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"
23 #include "FuncRequest.h"
24 #include "FuncStatus.h"
26 #include "LaTeXFeatures.h"
27 #include "output_xhtml.h"
28 #include "OutputParams.h"
29 #include "PDFOptions.h"
30 #include "TextClass.h"
32 #include "frontends/alert.h"
34 #include "support/convert.h"
35 #include "support/debug.h"
36 #include "support/docstream.h"
37 #include "support/ExceptionMessage.h"
38 #include "support/FileNameList.h"
39 #include "support/filetools.h"
40 #include "support/gettext.h"
41 #include "support/lstrings.h"
42 #include "support/os.h"
43 #include "support/Path.h"
44 #include "support/textutils.h"
49 using namespace lyx::support;
53 namespace Alert = frontend::Alert;
54 namespace os = support::os;
57 InsetBibtex::InsetBibtex(Buffer * buf, InsetCommandParams const & p)
58 : InsetCommand(buf, p)
60 buffer().invalidateBibinfoCache();
64 InsetBibtex::~InsetBibtex()
67 buffer().invalidateBibfileCache();
71 ParamInfo const & InsetBibtex::findInfo(string const & /* cmdName */)
73 static ParamInfo param_info_;
74 if (param_info_.empty()) {
75 param_info_.add("btprint", ParamInfo::LATEX_OPTIONAL);
76 param_info_.add("bibfiles", ParamInfo::LATEX_REQUIRED);
77 param_info_.add("options", ParamInfo::LYX_INTERNAL);
83 void InsetBibtex::doDispatch(Cursor & cur, FuncRequest & cmd)
85 switch (cmd.action()) {
91 case LFUN_INSET_MODIFY: {
92 InsetCommandParams p(BIBTEX_CODE);
94 if (!InsetCommand::string2params(to_utf8(cmd.argument()), p)) {
98 } catch (ExceptionMessage const & message) {
99 if (message.type_ == WarningException) {
100 Alert::warning(message.title_, message.details_);
101 cur.noScreenUpdate();
108 buffer().invalidateBibfileCache();
109 cur.forceBufferUpdate();
114 InsetCommand::doDispatch(cur, cmd);
120 bool InsetBibtex::getStatus(Cursor & cur, FuncRequest const & cmd,
121 FuncStatus & flag) const
123 switch (cmd.action()) {
124 case LFUN_INSET_EDIT:
125 flag.setEnabled(true);
129 return InsetCommand::getStatus(cur, cmd, flag);
134 void InsetBibtex::editDatabases() const
136 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
138 if (bibfilelist.empty())
141 int nr_databases = bibfilelist.size();
142 if (nr_databases > 1) {
143 docstring message = bformat(_("The BibTeX inset includes %1$s databases.\n"
144 "If you proceed, all of them will be opened."),
145 convert<docstring>(nr_databases));
146 int const ret = Alert::prompt(_("Open Databases?"),
147 message, 0, 1, _("&Cancel"), _("&Proceed"));
153 vector<docstring>::const_iterator it = bibfilelist.begin();
154 vector<docstring>::const_iterator en = bibfilelist.end();
155 for (; it != en; ++it) {
156 FileName const bibfile = getBibTeXPath(*it, buffer());
157 formats.edit(buffer(), bibfile,
158 formats.getFormatFromFile(bibfile));
163 docstring InsetBibtex::screenLabel() const
165 return _("BibTeX Generated Bibliography");
169 docstring InsetBibtex::toolTip(BufferView const & /*bv*/, int /*x*/, int /*y*/) const
171 docstring item = from_ascii("* ");
172 docstring tip = _("Databases:") + "\n";
173 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
175 if (bibfilelist.empty()) {
179 vector<docstring>::const_iterator it = bibfilelist.begin();
180 vector<docstring>::const_iterator en = bibfilelist.end();
181 for (; it != en; ++it) {
189 docstring style = getParam("options"); // maybe empty! and with bibtotoc
190 docstring bibtotoc = from_ascii("bibtotoc");
191 if (prefixIs(style, bibtotoc)) {
193 if (contains(style, char_type(',')))
194 style = split(style, bibtotoc, char_type(','));
197 tip += _("Style File:") +"\n";
204 tip += "\n" + _("Lists:") + " ";
205 docstring btprint = getParam("btprint");
206 if (btprint == "btPrintAll")
207 tip += _("all references");
208 else if (btprint == "btPrintNotCited")
209 tip += _("all uncited references");
211 tip += _("all cited references");
215 tip += _("included in TOC");
222 static string normalizeName(Buffer const & buffer,
223 OutputParams const & runparams, string const & name, string const & ext)
225 string const fname = makeAbsPath(name, buffer.filePath()).absFileName();
226 if (FileName::isAbsolute(name) || !FileName(fname + ext).isReadableFile())
232 return to_utf8(makeRelPath(from_utf8(fname),
233 from_utf8(buffer.masterBuffer()->filePath())));
237 int InsetBibtex::latex(odocstream & os, OutputParams const & runparams) const
239 // the sequence of the commands:
240 // 1. \bibliographystyle{style}
241 // 2. \addcontentsline{...} - if option bibtotoc set
242 // 3. \bibliography{database}
243 // and with bibtopic:
244 // 1. \bibliographystyle{style}
245 // 2. \begin{btSect}{database}
246 // 3. \btPrint{Cited|NotCited|All}
250 // If we are processing the LaTeX file in a temp directory then
251 // copy the .bib databases to this temp directory, mangling their
252 // names in the process. Store this mangled name in the list of
254 // (We need to do all this because BibTeX *really*, *really*
255 // can't handle "files with spaces" and Windows users tend to
256 // use such filenames.)
257 // Otherwise, store the (maybe absolute) path to the original,
258 // unmangled database name.
259 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
260 vector<docstring>::const_iterator it = bibfilelist.begin();
261 vector<docstring>::const_iterator en = bibfilelist.end();
262 odocstringstream dbs;
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());
281 bool const success = in_file.copyTo(out_file);
283 lyxerr << "Failed to copy '" << in_file
284 << "' to '" << out_file << "'"
287 } else if (!runparams.inComment && runparams.nice && not_from_texmf) {
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, ',');
333 if (!style.empty()) {
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")))
363 // Post this warning only once.
364 static bool warned_about_bst_spaces = false;
365 if (!warned_about_bst_spaces && runparams.nice && contains(style, ' ')) {
366 warned_about_bst_spaces = true;
367 Alert::warning(_("Export Warning!"),
368 _("There are spaces in the path to your BibTeX style file.\n"
369 "BibTeX will be unable to find it."));
372 if (!db_out.empty() && buffer().params().use_bibtopic) {
373 os << "\\begin{btSect}{" << db_out << "}\n";
374 docstring btprint = getParam("btprint");
377 btprint = from_ascii("btPrintCited");
378 os << "\\" << btprint << "\n"
379 << "\\end{btSect}\n";
384 if (!bibtotoc.empty() && !buffer().params().use_bibtopic) {
385 // set label for hyperref, see http://www.lyx.org/trac/ticket/6470
386 if (buffer().params().pdfoptions().use_hyperref)
387 os << "\\phantomsection";
388 if (buffer().params().documentClass().hasLaTeXLayout("chapter"))
389 os << "\\addcontentsline{toc}{chapter}{\\bibname}";
390 else if (buffer().params().documentClass().hasLaTeXLayout("section"))
391 os << "\\addcontentsline{toc}{section}{\\refname}";
394 if (!db_out.empty() && !buffer().params().use_bibtopic) {
395 docstring btprint = getParam("btprint");
396 if (btprint == "btPrintAll") {
397 os << "\\nocite{*}\n";
400 os << "\\bibliography{" << db_out << "}\n";
408 support::FileNameList InsetBibtex::getBibFiles() const
410 FileName path(buffer().filePath());
411 support::PathChanger p(path);
413 support::FileNameList vec;
415 vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
416 vector<docstring>::const_iterator it = bibfilelist.begin();
417 vector<docstring>::const_iterator en = bibfilelist.end();
418 for (; it != en; ++it) {
419 FileName const file = getBibTeXPath(*it, buffer());
424 LYXERR0("Couldn't find " + to_utf8(*it) + " in InsetBibtex::getBibFiles()!");
433 // methods for parsing bibtex files
435 typedef map<docstring, docstring> VarMap;
437 /// remove whitespace characters, optionally a single comma,
438 /// and further whitespace characters from the stream.
439 /// @return true if a comma was found, false otherwise
441 bool removeWSAndComma(ifdocstream & ifs) {
450 } while (ifs && isSpace(ch));
463 } while (ifs && isSpace(ch));
478 /// remove whitespace characters, read characer sequence
479 /// not containing whitespace characters or characters in
480 /// delimChars, and remove further whitespace characters.
482 /// @return true if a string of length > 0 could be read.
484 bool readTypeOrKey(docstring & val, ifdocstream & ifs,
485 docstring const & delimChars, docstring const & illegalChars,
498 } while (ifs && isSpace(ch));
504 bool legalChar = true;
505 while (ifs && !isSpace(ch) &&
506 delimChars.find(ch) == docstring::npos &&
507 (legalChar = (illegalChars.find(ch) == docstring::npos))
510 if (chCase == makeLowerCase)
511 val += lowercase(ch);
523 while (ifs && isSpace(ch)) {
531 return val.length() > 0;
534 /// read subsequent bibtex values that are delimited with a #-character.
535 /// Concatenate all parts and replace names with the associated string in
536 /// the variable strings.
537 /// @return true if reading was successfull (all single parts were delimited
539 bool readValue(docstring & val, ifdocstream & ifs, const VarMap & strings) {
552 } while (ifs && isSpace(ch));
557 // check for field type
560 // read integer value
564 } while (ifs && isDigit(ch));
569 } else if (ch == '"' || ch == '{') {
571 char_type delim = ch == '"' ? '"': '}';
576 } while (ifs && isSpace(ch));
581 // We now have the first non-whitespace character
582 // We'll collapse adjacent whitespace.
583 bool lastWasWhiteSpace = false;
585 // inside this delimited text braces must match.
586 // Thus we can have a closing delimiter only
587 // when nestLevel == 0
590 while (ifs && (nestLevel > 0 || ch != delim)) {
592 lastWasWhiteSpace = true;
596 // We output the space only after we stop getting
597 // whitespace so as not to output any whitespace
598 // at the end of the value.
599 if (lastWasWhiteSpace) {
600 lastWasWhiteSpace = false;
606 // update nesting level
625 // FIXME Why is this here?
633 // reading a string name
636 while (ifs && !isSpace(ch) && ch != '#' && ch != ',' && ch != '}' && ch != ')') {
637 strName += lowercase(ch);
644 // replace the string with its assigned value or
645 // discard it if it's not assigned
646 if (strName.length()) {
647 VarMap::const_iterator pos = strings.find(strName);
648 if (pos != strings.end()) {
655 while (ifs && isSpace(ch)) {
662 // continue reading next value on concatenate with '#'
672 // This method returns a comma separated list of Bibtex entries
673 void InsetBibtex::fillWithBibKeys(BiblioInfo & keylist,
674 InsetIterator const & /*di*/) const
676 // This bibtex parser is a first step to parse bibtex files
679 // - it reads the whole bibtex entry and does a syntax check
680 // (matching delimiters, missing commas,...
681 // - it recovers from errors starting with the next @-character
682 // - it reads @string definitions and replaces them in the
684 // - it accepts more characters in keys or value names than
687 // Officially bibtex does only support ASCII, but in practice
688 // you can use the encoding of the main document as long as
689 // some elements like keys and names are pure ASCII. Therefore
690 // we convert the file from the buffer encoding.
691 // We don't restrict keys to ASCII in LyX, since our own
692 // InsetBibitem can generate non-ASCII keys, and nonstandard
693 // 8bit clean bibtex forks exist.
694 support::FileNameList const files = getBibFiles();
695 support::FileNameList::const_iterator it = files.begin();
696 support::FileNameList::const_iterator en = files.end();
697 for (; it != en; ++ it) {
698 ifdocstream ifs(it->toFilesystemEncoding().c_str(),
699 ios_base::in, buffer().params().encoding().iconvName());
715 if (!readTypeOrKey(entryType, ifs, from_ascii("{("), docstring(), makeLowerCase)) {
716 lyxerr << "BibTeX Parser: Error reading entry type." << std::endl;
721 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
725 if (entryType == from_ascii("comment")) {
726 ifs.ignore(numeric_limits<int>::max(), '\n');
732 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
736 if ((ch != '(') && (ch != '{')) {
737 lyxerr << "BibTeX Parser: Invalid entry delimiter." << std::endl;
743 if (entryType == from_ascii("string")) {
745 // read string and add it to the strings map
746 // (or replace it's old value)
750 if (!readTypeOrKey(name, ifs, from_ascii("="), from_ascii("#{}(),"), makeLowerCase)) {
751 lyxerr << "BibTeX Parser: Error reading string name." << std::endl;
756 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
760 // next char must be an equal sign
762 if (!ifs || ch != '=') {
763 lyxerr << "BibTeX Parser: No `=' after string name: " <<
764 name << "." << std::endl;
768 if (!readValue(value, ifs, strings)) {
769 lyxerr << "BibTeX Parser: Unable to read value for string: " <<
770 name << "." << std::endl;
774 strings[name] = value;
776 } else if (entryType == from_ascii("preamble")) {
778 // preamble definitions are discarded.
779 // can they be of any use in lyx?
782 if (!readValue(value, ifs, strings)) {
783 lyxerr << "BibTeX Parser: Unable to read preamble value." << std::endl;
789 // Citation entry. Try to read the key.
792 if (!readTypeOrKey(key, ifs, from_ascii(","), from_ascii("}"), keepCase)) {
793 lyxerr << "BibTeX Parser: Unable to read key for entry type:" <<
794 entryType << "." << std::endl;
799 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
803 /////////////////////////////////////////////
804 // now we have a key, so we will add an entry
805 // (even if it's empty, as bibtex does)
807 // we now read the field = value pairs.
808 // all items must be separated by a comma. If
809 // it is missing the scanning of this entry is
810 // stopped and the next is searched.
814 BibTeXInfo keyvalmap(key, entryType);
816 bool readNext = removeWSAndComma(ifs);
818 while (ifs && readNext) {
821 if (!readTypeOrKey(name, ifs, from_ascii("="),
822 from_ascii("{}(),"), makeLowerCase) || !ifs)
825 // next char must be an equal sign
826 // FIXME Whitespace??
829 lyxerr << "BibTeX Parser: Unexpected end of file." << std::endl;
833 lyxerr << "BibTeX Parser: Missing `=' after field name: " <<
834 name << ", for key: " << key << "." << std::endl;
840 if (!readValue(value, ifs, strings)) {
841 lyxerr << "BibTeX Parser: Unable to read value for field: " <<
842 name << ", for key: " << key << "." << std::endl;
846 keyvalmap[name] = value;
847 data += "\n\n" + value;
848 keylist.addFieldName(name);
849 readNext = removeWSAndComma(ifs);
853 keylist.addEntryType(entryType);
854 keyvalmap.setAllData(data);
855 keylist[key] = keyvalmap;
856 } //< else (citation entry)
858 } //< for loop over files
862 FileName InsetBibtex::getBibTeXPath(docstring const & filename, Buffer const & buf)
864 string texfile = changeExtension(to_utf8(filename), "bib");
865 // note that, if the filename can be found directly from the path,
866 // findtexfile will just return a FileName object for that path.
867 FileName file(findtexfile(texfile, "bib"));
869 file = FileName(makeAbsPath(texfile, buf.filePath()));
874 bool InsetBibtex::addDatabase(docstring const & db)
876 docstring bibfiles = getParam("bibfiles");
877 if (tokenPos(bibfiles, ',', db) != -1)
879 if (!bibfiles.empty())
881 setParam("bibfiles", bibfiles + db);
886 bool InsetBibtex::delDatabase(docstring const & db)
888 docstring bibfiles = getParam("bibfiles");
889 if (contains(bibfiles, db)) {
890 int const n = tokenPos(bibfiles, ',', db);
893 // this is not the first database
894 docstring tmp = ',' + bd;
895 setParam("bibfiles", subst(bibfiles, tmp, docstring()));
897 // this is the first (or only) database
898 setParam("bibfiles", split(bibfiles, bd, ','));
906 void InsetBibtex::validate(LaTeXFeatures & features) const
908 if (features.bufferParams().use_bibtopic)
909 features.require("bibtopic");
911 // It'd be better to be able to get this from an InsetLayout, but at present
912 // InsetLayouts do not seem really to work for things that aren't InsetTexts.
913 if (features.runparams().flavor == OutputParams::HTML)
914 features.addPreambleSnippet("<style type=\"text/css\">\n"
915 "div.bibtexentry { margin-left: 2em; text-indent: -2em; }\n"
916 "span.bibtexlabel:before{ content: \"[\"; }\n"
917 "span.bibtexlabel:after{ content: \"] \"; }\n"
923 // docstring InsetBibtex::entriesAsXHTML(vector<docstring> const & entries)
924 // And then here just: entriesAsXHTML(buffer().masterBibInfo().citedEntries())
925 docstring InsetBibtex::xhtml(XHTMLStream & xs, OutputParams const &) const
927 BiblioInfo const & bibinfo = buffer().masterBibInfo();
928 vector<docstring> const & cites = bibinfo.citedEntries();
929 CiteEngine const engine = buffer().params().citeEngine();
931 (engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL);
933 docstring reflabel = from_ascii("References");
934 Language const * l = buffer().params().language;
936 reflabel = translateIfPossible(reflabel, l->code());
938 xs << html::StartTag("h2", "class='bibtex'")
940 << html::EndTag("h2")
941 << html::StartTag("div", "class='bibtex'");
943 // Now we loop over the entries
944 vector<docstring>::const_iterator vit = cites.begin();
945 vector<docstring>::const_iterator const ven = cites.end();
946 for (; vit != ven; ++vit) {
947 BiblioInfo::const_iterator const biit = bibinfo.find(*vit);
948 if (biit == bibinfo.end())
950 BibTeXInfo const & entry = biit->second;
951 xs << html::StartTag("div", "class='bibtexentry'");
953 // The same name/id problem we have elsewhere.
954 string const attr = "id='" + to_utf8(entry.key()) + "'";
955 xs << html::CompTag("a", attr);
958 citekey = entry.citeNumber();
960 docstring const auth = entry.getAbbreviatedAuthor();
961 // we do it this way so as to access the xref, if necessary
962 // note that this also gives us the modifier
963 docstring const year = bibinfo.getYear(*vit, true);
964 if (!auth.empty() && !year.empty())
965 citekey = auth + ' ' + year;
967 if (citekey.empty()) {
968 citekey = entry.label();
970 citekey = entry.key();
972 xs << html::StartTag("span", "class='bibtexlabel'")
974 << html::EndTag("span");
975 // FIXME Right now, we are calling BibInfo::getInfo on the key,
976 // which will give us all the cross-referenced info. But for every
977 // entry, so there's a lot of repitition. This should be fixed.
978 xs << html::StartTag("span", "class='bibtexinfo'")
979 << XHTMLStream::ESCAPE_AND
980 << bibinfo.getInfo(entry.key(), buffer(), true)
981 << html::EndTag("span")
982 << html::EndTag("div");
985 xs << html::EndTag("div");
990 docstring InsetBibtex::contextMenu(BufferView const &, int, int) const
992 return from_ascii("context-bibtex");