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"
17 #include "BufferParams.h"
18 #include "DispatchResult.h"
19 #include "support/debug.h"
20 #include "EmbeddedFiles.h"
22 #include "FuncRequest.h"
23 #include "support/gettext.h"
24 #include "LaTeXFeatures.h"
25 #include "MetricsInfo.h"
26 #include "OutputParams.h"
27 #include "TextClass.h"
29 #include "frontends/alert.h"
31 #include "support/ExceptionMessage.h"
32 #include "support/docstream.h"
33 #include "support/filetools.h"
34 #include "support/lstrings.h"
35 #include "support/os.h"
36 #include "support/Path.h"
37 #include "support/textutils.h"
39 #include <boost/tokenizer.hpp>
43 using namespace lyx::support;
47 namespace Alert = frontend::Alert;
48 namespace os = support::os;
51 InsetBibtex::InsetBibtex(InsetCommandParams const & p)
52 : InsetCommand(p, "bibtex")
56 CommandInfo const * InsetBibtex::findInfo(string const & /* cmdName */)
58 static const char * const paramnames[] =
59 {"options", "btprint", "bibfiles", "embed", ""};
60 static const bool isoptional[] = {true, true, false, false};
61 static const CommandInfo info = {4, paramnames, isoptional};
66 Inset * InsetBibtex::clone() const
68 return new InsetBibtex(*this);
72 void InsetBibtex::doDispatch(Cursor & cur, FuncRequest & cmd)
76 case LFUN_INSET_MODIFY: {
77 InsetCommandParams p(BIBTEX_CODE);
79 if (!InsetCommandMailer::string2params("bibtex",
80 to_utf8(cmd.argument()), p)) {
84 } catch (ExceptionMessage const & message) {
85 if (message.type_ == WarningException) {
86 Alert::warning(message.title_, message.details_);
93 InsetCommandParams orig = params();
94 // returned "embed" is composed of "true" or "false", which needs to be adjusted
99 string newEmbedStatus;
101 string bibfiles = to_utf8(p["bibfiles"]);
102 string embedStatus = to_utf8(p["embed"]);
104 bibfiles = split(bibfiles, tmp, ',');
105 embedStatus = split(embedStatus, emb, ',');
106 while (!tmp.empty()) {
107 EmbeddedFile file(changeExtension(tmp, "bib"), cur.buffer().filePath());
108 if (!newBibfiles.empty())
111 if (!newEmbedStatus.empty())
112 newEmbedStatus += ",";
114 newEmbedStatus += file.inzipName();
115 // Get next file name
116 bibfiles = split(bibfiles, tmp, ',');
117 embedStatus = split(embedStatus, emb, ',');
119 LYXERR(Debug::FILES, "Update parameters from " << p["bibfiles"]
120 << " " << p["embed"] << " to " << newBibfiles << " "
122 p["bibfiles"] = from_utf8(newBibfiles);
123 p["embed"] = from_utf8(newEmbedStatus);
127 // test parameter and copy files
128 getFiles(cur.buffer());
129 } catch (ExceptionMessage const & message) {
130 Alert::error(message.title_, message.details_);
131 // do not set parameter if an error happens
135 cur.buffer().updateBibfilesCache();
140 InsetCommand::doDispatch(cur, cmd);
146 docstring const InsetBibtex::getScreenLabel(Buffer const &) const
148 return _("BibTeX Generated Bibliography");
154 string normalizeName(Buffer const & buffer, OutputParams const & runparams,
155 string const & name, string const & ext)
157 string const fname = makeAbsPath(name, buffer.filePath()).absFilename();
158 if (FileName(name).isAbsolute() || !FileName(fname + ext).isReadableFile())
164 return to_utf8(makeRelPath(from_utf8(fname),
165 from_utf8(buffer.masterBuffer()->filePath())));
171 int InsetBibtex::latex(Buffer const & buffer, odocstream & os,
172 OutputParams const & runparams) const
174 // the sequence of the commands:
175 // 1. \bibliographystyle{style}
176 // 2. \addcontentsline{...} - if option bibtotoc set
177 // 3. \bibliography{database}
178 // and with bibtopic:
179 // 1. \bibliographystyle{style}
180 // 2. \begin{btSect}{database}
181 // 3. \btPrint{Cited|NotCited|All}
185 // If we are processing the LaTeX file in a temp directory then
186 // copy the .bib databases to this temp directory, mangling their
187 // names in the process. Store this mangled name in the list of
189 // (We need to do all this because BibTeX *really*, *really*
190 // can't handle "files with spaces" and Windows users tend to
191 // use such filenames.)
192 // Otherwise, store the (maybe absolute) path to the original,
193 // unmangled database name.
194 EmbeddedFileList const bibs = getFiles(buffer);
195 EmbeddedFileList::const_iterator it = bibs.begin();
196 EmbeddedFileList::const_iterator it_end = bibs.end();
197 odocstringstream dbs;
198 for (; it != it_end; ++it) {
199 string utf8input = removeExtension(it->availableFile().absFilename());
201 normalizeName(buffer, runparams, utf8input, ".bib");
202 FileName const try_in_file(makeAbsPath(database + ".bib", buffer.filePath()));
203 bool const not_from_texmf = try_in_file.isReadableFile();
205 if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
208 // mangledFilename() needs the extension
209 DocFileName const in_file = DocFileName(try_in_file);
210 database = removeExtension(in_file.mangledFilename());
211 FileName const out_file = makeAbsPath(database + ".bib",
212 buffer.masterBuffer()->temppath());
214 bool const success = in_file.copyTo(out_file);
216 lyxerr << "Failed to copy '" << in_file
217 << "' to '" << out_file << "'"
220 } else if (!runparams.inComment && runparams.nice && not_from_texmf &&
221 !isValidLaTeXFilename(database)) {
222 frontend::Alert::warning(_("Invalid filename"),
223 _("The following filename is likely to cause trouble "
224 "when running the exported file through LaTeX: ") +
225 from_utf8(database));
228 if (it != bibs.begin())
231 dbs << from_utf8(latex_path(database));
233 docstring const db_out = dbs.str();
235 // Post this warning only once.
236 static bool warned_about_spaces = false;
237 if (!warned_about_spaces &&
238 runparams.nice && db_out.find(' ') != docstring::npos) {
239 warned_about_spaces = true;
241 Alert::warning(_("Export Warning!"),
242 _("There are spaces in the paths to your BibTeX databases.\n"
243 "BibTeX will be unable to find them."));
247 string style = to_utf8(getParam("options")); // maybe empty! and with bibtotoc
249 if (prefixIs(style, "bibtotoc")) {
250 bibtotoc = "bibtotoc";
251 if (contains(style, ','))
252 style = split(style, bibtotoc, ',');
258 if (!style.empty()) {
259 string base = normalizeName(buffer, runparams, style, ".bst");
260 FileName const try_in_file(makeAbsPath(base + ".bst", buffer.filePath()));
261 bool const not_from_texmf = try_in_file.isReadableFile();
262 // If this style does not come from texmf and we are not
263 // exporting to .tex copy it to the tmp directory.
264 // This prevents problems with spaces and 8bit charcaters
266 if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
268 // use new style name
269 DocFileName const in_file = DocFileName(try_in_file);
270 base = removeExtension(in_file.mangledFilename());
271 FileName const out_file(makeAbsPath(base + ".bst",
272 buffer.masterBuffer()->temppath()));
273 bool const success = in_file.copyTo(out_file);
275 lyxerr << "Failed to copy '" << in_file
276 << "' to '" << out_file << "'"
281 os << "\\bibliographystyle{"
282 << from_utf8(latex_path(normalizeName(buffer, runparams, base, ".bst")))
287 // Post this warning only once.
288 static bool warned_about_bst_spaces = false;
289 if (!warned_about_bst_spaces && runparams.nice && contains(style, ' ')) {
290 warned_about_bst_spaces = true;
291 Alert::warning(_("Export Warning!"),
292 _("There are spaces in the path to your BibTeX style file.\n"
293 "BibTeX will be unable to find it."));
296 if (!db_out.empty() && buffer.params().use_bibtopic){
297 os << "\\begin{btSect}{" << db_out << "}\n";
298 docstring btprint = getParam("btprint");
301 btprint = from_ascii("btPrintCited");
302 os << "\\" << btprint << "\n"
303 << "\\end{btSect}\n";
308 if (!bibtotoc.empty() && !buffer.params().use_bibtopic) {
309 // maybe a problem when a textclass has no "art" as
310 // part of its name, because it's than book.
311 // For the "official" lyx-layouts it's no problem to support
313 if (!contains(buffer.params().getTextClass().name(),
315 if (buffer.params().sides == OneSide) {
320 os << "\\cleardoublepage";
324 os << "\\addcontentsline{toc}{chapter}{\\bibname}";
328 os << "\\addcontentsline{toc}{section}{\\refname}";
332 if (!db_out.empty() && !buffer.params().use_bibtopic){
333 docstring btprint = getParam("btprint");
334 if (btprint == "btPrintAll") {
335 os << "\\nocite{*}\n";
338 os << "\\bibliography{" << db_out << "}\n";
346 EmbeddedFileList const InsetBibtex::getFiles(Buffer const & buffer) const
348 FileName path(buffer.filePath());
351 EmbeddedFileList vec;
356 string bibfiles = to_utf8(getParam("bibfiles"));
357 string embedStatus = to_utf8(getParam("embed"));
358 bibfiles = split(bibfiles, tmp, ',');
359 embedStatus = split(embedStatus, emb, ',');
360 while (!tmp.empty()) {
362 EmbeddedFile file(changeExtension(tmp, "bib"), buffer.filePath());
363 // If the file structure is correct, this should not fail.
365 file.enable(buffer.embedded(), &buffer);
368 // this includes the cases when the embed parameter is empty
369 FileName const file = findtexfile(changeExtension(tmp, "bib"), "bib");
371 // If we didn't find a matching file name just fail silently
373 EmbeddedFile efile = EmbeddedFile(file.absFilename(), buffer.filePath());
374 efile.setEmbed(false);
375 efile.enable(buffer.embedded(), &buffer);
376 vec.push_back(efile);
380 // Get next file name
381 bibfiles = split(bibfiles, tmp, ',');
382 embedStatus = split(embedStatus, emb, ',');
390 // methods for parsing bibtex files
392 typedef map<docstring, docstring> VarMap;
394 /// remove whitespace characters, optionally a single comma,
395 /// and further whitespace characters from the stream.
396 /// @return true if a comma was found, false otherwise
398 bool removeWSAndComma(idocfstream & ifs) {
407 } while (ifs && isSpace(ch));
420 } while (ifs && isSpace(ch));
435 /// remove whitespace characters, read characer sequence
436 /// not containing whitespace characters or characters in
437 /// delimChars, and remove further whitespace characters.
439 /// @return true if a string of length > 0 could be read.
441 bool readTypeOrKey(docstring & val, idocfstream & ifs,
442 docstring const & delimChars, docstring const &illegalChars,
455 } while (ifs && isSpace(ch));
461 bool legalChar = true;
462 while (ifs && !isSpace(ch) &&
463 delimChars.find(ch) == docstring::npos &&
464 (legalChar = (illegalChars.find(ch) == docstring::npos))
467 if (chCase == makeLowerCase)
468 val += lowercase(ch);
480 while (ifs && isSpace(ch)) {
488 return val.length() > 0;
491 /// read subsequent bibtex values that are delimited with a #-character.
492 /// Concatenate all parts and replace names with the associated string in
493 /// the variable strings.
494 /// @return true if reading was successfull (all single parts were delimited
496 bool readValue(docstring & val, idocfstream & ifs, const VarMap & strings) {
509 } while (ifs && isSpace(ch));
514 // check for field type
517 // read integer value
521 } while (ifs && isDigit(ch));
526 } else if (ch == '"' || ch == '{') {
528 char_type delim = ch == '"' ? '"': '}';
533 } while (ifs && isSpace(ch));
538 //We now have the first non-whitespace character
539 //We'll collapse adjacent whitespace.
540 bool lastWasWhiteSpace = false;
542 // inside this delimited text braces must match.
543 // Thus we can have a closing delimiter only
544 // when nestLevel == 0
547 while (ifs && (nestLevel > 0 || ch != delim)) {
549 lastWasWhiteSpace = true;
553 //We output the space only after we stop getting
554 //whitespace so as not to output any whitespace
555 //at the end of the value.
556 if (lastWasWhiteSpace) {
557 lastWasWhiteSpace = false;
563 // update nesting level
570 if (nestLevel < 0) return false;
587 // reading a string name
590 while (ifs && !isSpace(ch) && ch != '#' && ch != ',' && ch != '}' && ch != ')') {
591 strName += lowercase(ch);
598 // replace the string with its assigned value or
599 // discard it if it's not assigned
600 if (strName.length()) {
601 VarMap::const_iterator pos = strings.find(strName);
602 if (pos != strings.end()) {
609 while (ifs && isSpace(ch)) {
616 // continue reading next value on concatenate with '#'
626 // This method returns a comma separated list of Bibtex entries
627 void InsetBibtex::fillWithBibKeys(Buffer const & buffer,
628 BiblioInfo & keylist, InsetIterator const & /*di*/) const
630 EmbeddedFileList const files = getFiles(buffer);
631 for (vector<EmbeddedFile>::const_iterator it = files.begin();
632 it != files.end(); ++ it) {
633 // This bibtex parser is a first step to parse bibtex files
636 // - it reads the whole bibtex entry and does a syntax check
637 // (matching delimiters, missing commas,...
638 // - it recovers from errors starting with the next @-character
639 // - it reads @string definitions and replaces them in the
641 // - it accepts more characters in keys or value names than
644 // Officially bibtex does only support ASCII, but in practice
645 // you can use the encoding of the main document as long as
646 // some elements like keys and names are pure ASCII. Therefore
647 // we convert the file from the buffer encoding.
648 // We don't restrict keys to ASCII in LyX, since our own
649 // InsetBibitem can generate non-ASCII keys, and nonstandard
650 // 8bit clean bibtex forks exist.
652 idocfstream ifs(it->availableFile().toFilesystemEncoding().c_str(),
654 buffer.params().encoding().iconvName());
670 if (!readTypeOrKey(entryType, ifs, from_ascii("{("),
671 docstring(), makeLowerCase) || !ifs)
674 if (entryType == from_ascii("comment")) {
676 ifs.ignore(numeric_limits<int>::max(), '\n');
684 if ((ch != '(') && (ch != '{')) {
685 // invalid entry delimiter
691 if (entryType == from_ascii("string")) {
693 // read string and add it to the strings map
694 // (or replace it's old value)
698 if (!readTypeOrKey(name, ifs, from_ascii("="),
699 from_ascii("#{}(),"), makeLowerCase) || !ifs)
702 // next char must be an equal sign
704 if (!ifs || ch != '=')
707 if (!readValue(value, ifs, strings))
710 strings[name] = value;
712 } else if (entryType == from_ascii("preamble")) {
714 // preamble definitions are discarded.
715 // can they be of any use in lyx?
718 if (!readValue(value, ifs, strings))
723 // Citation entry. Try to read the key.
726 if (!readTypeOrKey(key, ifs, from_ascii(","),
727 from_ascii("}"), keepCase) || !ifs)
730 /////////////////////////////////////////////
731 // now we have a key, so we will add an entry
732 // (even if it's empty, as bibtex does)
734 // we now read the field = value pairs.
735 // all items must be separated by a comma. If
736 // it is missing the scanning of this entry is
737 // stopped and the next is searched.
741 docstring commaNewline;
743 BibTeXInfo keyvalmap(key, entryType);
745 bool readNext = removeWSAndComma(ifs);
747 while (ifs && readNext) {
750 if (!readTypeOrKey(name, ifs, from_ascii("="),
751 from_ascii("{}(),"), makeLowerCase) || !ifs)
754 // next char must be an equal sign
764 if (!readValue(value, ifs, strings))
767 keyvalmap[name] = value;
768 data += "\n\n" + value;
769 keylist.addFieldName(name);
770 readNext = removeWSAndComma(ifs);
774 keylist.addEntryType(entryType);
775 keyvalmap.setAllData(data);
776 keylist[key] = keyvalmap;
779 } //< for loop over files
784 bool InsetBibtex::addDatabase(string const & db)
787 string bibfiles(to_utf8(getParam("bibfiles")));
788 if (tokenPos(bibfiles, ',', db) == -1) {
789 if (!bibfiles.empty())
791 setParam("bibfiles", from_utf8(bibfiles + db));
798 bool InsetBibtex::delDatabase(string const & db)
801 string bibfiles(to_utf8(getParam("bibfiles")));
802 if (contains(bibfiles, db)) {
803 int const n = tokenPos(bibfiles, ',', db);
806 // this is not the first database
807 string tmp = ',' + bd;
808 setParam("bibfiles", from_utf8(subst(bibfiles, tmp, string())));
810 // this is the first (or only) database
811 setParam("bibfiles", from_utf8(split(bibfiles, bd, ',')));
819 void InsetBibtex::validate(LaTeXFeatures & features) const
821 if (features.bufferParams().use_bibtopic)
822 features.require("bibtopic");
826 void InsetBibtex::registerEmbeddedFiles(Buffer const & buffer, EmbeddedFileList & files) const
828 EmbeddedFileList const dbs = getFiles(buffer);
829 for (vector<EmbeddedFile>::const_iterator it = dbs.begin();
830 it != dbs.end(); ++ it)
831 files.registerFile(*it, this, buffer);
835 void InsetBibtex::updateEmbeddedFile(Buffer const & buf, EmbeddedFile const & file)
837 // look for the item and update status
842 EmbeddedFileList dbs = getFiles(buf);
843 for (EmbeddedFileList::iterator it = dbs.begin();
844 it != dbs.end(); ++ it) {
846 if (it->absFilename() == file.absFilename())
847 it->setEmbed(file.embedded());
848 // write parameter string
854 bibfiles += from_utf8(it->outputFilename(buf.filePath()));
856 embed += from_utf8(it->inzipName());
858 setParam("bibfiles", bibfiles);
859 setParam("embed", embed);