2 * \file InsetBibtex.cpp
3 * This file is part of LyX, the document processor.
4 * Licence details can be found in the file COPYING.
6 * \author Alejandro Aguilar Sierra
7 * \author Richard Heck (BibTeX parser improvements)
9 * Full author contact details are available in file CREDITS.
14 #include "InsetBibtex.h"
16 #include "BiblioInfo.h"
18 #include "BufferParams.h"
20 #include "DispatchResult.h"
24 #include "FuncRequest.h"
25 #include "FuncStatus.h"
27 #include "LaTeXFeatures.h"
28 #include "output_xhtml.h"
29 #include "OutputParams.h"
30 #include "PDFOptions.h"
31 #include "TextClass.h"
33 #include "frontends/alert.h"
35 #include "support/convert.h"
36 #include "support/debug.h"
37 #include "support/docstream.h"
38 #include "support/ExceptionMessage.h"
39 #include "support/FileNameList.h"
40 #include "support/filetools.h"
41 #include "support/gettext.h"
42 #include "support/lstrings.h"
43 #include "support/os.h"
44 #include "support/Path.h"
45 #include "support/textutils.h"
50 using namespace lyx::support;
54 namespace Alert = frontend::Alert;
55 namespace os = support::os;
58 InsetBibtex::InsetBibtex(Buffer * buf, InsetCommandParams const & p)
59 : InsetCommand(buf, p)
61 buffer().invalidateBibinfoCache();
62 buffer().removeBiblioTempFiles();
66 InsetBibtex::~InsetBibtex()
68 if (isBufferLoaded()) {
69 buffer().invalidateBibfileCache();
70 buffer().removeBiblioTempFiles();
75 ParamInfo const & InsetBibtex::findInfo(string const & /* cmdName */)
77 static ParamInfo param_info_;
78 if (param_info_.empty()) {
79 param_info_.add("btprint", ParamInfo::LATEX_OPTIONAL);
80 param_info_.add("bibfiles", ParamInfo::LATEX_REQUIRED);
81 param_info_.add("options", ParamInfo::LYX_INTERNAL);
87 void InsetBibtex::doDispatch(Cursor & cur, FuncRequest & cmd)
89 switch (cmd.action()) {
95 case LFUN_INSET_MODIFY: {
96 InsetCommandParams p(BIBTEX_CODE);
98 if (!InsetCommand::string2params(to_utf8(cmd.argument()), p)) {
102 } catch (ExceptionMessage const & message) {
103 if (message.type_ == WarningException) {
104 Alert::warning(message.title_, message.details_);
105 cur.noScreenUpdate();
113 buffer().invalidateBibfileCache();
114 buffer().removeBiblioTempFiles();
115 cur.forceBufferUpdate();
120 InsetCommand::doDispatch(cur, cmd);
126 bool InsetBibtex::getStatus(Cursor & cur, FuncRequest const & cmd,
127 FuncStatus & flag) const
129 switch (cmd.action()) {
130 case LFUN_INSET_EDIT:
131 flag.setEnabled(true);
135 return InsetCommand::getStatus(cur, cmd, flag);
140 void InsetBibtex::editDatabases() const
142 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
144 if (bibfilelist.empty())
147 int nr_databases = bibfilelist.size();
148 if (nr_databases > 1) {
149 docstring message = bformat(_("The BibTeX inset includes %1$s databases.\n"
150 "If you proceed, all of them will be opened."),
151 convert<docstring>(nr_databases));
152 int const ret = Alert::prompt(_("Open Databases?"),
153 message, 0, 1, _("&Cancel"), _("&Proceed"));
159 vector<docstring>::const_iterator it = bibfilelist.begin();
160 vector<docstring>::const_iterator en = bibfilelist.end();
161 for (; it != en; ++it) {
162 FileName const bibfile = getBibTeXPath(*it, buffer());
163 formats.edit(buffer(), bibfile,
164 formats.getFormatFromFile(bibfile));
169 docstring InsetBibtex::screenLabel() const
171 return _("BibTeX Generated Bibliography");
175 docstring InsetBibtex::toolTip(BufferView const & /*bv*/, int /*x*/, int /*y*/) const
177 docstring item = from_ascii("* ");
178 docstring tip = _("Databases:") + "\n";
179 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
181 if (bibfilelist.empty()) {
185 vector<docstring>::const_iterator it = bibfilelist.begin();
186 vector<docstring>::const_iterator en = bibfilelist.end();
187 for (; it != en; ++it) {
195 docstring style = getParam("options"); // maybe empty! and with bibtotoc
196 docstring bibtotoc = from_ascii("bibtotoc");
197 if (prefixIs(style, bibtotoc)) {
199 if (contains(style, char_type(',')))
200 style = split(style, bibtotoc, char_type(','));
203 tip += _("Style File:") +"\n";
210 tip += "\n" + _("Lists:") + " ";
211 docstring btprint = getParam("btprint");
212 if (btprint == "btPrintAll")
213 tip += _("all references");
214 else if (btprint == "btPrintNotCited")
215 tip += _("all uncited references");
217 tip += _("all cited references");
221 tip += _("included in TOC");
228 static string normalizeName(Buffer const & buffer,
229 OutputParams const & runparams, string const & name, string const & ext)
231 string const fname = makeAbsPath(name, buffer.filePath()).absFileName();
232 if (FileName::isAbsolute(name) || !FileName(fname + ext).isReadableFile())
238 return to_utf8(makeRelPath(from_utf8(fname),
239 from_utf8(buffer.masterBuffer()->filePath())));
243 void InsetBibtex::latex(otexstream & os, OutputParams const & runparams) const
245 // the sequence of the commands:
246 // 1. \bibliographystyle{style}
247 // 2. \addcontentsline{...} - if option bibtotoc set
248 // 3. \bibliography{database}
249 // and with bibtopic:
250 // 1. \bibliographystyle{style}
251 // 2. \begin{btSect}{database}
252 // 3. \btPrint{Cited|NotCited|All}
256 // If we are processing the LaTeX file in a temp directory then
257 // copy the .bib databases to this temp directory, mangling their
258 // names in the process. Store this mangled name in the list of
260 // (We need to do all this because BibTeX *really*, *really*
261 // can't handle "files with spaces" and Windows users tend to
262 // use such filenames.)
263 // Otherwise, store the (maybe absolute) path to the original,
264 // unmangled database name.
265 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
266 vector<docstring>::const_iterator it = bibfilelist.begin();
267 vector<docstring>::const_iterator en = bibfilelist.end();
268 odocstringstream dbs;
271 // determine the export format
272 string const tex_format = flavor2format(runparams.flavor);
274 for (; it != en; ++it) {
275 string utf8input = to_utf8(*it);
277 normalizeName(buffer(), runparams, utf8input, ".bib");
278 FileName const try_in_file =
279 makeAbsPath(database + ".bib", buffer().filePath());
280 bool const not_from_texmf = try_in_file.isReadableFile();
282 if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
284 // mangledFileName() needs the extension
285 DocFileName const in_file = DocFileName(try_in_file);
286 database = removeExtension(in_file.mangledFileName());
287 FileName const out_file = makeAbsPath(database + ".bib",
288 buffer().masterBuffer()->temppath());
289 bool const success = in_file.copyTo(out_file);
291 lyxerr << "Failed to copy '" << in_file
292 << "' to '" << out_file << "'"
295 } else if (!runparams.inComment && runparams.nice && not_from_texmf) {
296 runparams.exportdata->addExternalFile(tex_format, try_in_file, database + ".bib");
297 if (!isValidLaTeXFileName(database)) {
298 frontend::Alert::warning(_("Invalid filename"),
299 _("The following filename will cause troubles "
300 "when running the exported file through LaTeX: ") +
301 from_utf8(database));
303 if (!isValidDVIFileName(database)) {
304 frontend::Alert::warning(_("Problematic filename for DVI"),
305 _("The following filename can cause troubles "
306 "when running the exported file through LaTeX "
307 "and opening the resulting DVI: ") +
308 from_utf8(database), true);
317 dbs << from_utf8(latex_path(database));
319 docstring const db_out = dbs.str();
321 // Post this warning only once.
322 static bool warned_about_spaces = false;
323 if (!warned_about_spaces &&
324 runparams.nice && db_out.find(' ') != docstring::npos) {
325 warned_about_spaces = true;
326 Alert::warning(_("Export Warning!"),
327 _("There are spaces in the paths to your BibTeX databases.\n"
328 "BibTeX will be unable to find them."));
331 string style = to_utf8(getParam("options")); // maybe empty! and with bibtotoc
333 if (prefixIs(style, "bibtotoc")) {
334 bibtotoc = "bibtotoc";
335 if (contains(style, ','))
336 style = split(style, bibtotoc, ',');
339 if (style == "default")
340 style = buffer().params().biblio_style;
342 if (!style.empty()) {
343 string base = normalizeName(buffer(), runparams, style, ".bst");
344 FileName const try_in_file =
345 makeAbsPath(base + ".bst", buffer().filePath());
346 bool const not_from_texmf = try_in_file.isReadableFile();
347 // If this style does not come from texmf and we are not
348 // exporting to .tex copy it to the tmp directory.
349 // This prevents problems with spaces and 8bit charcaters
351 if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
353 // use new style name
354 DocFileName const in_file = DocFileName(try_in_file);
355 base = removeExtension(in_file.mangledFileName());
356 FileName const out_file = makeAbsPath(base + ".bst",
357 buffer().masterBuffer()->temppath());
358 bool const success = in_file.copyTo(out_file);
360 lyxerr << "Failed to copy '" << in_file
361 << "' to '" << out_file << "'"
366 os << "\\bibliographystyle{"
367 << from_utf8(latex_path(normalizeName(buffer(), runparams, base, ".bst")))
371 // Post this warning only once.
372 static bool warned_about_bst_spaces = false;
373 if (!warned_about_bst_spaces && runparams.nice && contains(style, ' ')) {
374 warned_about_bst_spaces = true;
375 Alert::warning(_("Export Warning!"),
376 _("There are spaces in the path to your BibTeX style file.\n"
377 "BibTeX will be unable to find it."));
380 if (!db_out.empty() && buffer().params().use_bibtopic) {
381 os << "\\begin{btSect}{" << db_out << "}\n";
382 docstring btprint = getParam("btprint");
385 btprint = from_ascii("btPrintCited");
386 os << "\\" << btprint << "\n"
387 << "\\end{btSect}\n";
391 if (!bibtotoc.empty() && !buffer().params().use_bibtopic) {
392 // set label for hyperref, see http://www.lyx.org/trac/ticket/6470
393 if (buffer().params().pdfoptions().use_hyperref)
394 os << "\\phantomsection";
395 if (buffer().params().documentClass().hasLaTeXLayout("chapter"))
396 os << "\\addcontentsline{toc}{chapter}{\\bibname}";
397 else if (buffer().params().documentClass().hasLaTeXLayout("section"))
398 os << "\\addcontentsline{toc}{section}{\\refname}";
401 if (!db_out.empty() && !buffer().params().use_bibtopic) {
402 docstring btprint = getParam("btprint");
403 if (btprint == "btPrintAll") {
404 os << "\\nocite{*}\n";
406 os << "\\bibliography{" << db_out << "}\n";
411 support::FileNameList InsetBibtex::getBibFiles() const
413 FileName path(buffer().filePath());
414 support::PathChanger p(path);
416 support::FileNameList vec;
418 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
419 vector<docstring>::const_iterator it = bibfilelist.begin();
420 vector<docstring>::const_iterator en = bibfilelist.end();
421 for (; it != en; ++it) {
422 FileName const file = getBibTeXPath(*it, buffer());
427 LYXERR0("Couldn't find " + to_utf8(*it) + " in InsetBibtex::getBibFiles()!");
436 // methods for parsing bibtex files
438 typedef map<docstring, docstring> VarMap;
440 /// remove whitespace characters, optionally a single comma,
441 /// and further whitespace characters from the stream.
442 /// @return true if a comma was found, false otherwise
444 bool removeWSAndComma(ifdocstream & ifs) {
453 } while (ifs && isSpace(ch));
466 } while (ifs && isSpace(ch));
481 /// remove whitespace characters, read characer sequence
482 /// not containing whitespace characters or characters in
483 /// delimChars, and remove further whitespace characters.
485 /// @return true if a string of length > 0 could be read.
487 bool readTypeOrKey(docstring & val, ifdocstream & ifs,
488 docstring const & delimChars, docstring const & illegalChars,
501 } while (ifs && isSpace(ch));
507 bool legalChar = true;
508 while (ifs && !isSpace(ch) &&
509 delimChars.find(ch) == docstring::npos &&
510 (legalChar = (illegalChars.find(ch) == docstring::npos))
513 if (chCase == makeLowerCase)
514 val += lowercase(ch);
526 while (ifs && isSpace(ch)) {
534 return val.length() > 0;
537 /// read subsequent bibtex values that are delimited with a #-character.
538 /// Concatenate all parts and replace names with the associated string in
539 /// the variable strings.
540 /// @return true if reading was successfull (all single parts were delimited
542 bool readValue(docstring & val, ifdocstream & ifs, const VarMap & strings) {
555 } while (ifs && isSpace(ch));
560 // check for field type
561 if (isDigitASCII(ch)) {
563 // read integer value
567 } while (ifs && isDigitASCII(ch));
572 } else if (ch == '"' || ch == '{') {
574 char_type delim = ch == '"' ? '"': '}';
579 } while (ifs && isSpace(ch));
584 // We now have the first non-whitespace character
585 // We'll collapse adjacent whitespace.
586 bool lastWasWhiteSpace = false;
588 // inside this delimited text braces must match.
589 // Thus we can have a closing delimiter only
590 // when nestLevel == 0
593 while (ifs && (nestLevel > 0 || ch != delim)) {
595 lastWasWhiteSpace = true;
599 // We output the space only after we stop getting
600 // whitespace so as not to output any whitespace
601 // at the end of the value.
602 if (lastWasWhiteSpace) {
603 lastWasWhiteSpace = false;
609 // update nesting level
628 // FIXME Why is this here?
636 // reading a string name
639 while (ifs && !isSpace(ch) && ch != '#' && ch != ',' && ch != '}' && ch != ')') {
640 strName += lowercase(ch);
647 // replace the string with its assigned value or
648 // discard it if it's not assigned
649 if (strName.length()) {
650 VarMap::const_iterator pos = strings.find(strName);
651 if (pos != strings.end()) {
658 while (ifs && isSpace(ch)) {
665 // continue reading next value on concatenate with '#'
675 void InsetBibtex::collectBibKeys(InsetIterator const & /*di*/) const
681 void InsetBibtex::parseBibTeXFiles() const
683 // This bibtex parser is a first step to parse bibtex files
686 // - it reads the whole bibtex entry and does a syntax check
687 // (matching delimiters, missing commas,...
688 // - it recovers from errors starting with the next @-character
689 // - it reads @string definitions and replaces them in the
691 // - it accepts more characters in keys or value names than
694 // Officially bibtex does only support ASCII, but in practice
695 // you can use the encoding of the main document as long as
696 // some elements like keys and names are pure ASCII. Therefore
697 // we convert the file from the buffer encoding.
698 // We don't restrict keys to ASCII in LyX, since our own
699 // InsetBibitem can generate non-ASCII keys, and nonstandard
700 // 8bit clean bibtex forks exist.
704 support::FileNameList const files = getBibFiles();
705 support::FileNameList::const_iterator it = files.begin();
706 support::FileNameList::const_iterator en = files.end();
707 for (; it != en; ++ it) {
708 ifdocstream ifs(it->toFilesystemEncoding().c_str(),
709 ios_base::in, buffer().params().encoding().iconvName());
725 if (!readTypeOrKey(entryType, ifs, from_ascii("{("), docstring(), makeLowerCase)) {
726 lyxerr << "BibTeX Parser: Error reading entry type." << std::endl;
731 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
735 if (entryType == from_ascii("comment")) {
736 ifs.ignore(numeric_limits<int>::max(), '\n');
742 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
746 if ((ch != '(') && (ch != '{')) {
747 lyxerr << "BibTeX Parser: Invalid entry delimiter." << std::endl;
753 if (entryType == from_ascii("string")) {
755 // read string and add it to the strings map
756 // (or replace it's old value)
760 if (!readTypeOrKey(name, ifs, from_ascii("="), from_ascii("#{}(),"), makeLowerCase)) {
761 lyxerr << "BibTeX Parser: Error reading string name." << std::endl;
766 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
770 // next char must be an equal sign
772 if (!ifs || ch != '=') {
773 lyxerr << "BibTeX Parser: No `=' after string name: " <<
774 name << "." << std::endl;
778 if (!readValue(value, ifs, strings)) {
779 lyxerr << "BibTeX Parser: Unable to read value for string: " <<
780 name << "." << std::endl;
784 strings[name] = value;
786 } else if (entryType == from_ascii("preamble")) {
788 // preamble definitions are discarded.
789 // can they be of any use in lyx?
792 if (!readValue(value, ifs, strings)) {
793 lyxerr << "BibTeX Parser: Unable to read preamble value." << std::endl;
799 // Citation entry. Try to read the key.
802 if (!readTypeOrKey(key, ifs, from_ascii(","), from_ascii("}"), keepCase)) {
803 lyxerr << "BibTeX Parser: Unable to read key for entry type:" <<
804 entryType << "." << std::endl;
809 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
813 /////////////////////////////////////////////
814 // now we have a key, so we will add an entry
815 // (even if it's empty, as bibtex does)
817 // we now read the field = value pairs.
818 // all items must be separated by a comma. If
819 // it is missing the scanning of this entry is
820 // stopped and the next is searched.
824 BibTeXInfo keyvalmap(key, entryType);
826 bool readNext = removeWSAndComma(ifs);
828 while (ifs && readNext) {
831 if (!readTypeOrKey(name, ifs, from_ascii("="),
832 from_ascii("{}(),"), makeLowerCase) || !ifs)
835 // next char must be an equal sign
836 // FIXME Whitespace??
839 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
843 lyxerr << "BibTeX Parser: Missing `=' after field name: " <<
844 name << ", for key: " << key << "." << std::endl;
850 if (!readValue(value, ifs, strings)) {
851 lyxerr << "BibTeX Parser: Unable to read value for field: " <<
852 name << ", for key: " << key << "." << std::endl;
856 keyvalmap[name] = value;
857 data += "\n\n" + value;
858 keylist.addFieldName(name);
859 readNext = removeWSAndComma(ifs);
863 keylist.addEntryType(entryType);
864 keyvalmap.setAllData(data);
865 keylist[key] = keyvalmap;
866 } //< else (citation entry)
868 } //< for loop over files
870 buffer().addBiblioInfo(keylist);
874 FileName InsetBibtex::getBibTeXPath(docstring const & filename, Buffer const & buf)
876 string texfile = changeExtension(to_utf8(filename), "bib");
877 // note that, if the filename can be found directly from the path,
878 // findtexfile will just return a FileName object for that path.
879 FileName file(findtexfile(texfile, "bib"));
881 file = FileName(makeAbsPath(texfile, buf.filePath()));
886 bool InsetBibtex::addDatabase(docstring const & db)
888 docstring bibfiles = getParam("bibfiles");
889 if (tokenPos(bibfiles, ',', db) != -1)
891 if (!bibfiles.empty())
893 setParam("bibfiles", bibfiles + db);
898 bool InsetBibtex::delDatabase(docstring const & db)
900 docstring bibfiles = getParam("bibfiles");
901 if (contains(bibfiles, db)) {
902 int const n = tokenPos(bibfiles, ',', db);
905 // this is not the first database
906 docstring tmp = ',' + bd;
907 setParam("bibfiles", subst(bibfiles, tmp, docstring()));
909 // this is the first (or only) database
910 setParam("bibfiles", split(bibfiles, bd, ','));
918 void InsetBibtex::validate(LaTeXFeatures & features) const
920 if (features.bufferParams().use_bibtopic)
921 features.require("bibtopic");
923 // It'd be better to be able to get this from an InsetLayout, but at present
924 // InsetLayouts do not seem really to work for things that aren't InsetTexts.
925 if (features.runparams().flavor == OutputParams::HTML)
926 features.addCSSSnippet("div.bibtexentry { margin-left: 2em; text-indent: -2em; }\n"
927 "span.bibtexlabel:before{ content: \"[\"; }\n"
928 "span.bibtexlabel:after{ content: \"] \"; }");
933 // docstring InsetBibtex::entriesAsXHTML(vector<docstring> const & entries)
934 // And then here just: entriesAsXHTML(buffer().masterBibInfo().citedEntries())
935 docstring InsetBibtex::xhtml(XHTMLStream & xs, OutputParams const &) const
937 BiblioInfo const & bibinfo = buffer().masterBibInfo();
938 vector<docstring> const & cites = bibinfo.citedEntries();
939 CiteEngine const engine = buffer().params().citeEngine();
941 (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL);
943 docstring reflabel = from_ascii("References");
944 Language const * l = buffer().params().language;
946 reflabel = translateIfPossible(reflabel, l->code());
948 xs << html::StartTag("h2", "class='bibtex'")
950 << html::EndTag("h2")
951 << html::StartTag("div", "class='bibtex'");
953 // Now we loop over the entries
954 vector<docstring>::const_iterator vit = cites.begin();
955 vector<docstring>::const_iterator const ven = cites.end();
956 for (; vit != ven; ++vit) {
957 BiblioInfo::const_iterator const biit = bibinfo.find(*vit);
958 if (biit == bibinfo.end())
960 BibTeXInfo const & entry = biit->second;
961 xs << html::StartTag("div", "class='bibtexentry'");
963 // The same name/id problem we have elsewhere.
964 string const attr = "id='" + to_utf8(entry.key()) + "'";
965 xs << html::CompTag("a", attr);
968 citekey = entry.citeNumber();
970 docstring const auth = entry.getAbbreviatedAuthor();
971 // we do it this way so as to access the xref, if necessary
972 // note that this also gives us the modifier
973 docstring const year = bibinfo.getYear(*vit, true);
974 if (!auth.empty() && !year.empty())
975 citekey = auth + ' ' + year;
977 if (citekey.empty()) {
978 citekey = entry.label();
980 citekey = entry.key();
982 xs << html::StartTag("span", "class='bibtexlabel'")
984 << html::EndTag("span");
985 // FIXME Right now, we are calling BibInfo::getInfo on the key,
986 // which will give us all the cross-referenced info. But for every
987 // entry, so there's a lot of repitition. This should be fixed.
988 xs << html::StartTag("span", "class='bibtexinfo'")
989 << XHTMLStream::ESCAPE_AND
990 << bibinfo.getInfo(entry.key(), buffer(), true)
991 << html::EndTag("span")
992 << html::EndTag("div")
995 xs << html::EndTag("div");
1000 string InsetBibtex::contextMenuName() const
1002 return "context-bibtex";