfo::const_iterator it  = begin();
 593         for (; it != end(); ++it)
 594                 bibkeys.push_back(it->first);
 595         sort(bibkeys.begin(), bibkeys.end(), compareNoCase());
 596         return bibkeys;
 597 }
 598
 599
 600 vector<docstring> const BiblioInfo::getFields() const
 601 {
 602         vector<docstring> bibfields;
 603         set<docstring>::const_iterator it = field_names_.begin();
 604         set<docstring>::const_iterator end = field_names_.end();
 605         for (; it != end; ++it)
 606                 bibfields.push_back(*it);
 607         sort(bibfields.begin(), bibfields.end());
 608         return bibfields;
 609 }
 610
 611
 612 vector<docstring> const BiblioInfo::getEntries() const
 613 {
 614         vector<docstring> bibentries;
 615         set<docstring>::const_iterator it = entry_types_.begin();
 616         set<docstring>::const_iterator end = entry_types_.end();
 617         for (; it != end; ++it)
 618                 bibentries.push_back(*it);
 619         sort(bibentries.begin(), bibentries.end());
 620         return bibentries;
 621 }
 622
 623
 624 docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key) const
 625 {
 626         BiblioInfo::const_iterator it = find(key);
 627         if (it == end())
 628                 return docstring();
 629         BibTeXInfo const & data = it->second;
 630         return data.getAbbreviatedAuthor();
 631 }
 632
 633
 634 docstring const BiblioInfo::getCiteNumber(docstring const & key) const
 635 {
 636         BiblioInfo::const_iterator it = find(key);
 637         if (it == end())
 638                 return docstring();
 639         BibTeXInfo const & data = it->second;
 640         return data.citeNumber();
 641 }
 642
 643
 644 docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier) const
 645 {
 646         BiblioInfo::const_iterator it = find(key);
 647         if (it == end())
 648                 return docstring();
 649         BibTeXInfo const & data = it->second;
 650         docstring year = data.getYear();
 651         if (year.empty()) {
 652                 // let's try the crossref
 653                 docstring const xref = data.getXRef();
 654                 if (xref.empty())
 655                         return _("No year"); // no luck
 656                 BiblioInfo::const_iterator const xrefit = find(xref);
 657                 if (xrefit == end())
 658                         return _("No year"); // no luck again
 659                 BibTeXInfo const & xref_data = xrefit->second;
 660                 year = xref_data.getYear();
 661         }
 662         if (use_modifier && data.modifier() != 0)
 663                 year += data.modifier();
 664         return year;
 665 }
 666
 667
 668 docstring const BiblioInfo::getInfo(docstring const & key, bool richtext) const
 669 {
 670         BiblioInfo::const_iterator it = find(key);
 671         if (it == end())
 672                 return docstring();
 673         BibTeXInfo const & data = it->second;
 674         BibTeXInfo const * xrefptr = 0;
 675         docstring const xref = data.getXRef();
 676         if (!xref.empty()) {
 677                 BiblioInfo::const_iterator const xrefit = find(xref);
 678                 if (xrefit != end())
 679                         xrefptr = &(xrefit->second);
 680         }
 681         return data.getInfo(xrefptr, richtext);
 682 }
 683
 684
 685 bool BiblioInfo::isBibtex(docstring const & key) const
 686 {
 687         BiblioInfo::const_iterator it = find(key);
 688         if (it == end())
 689                 return false;
 690         return it->second.isBibTeX();
 691 }
 692
 693
 694
 695 vector<docstring> const BiblioInfo::getCiteStrings(
 696         docstring const & key, Buffer const & buf) const
 697 {
 698         CiteEngine const engine = buf.params().citeEngine();
 699         if (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL)
 700                 return getNumericalStrings(key, buf);
 701         else
 702                 return getAuthorYearStrings(key, buf);
 703 }
 704
 705
 706 vector<docstring> const BiblioInfo::getNumericalStrings(
 707         docstring const & key, Buffer const & buf) const
 708 {
 709         if (empty())
 710                 return vector<docstring>();
 711
 712         docstring const author = getAbbreviatedAuthor(key);
 713         docstring const year   = getYear(key);
 714         if (author.empty() || year.empty())
 715                 return vector<docstring>();
 716
 717         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 718
 719         vector<docstring> vec(styles.size());
 720         for (size_t i = 0; i != vec.size(); ++i) {
 721                 docstring str;
 722
 723                 switch (styles[i]) {
 724                         case CITE:
 725                         case CITEP:
 726                                 str = from_ascii("[#ID]");
 727                                 break;
 728
 729                         case NOCITE:
 730                                 str = _("Add to bibliography only.");
 731                                 break;
 732
 733                         case CITET:
 734                                 str = author + " [#ID]";
 735                                 break;
 736
 737                         case CITEALT:
 738                                 str = author + " #ID";
 739                                 break;
 740
 741                         case CITEALP:
 742                                 str = from_ascii("#ID");
 743                                 break;
 744
 745                         case CITEAUTHOR:
 746                                 str = author;
 747                                 break;
 748
 749                         case CITEYEAR:
 750                                 str = year;
 751                                 break;
 752
 753                         case CITEYEARPAR:
 754                                 str = '(' + year + ')';
 755                                 break;
 756                 }
 757
 758                 vec[i] = str;
 759         }
 760
 761         return vec;
 762 }
 763
 764
 765 vector<docstring> const BiblioInfo::getAuthorYearStrings(
 766         docstring const & key, Buffer const & buf) const
 767 {
 768         if (empty())
 769                 return vector<docstring>();
 770
 771         docstring const author = getAbbreviatedAuthor(key);
 772         docstring const year   = getYear(key);
 773         if (author.empty() || year.empty())
 774                 return vector<docstring>();
 775
 776         vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
 777
 778         vector<docstring> vec(styles.size());
 779         for (size_t i = 0; i != vec.size(); ++i) {
 780                 docstring str;
 781
 782                 switch (styles[i]) {
 783                         case CITE:
 784                 // jurabib only: Author/Annotator
 785                 // (i.e. the "before" field, 2nd opt arg)
 786                                 str = author + "/<" + _("before") + '>';
 787                                 break;
 788
 789                         case NOCITE:
 790                                 str = _("Add to bibliography only.");
 791                                 break;
 792
 793                         case CITET:
 794                                 str = author + " (" + year + ')';
 795                                 break;
 796
 797                         case CITEP:
 798                                 str = '(' + author + ", " + year + ')';
 799                                 break;
 800
 801                         case CITEALT:
 802                                 str = author + ' ' + year ;
 803                                 break;
 804
 805                         case CITEALP:
 806                                 str = author + ", " + year ;
 807                                 break;
 808
 809                         case CITEAUTHOR:
 810                                 str = author;
 811                                 break;
 812
 813                         case CITEYEAR:
 814                                 str = year;
 815                                 break;
 816
 817                         case CITEYEARPAR:
 818                                 str = '(' + year + ')';
 819                                 break;
 820                 }
 821                 vec[i] = str;
 822         }
 823         return vec;
 824 }
 825
 826
 827 void BiblioInfo::mergeBiblioInfo(BiblioInfo const & info)
 828 {
 829         bimap_.insert(info.begin(), info.end());
 830 }
 831
 832
 833 namespace {
 834         // used in xhtml to sort a list of BibTeXInfo objects
 835         bool lSorter(BibTeXInfo const * lhs, BibTeXInfo const * rhs)
 836         {
 837                 docstring const lauth = lhs->getAbbreviatedAuthor();
 838                 docstring const rauth = rhs->getAbbreviatedAuthor();
 839                 docstring const lyear = lhs->getYear();
 840                 docstring const ryear = rhs->getYear();
 841                 docstring const ltitl = lhs->operator[]("title");
 842                 docstring const rtitl = rhs->operator[]("title");
 843                 return  (lauth < rauth)
 844                                 || (lauth == rauth && lyear < ryear)
 845                                 || (lauth == rauth && lyear == ryear && ltitl < rtitl);
 846         }
 847 }
 848
 849
 850 void BiblioInfo::collectCitedEntries(Buffer const & buf)
 851 {
 852         cited_entries_.clear();
 853         // We are going to collect all the citation keys used in the document,
 854         // getting them from the TOC.
 855         // FIXME We may want to collect these differently, in the first case,
 856         // so that we might have them in order of appearance.
 857         set<docstring> citekeys;
 858         Toc const & toc = buf.tocBackend().toc("citation");
 859         Toc::const_iterator it = toc.begin();
 860         Toc::const_iterator const en = toc.end();
 861         for (; it != en; ++it) {
 862                 if (it->str().empty())
 863                         continue;
 864                 vector<docstring> const keys = getVectorFromString(it->str());
 865                 citekeys.insert(keys.begin(), keys.end());
 866         }
 867         if (citekeys.empty())
 868                 return;
 869
 870         // We have a set of the keys used in this document.
 871         // We will now convert it to a list of the BibTeXInfo objects used in
 872         // this document...
 873         vector<BibTeXInfo const *> bi;
 874         set<docstring>::const_iterator cit = citekeys.begin();
 875         set<docstring>::const_iterator const cen = citekeys.end();
 876         for (; cit != cen; ++cit) {
 877                 BiblioInfo::const_iterator const bt = find(*cit);
 878                 if (bt == end() || !bt->second.isBibTeX())
 879                         continue;
 880                 bi.push_back(&(bt->second));
 881         }
 882         // ...and sort it.
 883         sort(bi.begin(), bi.end(), lSorter);
 884
 885         // Now we can write the sorted keys
 886         vector<BibTeXInfo const *>::const_iterator bit = bi.begin();
 887         vector<BibTeXInfo const *>::const_iterator ben = bi.end();
 888         for (; bit != ben; ++bit)
 889                 cited_entries_.push_back((*bit)->key());
 890 }
 891
 892
 893 void BiblioInfo::makeCitationLabels(Buffer const & buf)
 894 {
 895         collectCitedEntries(buf);
 896         CiteEngine const engine = buf.params().citeEngine();
 897         bool const numbers =
 898                 (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL);
 899
 900         int keynumber = 0;
 901         char modifier = 0;
 902         // used to remember the last one we saw
 903         // we'll be comparing entries to see if we need to add
 904         // modifiers, like "1984a"
 905         map<docstring, BibTeXInfo>::iterator last;
 906
 907         vector<docstring>::const_iterator it = cited_entries_.begin();
 908         vector<docstring>::const_iterator const en = cited_entries_.end();
 909         for (; it != en; ++it) {
 910                 map<docstring, BibTeXInfo>::iterator const biit = bimap_.find(*it);
 911                 // this shouldn't happen, but...
 912                 if (biit == bimap_.end())
 913                         // ...fail gracefully, anyway.
 914                         continue;
 915                 BibTeXInfo & entry = biit->second;
 916                 if (numbers) {
 917                         docstring const num = convert<docstring>(++keynumber);
 918                         entry.setCiteNumber(num);
 919                 } else {
 920                         if (it != cited_entries_.begin()
 921                             && entry.getAbbreviatedAuthor() == last->second.getAbbreviatedAuthor()
 922                             // we access the year via getYear() so as to get it from the xref,
 923                             // if we need to do so
 924                             && getYear(entry.key()) == getYear(last->second.key())) {
 925                                 if (modifier == 0) {
 926                                         // so the last one should have been 'a'
 927                                         last->second.setModifier('a');
 928                                         modifier = 'b';
 929                                 } else if (modifier == 'z')
 930                                         modifier = 'A';
 931                                 else
 932                                         modifier++;
 933                         } else {
 934                                 modifier = 0;
 935                         }
 936                         entry.setModifier(modifier);
 937                         // remember the last one
 938                         last = biit;
 939                 }
 940         }
 941 }
 942
 943
 944 //////////////////////////////////////////////////////////////////////
 945 //
 946 // CitationStyle
 947 //
 948 //////////////////////////////////////////////////////////////////////
 949
 950 namespace {
 951
 952
 953 char const * const citeCommands[] = {
 954         "cite", "citet", "citep", "citealt", "citealp",
 955         "citeauthor", "citeyear", "citeyearpar", "nocite" };
 956
 957 unsigned int const nCiteCommands =
 958                 sizeof(citeCommands) / sizeof(char *);
 959
 960 CiteStyle const citeStylesArray[] = {
 961         CITE, CITET, CITEP, CITEALT, CITEALP,
 962         CITEAUTHOR, CITEYEAR, CITEYEARPAR, NOCITE };
 963
 964 unsigned int const nCiteStyles =
 965                 sizeof(citeStylesArray) / sizeof(CiteStyle);
 966
 967 CiteStyle const citeStylesFull[] = {
 968         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 969
 970 unsigned int const nCiteStylesFull =
 971                 sizeof(citeStylesFull) / sizeof(CiteStyle);
 972
 973 CiteStyle const citeStylesUCase[] = {
 974         CITET, CITEP, CITEALT, CITEALP, CITEAUTHOR };
 975
 976 unsigned int const nCiteStylesUCase =
 977         sizeof(citeStylesUCase) / sizeof(CiteStyle);
 978
 979 } // namespace anon
 980
 981
 982 CitationStyle citationStyleFromString(string const & command)
 983 {
 984         CitationStyle s;
 985         if (command.empty())
 986                 return s;
 987
 988         string cmd = command;
 989         if (cmd[0] == 'C') {
 990                 s.forceUpperCase = true;
 991                 cmd[0] = 'c';
 992         }
 993
 994         size_t const n = cmd.size() - 1;
 995         if (cmd != "cite" && cmd[n] == '*') {
 996                 s.full = true;
 997                 cmd = cmd.substr(0, n);
 998         }
 999
1000         char const * const * const last = citeCommands + nCiteCommands;
1001         char const * const * const ptr = find(citeCommands, last, cmd);
1002
1003         if (ptr != last) {
1004                 size_t idx = ptr - citeCommands;
1005                 s.style = citeStylesArray[idx];
1006         }
1007         return s;
1008 }
1009
1010
1011 string citationStyleToString(const CitationStyle & s)
1012 {
1013         string cite = citeCommands[s.style];
1014         if (s.full) {
1015                 CiteStyle const * last = citeStylesFull + nCiteStylesFull;
1016                 if (std::find(citeStylesFull, last, s.style) != last)
1017                         cite += '*';
1018         }
1019
1020         if (s.forceUpperCase) {
1021                 CiteStyle const * last = citeStylesUCase + nCiteStylesUCase;
1022                 if (std::find(citeStylesUCase, last, s.style) != last)
1023                         cite[0] = 'C';
1024         }
1025
1026         return cite;
1027 }
1028
1029 vector<CiteStyle> citeStyles(CiteEngine engine)
1030 {
1031         unsigned int nStyles = 0;
1032         unsigned int start = 0;
1033
1034         switch (engine) {
1035                 case ENGINE_BASIC:
1036                         nStyles = 2;
1037                         start = 0;
1038                         break;
1039                 case ENGINE_NATBIB_AUTHORYEAR:
1040                 case ENGINE_NATBIB_NUMERICAL:
1041                         nStyles = nCiteStyles - 1;
1042                         start = 1;
1043                         break;
1044                 case ENGINE_JURABIB:
1045                         nStyles = nCiteStyles;
1046                         start = 0;
1047                         break;
1048         }
1049
1050         vector<CiteStyle> styles(nStyles);
1051         size_t i = 0;
1052         int j = start;
1053         for (; i != styles.size(); ++i, ++j)
1054                 styles[i] = citeStylesArray[j];
1055
1056         return styles;
1057 }
1058
1059 } // namespace lyx
1060