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"
26 #include "LaTeXFeatures.h"
27 #include "output_xhtml.h"
28 #include "OutputParams.h"
29 #include "PDFOptions.h"
30 #include "texstream.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/PathChanger.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().invalidateBibfileCache();
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 tip = _("Databases:");
178 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
181 if (bibfilelist.empty())
182 tip += "<li>" + _("none") + "</li>";
184 for (docstring const & bibfile : bibfilelist)
185 tip += "<li>" + bibfile + "</li>";
190 docstring style = getParam("options"); // maybe empty! and with bibtotoc
191 docstring bibtotoc = from_ascii("bibtotoc");
192 if (prefixIs(style, bibtotoc)) {
194 if (contains(style, char_type(',')))
195 style = split(style, bibtotoc, char_type(','));
198 tip += _("Style File:");
199 tip += "<ul><li>" + (style.empty() ? _("none") : style) + "</li></ul>";
201 tip += _("Lists:") + " ";
202 docstring btprint = getParam("btprint");
203 if (btprint == "btPrintAll")
204 tip += _("all references");
205 else if (btprint == "btPrintNotCited")
206 tip += _("all uncited references");
208 tip += _("all cited references");
212 tip += _("included in TOC");
219 static string normalizeName(Buffer const & buffer,
220 OutputParams const & runparams, string const & name, string const & ext)
222 string const fname = makeAbsPath(name, buffer.filePath()).absFileName();
223 if (FileName::isAbsolute(name) || !FileName(fname + ext).isReadableFile())
229 return to_utf8(makeRelPath(from_utf8(fname),
230 from_utf8(buffer.masterBuffer()->filePath())));
234 void InsetBibtex::latex(otexstream & os, OutputParams const & runparams) const
236 // the sequence of the commands:
237 // 1. \bibliographystyle{style}
238 // 2. \addcontentsline{...} - if option bibtotoc set
239 // 3. \bibliography{database}
240 // and with bibtopic:
241 // 1. \bibliographystyle{style}
242 // 2. \begin{btSect}{database}
243 // 3. \btPrint{Cited|NotCited|All}
247 // If we are processing the LaTeX file in a temp directory then
248 // copy the .bib databases to this temp directory, mangling their
249 // names in the process. Store this mangled name in the list of
251 // (We need to do all this because BibTeX *really*, *really*
252 // can't handle "files with spaces" and Windows users tend to
253 // use such filenames.)
254 // Otherwise, store the (maybe absolute) path to the original,
255 // unmangled database name.
256 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
257 vector<docstring>::const_iterator it = bibfilelist.begin();
258 vector<docstring>::const_iterator en = bibfilelist.end();
259 odocstringstream dbs;
262 // determine the export format
263 string const tex_format = flavor2format(runparams.flavor);
265 for (; it != en; ++it) {
266 string utf8input = to_utf8(*it);
268 normalizeName(buffer(), runparams, utf8input, ".bib");
269 FileName const try_in_file =
270 makeAbsPath(database + ".bib", buffer().filePath());
271 bool const not_from_texmf = try_in_file.isReadableFile();
273 if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
275 // mangledFileName() needs the extension
276 DocFileName const in_file = DocFileName(try_in_file);
277 database = removeExtension(in_file.mangledFileName());
278 FileName const out_file = makeAbsPath(database + ".bib",
279 buffer().masterBuffer()->temppath());
280 bool const success = in_file.copyTo(out_file);
282 lyxerr << "Failed to copy '" << in_file
283 << "' to '" << out_file << "'"
286 } else if (!runparams.inComment && runparams.nice && not_from_texmf) {
287 runparams.exportdata->addExternalFile(tex_format, try_in_file, database + ".bib");
288 if (!isValidLaTeXFileName(database)) {
289 frontend::Alert::warning(_("Invalid filename"),
290 _("The following filename will cause troubles "
291 "when running the exported file through LaTeX: ") +
292 from_utf8(database));
294 if (!isValidDVIFileName(database)) {
295 frontend::Alert::warning(_("Problematic filename for DVI"),
296 _("The following filename can cause troubles "
297 "when running the exported file through LaTeX "
298 "and opening the resulting DVI: ") +
299 from_utf8(database), true);
308 dbs << from_utf8(latex_path(database));
310 docstring const db_out = dbs.str();
312 // Post this warning only once.
313 static bool warned_about_spaces = false;
314 if (!warned_about_spaces &&
315 runparams.nice && db_out.find(' ') != docstring::npos) {
316 warned_about_spaces = true;
317 Alert::warning(_("Export Warning!"),
318 _("There are spaces in the paths to your BibTeX databases.\n"
319 "BibTeX will be unable to find them."));
322 string style = to_utf8(getParam("options")); // maybe empty! and with bibtotoc
324 if (prefixIs(style, "bibtotoc")) {
325 bibtotoc = "bibtotoc";
326 if (contains(style, ','))
327 style = split(style, bibtotoc, ',');
330 if (style == "default")
331 style = buffer().params().biblio_style;
333 if (!style.empty() && !buffer().params().use_bibtopic) {
334 string base = normalizeName(buffer(), runparams, style, ".bst");
335 FileName const try_in_file =
336 makeAbsPath(base + ".bst", buffer().filePath());
337 bool const not_from_texmf = try_in_file.isReadableFile();
338 // If this style does not come from texmf and we are not
339 // exporting to .tex copy it to the tmp directory.
340 // This prevents problems with spaces and 8bit charcaters
342 if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
344 // use new style name
345 DocFileName const in_file = DocFileName(try_in_file);
346 base = removeExtension(in_file.mangledFileName());
347 FileName const out_file = makeAbsPath(base + ".bst",
348 buffer().masterBuffer()->temppath());
349 bool const success = in_file.copyTo(out_file);
351 lyxerr << "Failed to copy '" << in_file
352 << "' to '" << out_file << "'"
357 os << "\\bibliographystyle{"
358 << from_utf8(latex_path(normalizeName(buffer(), runparams, base, ".bst")))
362 // Post this warning only once.
363 static bool warned_about_bst_spaces = false;
364 if (!warned_about_bst_spaces && runparams.nice && contains(style, ' ')) {
365 warned_about_bst_spaces = true;
366 Alert::warning(_("Export Warning!"),
367 _("There are spaces in the path to your BibTeX style file.\n"
368 "BibTeX will be unable to find it."));
371 if (!db_out.empty() && buffer().params().use_bibtopic) {
372 os << "\\begin{btSect}";
374 os << "[" << style << "]";
375 os << "{" << db_out << "}\n";
376 docstring btprint = getParam("btprint");
379 btprint = from_ascii("btPrintCited");
380 os << "\\" << btprint << "\n"
381 << "\\end{btSect}\n";
385 if (!bibtotoc.empty() && !buffer().params().use_bibtopic) {
386 // set label for hyperref, see http://www.lyx.org/trac/ticket/6470
387 if (buffer().params().pdfoptions().use_hyperref)
388 os << "\\phantomsection";
389 if (buffer().params().documentClass().hasLaTeXLayout("chapter"))
390 os << "\\addcontentsline{toc}{chapter}{\\bibname}";
391 else if (buffer().params().documentClass().hasLaTeXLayout("section"))
392 os << "\\addcontentsline{toc}{section}{\\refname}";
395 if (!db_out.empty() && !buffer().params().use_bibtopic) {
396 docstring btprint = getParam("btprint");
397 if (btprint == "btPrintAll") {
398 os << "\\nocite{*}\n";
400 os << "\\bibliography{" << db_out << "}\n";
405 support::FileNameList InsetBibtex::getBibFiles() const
407 FileName path(buffer().filePath());
408 support::PathChanger p(path);
410 support::FileNameList vec;
412 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
413 vector<docstring>::const_iterator it = bibfilelist.begin();
414 vector<docstring>::const_iterator en = bibfilelist.end();
415 for (; it != en; ++it) {
416 FileName const file = getBibTeXPath(*it, buffer());
421 LYXERR0("Couldn't find " + to_utf8(*it) + " in InsetBibtex::getBibFiles()!");
430 // methods for parsing bibtex files
432 typedef map<docstring, docstring> VarMap;
434 /// remove whitespace characters, optionally a single comma,
435 /// and further whitespace characters from the stream.
436 /// @return true if a comma was found, false otherwise
438 bool removeWSAndComma(ifdocstream & ifs) {
447 } while (ifs && isSpace(ch));
460 } while (ifs && isSpace(ch));
475 /// remove whitespace characters, read characer sequence
476 /// not containing whitespace characters or characters in
477 /// delimChars, and remove further whitespace characters.
479 /// @return true if a string of length > 0 could be read.
481 bool readTypeOrKey(docstring & val, ifdocstream & ifs,
482 docstring const & delimChars, docstring const & illegalChars,
495 } while (ifs && isSpace(ch));
501 bool legalChar = true;
502 while (ifs && !isSpace(ch) &&
503 delimChars.find(ch) == docstring::npos &&
504 (legalChar = (illegalChars.find(ch) == docstring::npos))
507 if (chCase == makeLowerCase)
508 val += lowercase(ch);
520 while (ifs && isSpace(ch)) {
528 return val.length() > 0;
531 /// read subsequent bibtex values that are delimited with a #-character.
532 /// Concatenate all parts and replace names with the associated string in
533 /// the variable strings.
534 /// @return true if reading was successfull (all single parts were delimited
536 bool readValue(docstring & val, ifdocstream & ifs, const VarMap & strings) {
549 } while (ifs && isSpace(ch));
554 // check for field type
555 if (isDigitASCII(ch)) {
557 // read integer value
561 } while (ifs && isDigitASCII(ch));
566 } else if (ch == '"' || ch == '{') {
568 char_type delim = ch == '"' ? '"': '}';
573 } while (ifs && isSpace(ch));
578 // We now have the first non-whitespace character
579 // We'll collapse adjacent whitespace.
580 bool lastWasWhiteSpace = false;
582 // inside this delimited text braces must match.
583 // Thus we can have a closing delimiter only
584 // when nestLevel == 0
587 while (ifs && (nestLevel > 0 || ch != delim)) {
589 lastWasWhiteSpace = true;
593 // We output the space only after we stop getting
594 // whitespace so as not to output any whitespace
595 // at the end of the value.
596 if (lastWasWhiteSpace) {
597 lastWasWhiteSpace = false;
603 // update nesting level
622 // FIXME Why is this here?
630 // reading a string name
633 while (ifs && !isSpace(ch) && ch != '#' && ch != ',' && ch != '}' && ch != ')') {
634 strName += lowercase(ch);
641 // replace the string with its assigned value or
642 // discard it if it's not assigned
643 if (strName.length()) {
644 VarMap::const_iterator pos = strings.find(strName);
645 if (pos != strings.end()) {
652 while (ifs && isSpace(ch)) {
659 // continue reading next value on concatenate with '#'
669 void InsetBibtex::collectBibKeys(InsetIterator const & /*di*/) const
675 void InsetBibtex::parseBibTeXFiles() const
677 // This bibtex parser is a first step to parse bibtex files
680 // - it reads the whole bibtex entry and does a syntax check
681 // (matching delimiters, missing commas,...
682 // - it recovers from errors starting with the next @-character
683 // - it reads @string definitions and replaces them in the
685 // - it accepts more characters in keys or value names than
688 // Officially bibtex does only support ASCII, but in practice
689 // you can use the encoding of the main document as long as
690 // some elements like keys and names are pure ASCII. Therefore
691 // we convert the file from the buffer encoding.
692 // We don't restrict keys to ASCII in LyX, since our own
693 // InsetBibitem can generate non-ASCII keys, and nonstandard
694 // 8bit clean bibtex forks exist.
698 support::FileNameList const files = getBibFiles();
699 support::FileNameList::const_iterator it = files.begin();
700 support::FileNameList::const_iterator en = files.end();
701 for (; it != en; ++ it) {
702 ifdocstream ifs(it->toFilesystemEncoding().c_str(),
703 ios_base::in, buffer().params().encoding().iconvName());
719 if (!readTypeOrKey(entryType, ifs, from_ascii("{("), docstring(), makeLowerCase)) {
720 lyxerr << "BibTeX Parser: Error reading entry type." << std::endl;
725 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
729 if (entryType == from_ascii("comment")) {
730 ifs.ignore(numeric_limits<int>::max(), '\n');
736 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
740 if ((ch != '(') && (ch != '{')) {
741 lyxerr << "BibTeX Parser: Invalid entry delimiter." << std::endl;
747 if (entryType == from_ascii("string")) {
749 // read string and add it to the strings map
750 // (or replace it's old value)
754 if (!readTypeOrKey(name, ifs, from_ascii("="), from_ascii("#{}(),"), makeLowerCase)) {
755 lyxerr << "BibTeX Parser: Error reading string name." << std::endl;
760 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
764 // next char must be an equal sign
766 if (!ifs || ch != '=') {
767 lyxerr << "BibTeX Parser: No `=' after string name: " <<
768 name << "." << std::endl;
772 if (!readValue(value, ifs, strings)) {
773 lyxerr << "BibTeX Parser: Unable to read value for string: " <<
774 name << "." << std::endl;
778 strings[name] = value;
780 } else if (entryType == from_ascii("preamble")) {
782 // preamble definitions are discarded.
783 // can they be of any use in lyx?
786 if (!readValue(value, ifs, strings)) {
787 lyxerr << "BibTeX Parser: Unable to read preamble value." << std::endl;
793 // Citation entry. Try to read the key.
796 if (!readTypeOrKey(key, ifs, from_ascii(","), from_ascii("}"), keepCase)) {
797 lyxerr << "BibTeX Parser: Unable to read key for entry type:" <<
798 entryType << "." << std::endl;
803 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
807 /////////////////////////////////////////////
808 // now we have a key, so we will add an entry
809 // (even if it's empty, as bibtex does)
811 // we now read the field = value pairs.
812 // all items must be separated by a comma. If
813 // it is missing the scanning of this entry is
814 // stopped and the next is searched.
818 BibTeXInfo keyvalmap(key, entryType);
820 bool readNext = removeWSAndComma(ifs);
822 while (ifs && readNext) {
825 if (!readTypeOrKey(name, ifs, from_ascii("="),
826 from_ascii("{}(),"), makeLowerCase) || !ifs)
829 // next char must be an equal sign
830 // FIXME Whitespace??
833 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
837 lyxerr << "BibTeX Parser: Missing `=' after field name: " <<
838 name << ", for key: " << key << "." << std::endl;
844 if (!readValue(value, ifs, strings)) {
845 lyxerr << "BibTeX Parser: Unable to read value for field: " <<
846 name << ", for key: " << key << "." << std::endl;
850 keyvalmap[name] = value;
851 data += "\n\n" + value;
852 keylist.addFieldName(name);
853 readNext = removeWSAndComma(ifs);
857 keylist.addEntryType(entryType);
858 keyvalmap.setAllData(data);
859 keylist[key] = keyvalmap;
860 } //< else (citation entry)
862 } //< for loop over files
864 buffer().addBiblioInfo(keylist);
868 FileName InsetBibtex::getBibTeXPath(docstring const & filename, Buffer const & buf)
870 string texfile = changeExtension(to_utf8(filename), "bib");
871 // note that, if the filename can be found directly from the path,
872 // findtexfile will just return a FileName object for that path.
873 FileName file(findtexfile(texfile, "bib"));
875 file = FileName(makeAbsPath(texfile, buf.filePath()));
880 bool InsetBibtex::addDatabase(docstring const & db)
882 docstring bibfiles = getParam("bibfiles");
883 if (tokenPos(bibfiles, ',', db) != -1)
885 if (!bibfiles.empty())
887 setParam("bibfiles", bibfiles + db);
892 bool InsetBibtex::delDatabase(docstring const & db)
894 docstring bibfiles = getParam("bibfiles");
895 if (contains(bibfiles, db)) {
896 int const n = tokenPos(bibfiles, ',', db);
899 // this is not the first database
900 docstring tmp = ',' + bd;
901 setParam("bibfiles", subst(bibfiles, tmp, docstring()));
903 // this is the first (or only) database
904 setParam("bibfiles", split(bibfiles, bd, ','));
912 void InsetBibtex::validate(LaTeXFeatures & features) const
914 if (features.bufferParams().use_bibtopic)
915 features.require("bibtopic");
917 // It'd be better to be able to get this from an InsetLayout, but at present
918 // InsetLayouts do not seem really to work for things that aren't InsetTexts.
919 if (features.runparams().flavor == OutputParams::HTML)
920 features.addCSSSnippet("div.bibtexentry { margin-left: 2em; text-indent: -2em; }\n"
921 "span.bibtexlabel:before{ content: \"[\"; }\n"
922 "span.bibtexlabel:after{ content: \"] \"; }");
926 int InsetBibtex::plaintext(odocstringstream & os,
927 OutputParams const & op, size_t max_length) const
929 docstring const reflabel = buffer().B_("References");
931 // We could output more information here, e.g., what databases are included
932 // and information about options. But I don't necessarily see any reason to
933 // do this right now.
934 if (op.for_tooltip || op.for_toc || op.for_search) {
935 os << '[' << reflabel << ']' << '\n';
936 return PLAINTEXT_NEWLINE;
939 BiblioInfo bibinfo = buffer().masterBibInfo();
940 bibinfo.makeCitationLabels(buffer());
941 vector<docstring> const & cites = bibinfo.citedEntries();
943 size_t start_size = os.str().size();
945 refoutput += reflabel + "\n\n";
947 // Now we loop over the entries
948 vector<docstring>::const_iterator vit = cites.begin();
949 vector<docstring>::const_iterator const ven = cites.end();
950 for (; vit != ven; ++vit) {
951 if (start_size + refoutput.size() >= max_length)
953 BiblioInfo::const_iterator const biit = bibinfo.find(*vit);
954 if (biit == bibinfo.end())
956 BibTeXInfo const & entry = biit->second;
957 refoutput += "[" + entry.label() + "] ";
958 // FIXME Right now, we are calling BibInfo::getInfo on the key,
959 // which will give us all the cross-referenced info. But for every
960 // entry, so there's a lot of repitition. This should be fixed.
961 refoutput += bibinfo.getInfo(entry.key(), buffer(), false) + "\n\n";
964 return refoutput.size();
969 // docstring InsetBibtex::entriesAsXHTML(vector<docstring> const & entries)
970 // And then here just: entriesAsXHTML(buffer().masterBibInfo().citedEntries())
971 docstring InsetBibtex::xhtml(XHTMLStream & xs, OutputParams const &) const
973 BiblioInfo const & bibinfo = buffer().masterBibInfo();
974 bool const all_entries = getParam("btprint") == "btPrintAll";
975 vector<docstring> const & cites =
976 all_entries ? bibinfo.getKeys() : bibinfo.citedEntries();
978 docstring const reflabel = buffer().B_("References");
980 xs << html::StartTag("h2", "class='bibtex'")
982 << html::EndTag("h2")
983 << html::StartTag("div", "class='bibtex'");
985 // Now we loop over the entries
986 vector<docstring>::const_iterator vit = cites.begin();
987 vector<docstring>::const_iterator const ven = cites.end();
988 for (; vit != ven; ++vit) {
989 BiblioInfo::const_iterator const biit = bibinfo.find(*vit);
990 if (biit == bibinfo.end())
993 BibTeXInfo const & entry = biit->second;
994 string const attr = "class='bibtexentry' id='LyXCite-"
995 + to_utf8(html::cleanAttr(entry.key())) + "'";
996 xs << html::StartTag("div", attr);
998 // don't print labels if we're outputting all entries
1000 xs << html::StartTag("span", "class='bibtexlabel'")
1002 << html::EndTag("span");
1005 // FIXME Right now, we are calling BibInfo::getInfo on the key,
1006 // which will give us all the cross-referenced info. But for every
1007 // entry, so there's a lot of repitition. This should be fixed.
1008 xs << html::StartTag("span", "class='bibtexinfo'")
1009 << XHTMLStream::ESCAPE_AND
1010 << bibinfo.getInfo(entry.key(), buffer(), true)
1011 << html::EndTag("span")
1012 << html::EndTag("div")
1015 xs << html::EndTag("div");
1020 void InsetBibtex::write(ostream & os) const
1022 params().Write(os, &buffer());
1026 string InsetBibtex::contextMenuName() const
1028 return "context-bibtex";