public void checkEntryTypes(ParserResult _pr) {
for (BibtexEntry be : _db.getEntries()) {
if (be.getType() instanceof UnknownEntryType) {
// Look up the unknown type name in our map of parsed types:
Object o = entryTypes.get(be.getType().getName().toLowerCase());
if (o != null) {
BibtexEntryType type = (BibtexEntryType) o;
be.setType(type);
} else {
// System.out.println("Unknown entry type:
// "+be.getType().getName());
_pr.addWarning(
Globals.lang("unknown entry type")
+ ": "
+ be.getType().getName()
+ ". "
+ Globals.lang("Type set to 'other'")
+ ".");
be.setType(BibtexEntryType.OTHER);
}
}
}
}
/**
* Tries to restore the key
*
* @return rest of key on success, otherwise empty string
* @throws IOException on Reader-Error
*/
private String fixKey() throws IOException {
StringBuilder key = new StringBuilder();
int lookahead_used = 0;
char currentChar;
// Find a char which ends key (','&&'\n') or entryfield ('='):
do {
currentChar = (char) read();
key.append(currentChar);
lookahead_used++;
} while ((currentChar != ',' && currentChar != '\n' && currentChar != '=')
&& (lookahead_used < LOOKAHEAD));
// Consumed a char too much, back into reader and remove from key:
unread(currentChar);
key.deleteCharAt(key.length() - 1);
// Restore if possible:
switch (currentChar) {
case '=':
// Get entryfieldname, push it back and take rest as key
key = key.reverse();
boolean matchedAlpha = false;
for (int i = 0; i < key.length(); i++) {
currentChar = key.charAt(i);
/// Skip spaces:
if (!matchedAlpha && currentChar == ' ') {
continue;
}
matchedAlpha = true;
// Begin of entryfieldname (e.g. author) -> push back:
unread(currentChar);
if (currentChar == ' ' || currentChar == '\n') {
/*
* found whitespaces, entryfieldname completed -> key in
* keybuffer, skip whitespaces
*/
StringBuilder newKey = new StringBuilder();
for (int j = i; j < key.length(); j++) {
currentChar = key.charAt(j);
if (!Character.isWhitespace(currentChar)) {
newKey.append(currentChar);
}
}
// Finished, now reverse newKey and remove whitespaces:
_pr.addWarning(
Globals.lang("Line %0: Found corrupted BibTeX-key.", String.valueOf(line)));
key = newKey.reverse();
}
}
break;
case ',':
_pr.addWarning(
Globals.lang(
"Line %0: Found corrupted BibTeX-key (contains whitespaces).",
String.valueOf(line)));
case '\n':
_pr.addWarning(
Globals.lang(
"Line %0: Found corrupted BibTeX-key (comma missing).", String.valueOf(line)));
break;
default:
// No more lookahead, give up:
unreadBuffer(key);
return "";
}
return removeWhitespaces(key).toString();
}
/**
* Will parse the BibTex-Data found when reading from reader.
*
* <p>The reader will be consumed.
*
* <p>Multiple calls to parse() return the same results
*
* @return ParserResult
* @throws IOException
*/
public ParserResult parse() throws IOException {
// If we already parsed this, just return it.
if (_pr != null) return _pr;
_db = new BibtexDatabase(); // Bibtex related contents.
_meta = new HashMap<String, String>(); // Metadata in comments for Bibkeeper.
entryTypes = new HashMap<String, BibtexEntryType>(); // To store custem entry types parsed.
_pr = new ParserResult(_db, _meta, entryTypes);
// First see if we can find the version number of the JabRef version that
// wrote the file:
String versionNum = readJabRefVersionNumber();
if (versionNum != null) {
_pr.setJabrefVersion(versionNum);
setMajorMinorVersions();
} else {
// No version number found. However, we have only
}
skipWhitespace();
try {
while (!_eof) {
boolean found = consumeUncritically('@');
if (!found) break;
skipWhitespace();
String entryType = parseTextToken();
BibtexEntryType tp = BibtexEntryType.getType(entryType);
boolean isEntry = (tp != null);
// Util.pr(tp.getName());
if (!isEntry) {
// The entry type name was not recognized. This can mean
// that it is a string, preamble, or comment. If so,
// parse and set accordingly. If not, assume it is an entry
// with an unknown type.
if (entryType.toLowerCase().equals("preamble")) {
_db.setPreamble(parsePreamble());
} else if (entryType.toLowerCase().equals("string")) {
BibtexString bs = parseString();
try {
_db.addString(bs);
} catch (KeyCollisionException ex) {
_pr.addWarning(Globals.lang("Duplicate string name") + ": " + bs.getName());
// ex.printStackTrace();
}
} else if (entryType.toLowerCase().equals("comment")) {
StringBuffer commentBuf = parseBracketedTextExactly();
/**
* Metadata are used to store Bibkeeper-specific information in .bib files.
*
* <p>Metadata are stored in bibtex files in the format
*
* @comment{jabref-meta: type:data0;data1;data2;...}
* <p>Each comment that starts with the META_FLAG is stored in the meta HashMap,
* with type as key. Unluckily, the old META_FLAG bibkeeper-meta: was used in JabRef
* 1.0 and 1.1, so we need to support it as well. At least for a while. We'll always
* save with the new one.
*/
String comment = commentBuf.toString().replaceAll("[\\x0d\\x0a]", "");
if (comment
.substring(0, Math.min(comment.length(), GUIGlobals.META_FLAG.length()))
.equals(GUIGlobals.META_FLAG)
|| comment
.substring(0, Math.min(comment.length(), GUIGlobals.META_FLAG_OLD.length()))
.equals(GUIGlobals.META_FLAG_OLD)) {
String rest;
if (comment.substring(0, GUIGlobals.META_FLAG.length()).equals(GUIGlobals.META_FLAG))
rest = comment.substring(GUIGlobals.META_FLAG.length());
else rest = comment.substring(GUIGlobals.META_FLAG_OLD.length());
int pos = rest.indexOf(':');
if (pos > 0) _meta.put(rest.substring(0, pos), rest.substring(pos + 1));
// We remove all line breaks in the metadata - these
// will have been inserted
// to prevent too long lines when the file was
// saved, and are not part of the data.
}
/**
* A custom entry type can also be stored in a
*
* @comment:
*/
if (comment
.substring(0, Math.min(comment.length(), GUIGlobals.ENTRYTYPE_FLAG.length()))
.equals(GUIGlobals.ENTRYTYPE_FLAG)) {
CustomEntryType typ = CustomEntryType.parseEntryType(comment);
entryTypes.put(typ.getName().toLowerCase(), typ);
}
} else {
// The entry type was not recognized. This may mean that
// it is a custom entry type whose definition will
// appear
// at the bottom of the file. So we use an
// UnknownEntryType
// to remember the type name by.
tp = new UnknownEntryType(entryType.toLowerCase());
// System.out.println("unknown type: "+entryType);
isEntry = true;
}
}
if (isEntry) // True if not comment, preamble or string.
{
/**
* Morten Alver 13 Aug 2006: Trying to make the parser more robust. If an exception is
* thrown when parsing an entry, drop the entry and try to resume parsing. Add a warning
* for the user.
*
* <p>An alternative solution is to try rescuing the entry for which parsing failed, by
* returning the entry with the exception and adding it before parsing is continued.
*/
try {
BibtexEntry be = parseEntry(tp);
boolean duplicateKey = _db.insertEntry(be);
if (duplicateKey) // JZTODO lyrics
_pr.addWarning(
Globals.lang("duplicate BibTeX key")
+ ": "
+ be.getCiteKey()
+ " ("
+ Globals.lang("grouping may not work for this entry")
+ ")");
else if (be.getCiteKey() == null || be.getCiteKey().equals("")) {
_pr.addWarning(
Globals.lang("empty BibTeX key")
+ ": "
+ be.getAuthorTitleYear(40)
+ " ("
+ Globals.lang("grouping may not work for this entry")
+ ")");
}
} catch (IOException ex) {
ex.printStackTrace();
_pr.addWarning(
Globals.lang("Error occured when parsing entry")
+ ": '"
+ ex.getMessage()
+ "'. "
+ Globals.lang("Skipped entry."));
}
}
skipWhitespace();
}
// Before returning the database, update entries with unknown type
// based on parsed type definitions, if possible.
checkEntryTypes(_pr);
return _pr;
} catch (KeyCollisionException kce) {
// kce.printStackTrace();
throw new IOException("Duplicate ID in bibtex file: " + kce.toString());
}
}