private StringBuilder parseBracketedTextExactly() throws IOException {
StringBuilder value = new StringBuilder();
consume('{');
int brackets = 0;
char character;
char lastCharacter = '\0';
while (true) {
character = (char) read();
boolean isClosingBracket = (character == '}') && (lastCharacter != '\\');
if (isClosingBracket && (brackets == 0)) {
return value;
} else if (isEOFCharacter(character)) {
throw new IOException("Error in line " + line + ": EOF in mid-string");
} else if ((character == '{') && (!isEscapeSymbol(lastCharacter))) {
brackets++;
} else if (isClosingBracket) {
brackets--;
}
value.append(character);
lastCharacter = character;
}
}
private void parseField(BibEntry entry) throws IOException {
String key = parseTextToken().toLowerCase();
skipWhitespace();
consume('=');
String content = parseFieldContent(key);
if (!content.isEmpty()) {
if (entry.hasField(key)) {
// The following hack enables the parser to deal with multiple
// author or
// editor lines, stringing them together instead of getting just
// one of them.
// Multiple author or editor lines are not allowed by the bibtex
// format, but
// at least one online database exports bibtex like that, making
// it inconvenient
// for users if JabRef didn't accept it.
if (InternalBibtexFields.getFieldExtras(key).contains(FieldProperties.PERSON_NAMES)) {
entry.setField(key, entry.getFieldOptional(key).get() + " and " + content);
} else if (FieldName.KEYWORDS.equals(key)) {
// multiple keywords fields should be combined to one
entry.addKeyword(content, Globals.prefs.get(JabRefPreferences.KEYWORD_SEPARATOR));
}
} else {
entry.setField(key, content);
}
}
}
private StringBuffer parseBracketedText() throws IOException {
StringBuffer value = new StringBuffer();
consume('{', '(');
int brackets = 0;
while (!((isClosingBracketNext()) && (brackets == 0))) {
int character = read();
if (isEOFCharacter(character)) {
throw new IOException("Error in line " + line + ": EOF in mid-string");
} else if ((character == '{') || (character == '(')) {
brackets++;
} else if ((character == '}') || (character == ')')) {
brackets--;
}
// If we encounter whitespace of any kind, read it as a
// simple space, and ignore any others that follow immediately.
/*
* if (j == '\n') { if (peek() == '\n') value.append('\n'); } else
*/
if (Character.isWhitespace((char) character)) {
String whitespacesReduced = skipAndRecordWhitespace(character);
if (!(whitespacesReduced.isEmpty()) && !"\n\t".equals(whitespacesReduced)) { // &&
whitespacesReduced = whitespacesReduced.replace("\t", ""); // Remove tabulators.
value.append(whitespacesReduced);
} else {
value.append(' ');
}
} else {
value.append((char) character);
}
}
consume('}', ')');
return value;
}
private BibtexString parseString() throws IOException {
skipWhitespace();
consume('{', '(');
skipWhitespace();
LOGGER.debug("Parsing string name");
String name = parseTextToken();
LOGGER.debug("Parsed string name");
skipWhitespace();
LOGGER.debug("Now the contents");
consume('=');
String content = parseFieldContent(name);
LOGGER.debug("Now I'm going to consume a }");
consume('}', ')');
// Consume new line which signals end of entry
skipOneNewline();
LOGGER.debug("Finished string parsing.");
String id = IdGenerator.next();
return new BibtexString(id, name, content);
}
private BibEntry parseEntry(String entryType) throws IOException {
String id = IdGenerator.next();
BibEntry result = new BibEntry(id, entryType);
skipWhitespace();
consume('{', '(');
int character = peek();
if ((character != '\n') && (character != '\r')) {
skipWhitespace();
}
String key = parseKey();
result.setCiteKey(key);
skipWhitespace();
while (true) {
character = peek();
if ((character == '}') || (character == ')')) {
break;
}
if (character == ',') {
consume(',');
}
skipWhitespace();
character = peek();
if ((character == '}') || (character == ')')) {
break;
}
parseField(result);
}
consume('}', ')');
// Consume new line which signals end of entry
skipOneNewline();
return result;
}
private StringBuilder parseQuotedFieldExactly() throws IOException {
StringBuilder value = new StringBuilder();
consume('"');
int brackets = 0;
while (!((peek() == '"') && (brackets == 0))) {
int j = read();
if (isEOFCharacter(j)) {
throw new IOException("Error in line " + line + ": EOF in mid-string");
} else if (j == '{') {
brackets++;
} else if (j == '}') {
brackets--;
}
value.append((char) j);
}
consume('"');
return value;
}
private String parseFieldContent(String key) throws IOException {
skipWhitespace();
StringBuilder value = new StringBuilder();
int character;
while (((character = peek()) != ',') && (character != '}') && (character != ')')) {
if (eof) {
throw new IOException("Error in line " + line + ": EOF in mid-string");
}
if (character == '"') {
StringBuilder text = parseQuotedFieldExactly();
value.append(fieldContentParser.format(text, key));
} else if (character == '{') {
// Value is a string enclosed in brackets. There can be pairs
// of brackets inside of a field, so we need to count the
// brackets to know when the string is finished.
StringBuilder text = parseBracketedTextExactly();
value.append(fieldContentParser.format(text, key));
} else if (Character.isDigit((char) character)) { // value is a number
String number = parseTextToken();
value.append(number);
} else if (character == '#') {
consume('#');
} else {
String textToken = parseTextToken();
if (textToken.isEmpty()) {
throw new IOException(
"Error in line "
+ line
+ " or above: "
+ "Empty text token.\nThis could be caused "
+ "by a missing comma between two fields.");
}
value.append('#').append(textToken).append('#');
}
skipWhitespace();
}
return value.toString();
}
