private static Token token(
String term, int posInc, int posLength, int startOffset, int endOffset) {
final Token t = new Token(term, startOffset, endOffset);
t.setPositionIncrement(posInc);
t.setPositionLength(posLength);
return t;
}
@Test
public void testCountPositions() throws IOException {
// We're looking to make sure that we:
Token t1 = new Token(); // Don't count tokens without an increment
t1.setPositionIncrement(0);
Token t2 = new Token();
t2.setPositionIncrement(1); // Count normal tokens with one increment
Token t3 = new Token();
t2.setPositionIncrement(2); // Count funny tokens with more than one increment
int finalTokenIncrement =
4; // Count the final token increment on the rare token streams that have them
Token[] tokens = new Token[] {t1, t2, t3};
Collections.shuffle(Arrays.asList(tokens), getRandom());
TokenStream tokenStream = new CannedTokenStream(finalTokenIncrement, 0, tokens);
assertThat(TokenCountFieldMapper.countPositions(tokenStream), equalTo(7));
}
private void fillTokens() throws IOException {
final StringBuilder sb = new StringBuilder();
final char[] buffer = new char[256];
while (true) {
final int count = input.read(buffer);
if (count == -1) {
break;
}
sb.append(buffer, 0, count);
// System.out.println("got count=" + count);
}
// System.out.println("fillTokens: " + sb);
inputLength = sb.length();
final String[] parts = sb.toString().split(" ");
tokens = new ArrayList<Token>();
int pos = 0;
int maxPos = -1;
int offset = 0;
// System.out.println("again");
for (String part : parts) {
final String[] overlapped = part.split("/");
boolean firstAtPos = true;
int minPosLength = Integer.MAX_VALUE;
for (String part2 : overlapped) {
final int colonIndex = part2.indexOf(':');
final String token;
final int posLength;
if (colonIndex != -1) {
token = part2.substring(0, colonIndex);
posLength = Integer.parseInt(part2.substring(1 + colonIndex));
} else {
token = part2;
posLength = 1;
}
maxPos = Math.max(maxPos, pos + posLength);
minPosLength = Math.min(minPosLength, posLength);
final Token t = new Token(token, offset, offset + 2 * posLength - 1);
t.setPositionLength(posLength);
t.setPositionIncrement(firstAtPos ? 1 : 0);
firstAtPos = false;
// System.out.println(" add token=" + t + " startOff=" + t.startOffset() + " endOff=" +
// t.endOffset());
tokens.add(t);
}
pos += minPosLength;
offset = 2 * pos;
}
assert maxPos <= pos : "input string mal-formed: posLength>1 tokens hang over the end";
}
/**
* For languages with canonical form
*
* @return canonical token (or null if none)
*/
public Token canonizeToken(Token t) {
if (!hasCanonicalFilter) return null;
if (lang.equals("sr")) {
String nt = new SerbianFilter(null).convert(t.termText());
if (!t.equals(nt)) {
Token tt = new Token(nt, t.startOffset(), t.endOffset());
tt.setPositionIncrement(0);
tt.setType("alias");
return tt;
}
}
return null;
}
/**
* Merge two lists of tokens, producing a single list with manipulated positionIncrements so that
* the tokens end up at the same position.
*
* <p>Example: [a b] merged with [c d] produces [a/b c/d] ('/' denotes tokens in the same
* position) Example: [a,5 b,2] merged with [c d,4 e,4] produces [c a,5/d b,2 e,2] (a,n means a
* has posInc=n)
*/
public static List<Token> mergeTokens(List<Token> lst1, List<Token> lst2) {
ArrayList<Token> result = new ArrayList<Token>();
if (lst1 == null || lst2 == null) {
if (lst2 != null) result.addAll(lst2);
if (lst1 != null) result.addAll(lst1);
return result;
}
int pos = 0;
Iterator<Token> iter1 = lst1.iterator();
Iterator<Token> iter2 = lst2.iterator();
Token tok1 = iter1.hasNext() ? iter1.next() : null;
Token tok2 = iter2.hasNext() ? iter2.next() : null;
int pos1 = tok1 != null ? tok1.getPositionIncrement() : 0;
int pos2 = tok2 != null ? tok2.getPositionIncrement() : 0;
while (tok1 != null || tok2 != null) {
while (tok1 != null && (pos1 <= pos2 || tok2 == null)) {
Token tok = new Token(tok1.startOffset(), tok1.endOffset(), tok1.type());
tok.copyBuffer(tok1.buffer(), 0, tok1.length());
tok.setPositionIncrement(pos1 - pos);
result.add(tok);
pos = pos1;
tok1 = iter1.hasNext() ? iter1.next() : null;
pos1 += tok1 != null ? tok1.getPositionIncrement() : 0;
}
while (tok2 != null && (pos2 <= pos1 || tok1 == null)) {
Token tok = new Token(tok2.startOffset(), tok2.endOffset(), tok2.type());
tok.copyBuffer(tok2.buffer(), 0, tok2.length());
tok.setPositionIncrement(pos2 - pos);
result.add(tok);
pos = pos2;
tok2 = iter2.hasNext() ? iter2.next() : null;
pos2 += tok2 != null ? tok2.getPositionIncrement() : 0;
}
}
return result;
}
private void splitIntoTokens() {
String term = termAtt.term();
String[] termParts = splitTerm(term);
if (termParts.length > 1) {
int termPos = offsetAtt.startOffset();
for (int i = 0; i < termParts.length; i++) {
String termPart = termParts[i];
int termPartPos = termPos + term.indexOf(termPart);
int termPartEndPos = termPartPos + termPart.length();
Token newToken = new Token(termPart, termPartPos, termPartEndPos);
newToken.setPositionIncrement(0); // in the same position
tokens.add(newToken);
}
}
}
private static Token token(String term, int posInc, int posLength) {
final Token t = new Token(term, 0, 0);
t.setPositionIncrement(posInc);
t.setPositionLength(posLength);
return t;
}