1 <!doctype html public "-//w3c//dtd html 4.0 transitional//en">
3 Licensed to the Apache Software Foundation (ASF) under one or more
4 contributor license agreements. See the NOTICE file distributed with
5 this work for additional information regarding copyright ownership.
6 The ASF licenses this file to You under the Apache License, Version 2.0
7 (the "License"); you may not use this file except in compliance with
8 the License. You may obtain a copy of the License at
10 http://www.apache.org/licenses/LICENSE-2.0
12 Unless required by applicable law or agreed to in writing, software
13 distributed under the License is distributed on an "AS IS" BASIS,
14 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15 See the License for the specific language governing permissions and
16 limitations under the License.
20 <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
23 <p>API and code to convert text into indexable/searchable tokens. Covers {@link org.apache.lucene.analysis.Analyzer} and related classes.</p>
24 <h2>Parsing? Tokenization? Analysis!</h2>
26 Lucene, indexing and search library, accepts only plain text input.
30 Applications that build their search capabilities upon Lucene may support documents in various formats – HTML, XML, PDF, Word – just to name a few.
31 Lucene does not care about the <i>Parsing</i> of these and other document formats, and it is the responsibility of the
32 application using Lucene to use an appropriate <i>Parser</i> to convert the original format into plain text before passing that plain text to Lucene.
36 Plain text passed to Lucene for indexing goes through a process generally called tokenization. Tokenization is the process
37 of breaking input text into small indexing elements – tokens.
38 The way input text is broken into tokens heavily influences how people will then be able to search for that text.
39 For instance, sentences beginnings and endings can be identified to provide for more accurate phrase
40 and proximity searches (though sentence identification is not provided by Lucene).
42 In some cases simply breaking the input text into tokens is not enough – a deeper <i>Analysis</i> may be needed.
43 There are many post tokenization steps that can be done, including (but not limited to):
45 <li><a href="http://en.wikipedia.org/wiki/Stemming">Stemming</a> –
46 Replacing of words by their stems.
47 For instance with English stemming "bikes" is replaced by "bike";
48 now query "bike" can find both documents containing "bike" and those containing "bikes".
50 <li><a href="http://en.wikipedia.org/wiki/Stop_words">Stop Words Filtering</a> –
51 Common words like "the", "and" and "a" rarely add any value to a search.
52 Removing them shrinks the index size and increases performance.
53 It may also reduce some "noise" and actually improve search quality.
55 <li><a href="http://en.wikipedia.org/wiki/Text_normalization">Text Normalization</a> –
56 Stripping accents and other character markings can make for better searching.
58 <li><a href="http://en.wikipedia.org/wiki/Synonym">Synonym Expansion</a> –
59 Adding in synonyms at the same token position as the current word can mean better
60 matching when users search with words in the synonym set.
64 <h2>Core Analysis</h2>
66 The analysis package provides the mechanism to convert Strings and Readers into tokens that can be indexed by Lucene. There
67 are three main classes in the package from which all analysis processes are derived. These are:
69 <li>{@link org.apache.lucene.analysis.Analyzer} – An Analyzer is responsible for building a {@link org.apache.lucene.analysis.TokenStream} which can be consumed
70 by the indexing and searching processes. See below for more information on implementing your own Analyzer.</li>
71 <li>{@link org.apache.lucene.analysis.Tokenizer} – A Tokenizer is a {@link org.apache.lucene.analysis.TokenStream} and is responsible for breaking
72 up incoming text into tokens. In most cases, an Analyzer will use a Tokenizer as the first step in
73 the analysis process.</li>
74 <li>{@link org.apache.lucene.analysis.TokenFilter} – A TokenFilter is also a {@link org.apache.lucene.analysis.TokenStream} and is responsible
75 for modifying tokens that have been created by the Tokenizer. Common modifications performed by a
76 TokenFilter are: deletion, stemming, synonym injection, and down casing. Not all Analyzers require TokenFilters</li>
78 <b>Lucene 2.9 introduces a new TokenStream API. Please see the section "New TokenStream API" below for more details.</b>
80 <h2>Hints, Tips and Traps</h2>
82 The synergy between {@link org.apache.lucene.analysis.Analyzer} and {@link org.apache.lucene.analysis.Tokenizer}
83 is sometimes confusing. To ease on this confusion, some clarifications:
85 <li>The {@link org.apache.lucene.analysis.Analyzer} is responsible for the entire task of
86 <u>creating</u> tokens out of the input text, while the {@link org.apache.lucene.analysis.Tokenizer}
87 is only responsible for <u>breaking</u> the input text into tokens. Very likely, tokens created
88 by the {@link org.apache.lucene.analysis.Tokenizer} would be modified or even omitted
89 by the {@link org.apache.lucene.analysis.Analyzer} (via one or more
90 {@link org.apache.lucene.analysis.TokenFilter}s) before being returned.
92 <li>{@link org.apache.lucene.analysis.Tokenizer} is a {@link org.apache.lucene.analysis.TokenStream},
93 but {@link org.apache.lucene.analysis.Analyzer} is not.
95 <li>{@link org.apache.lucene.analysis.Analyzer} is "field aware", but
96 {@link org.apache.lucene.analysis.Tokenizer} is not.
101 Lucene Java provides a number of analysis capabilities, the most commonly used one being the {@link
102 org.apache.lucene.analysis.standard.StandardAnalyzer}. Many applications will have a long and industrious life with nothing more
103 than the StandardAnalyzer. However, there are a few other classes/packages that are worth mentioning:
105 <li>{@link org.apache.lucene.analysis.PerFieldAnalyzerWrapper} – Most Analyzers perform the same operation on all
106 {@link org.apache.lucene.document.Field}s. The PerFieldAnalyzerWrapper can be used to associate a different Analyzer with different
107 {@link org.apache.lucene.document.Field}s.</li>
108 <li>The contrib/analyzers library located at the root of the Lucene distribution has a number of different Analyzer implementations to solve a variety
109 of different problems related to searching. Many of the Analyzers are designed to analyze non-English languages.</li>
110 <li>The contrib/snowball library
111 located at the root of the Lucene distribution has Analyzer and TokenFilter
112 implementations for a variety of Snowball stemmers.
113 See <a href="http://snowball.tartarus.org">http://snowball.tartarus.org</a>
114 for more information on Snowball stemmers.</li>
115 <li>There are a variety of Tokenizer and TokenFilter implementations in this package. Take a look around, chances are someone has implemented what you need.</li>
119 Analysis is one of the main causes of performance degradation during indexing. Simply put, the more you analyze the slower the indexing (in most cases).
120 Perhaps your application would be just fine using the simple {@link org.apache.lucene.analysis.WhitespaceTokenizer} combined with a
121 {@link org.apache.lucene.analysis.StopFilter}. The contrib/benchmark library can be useful for testing out the speed of the analysis process.
123 <h2>Invoking the Analyzer</h2>
125 Applications usually do not invoke analysis – Lucene does it for them:
127 <li>At indexing, as a consequence of
128 {@link org.apache.lucene.index.IndexWriter#addDocument(org.apache.lucene.document.Document) addDocument(doc)},
129 the Analyzer in effect for indexing is invoked for each indexed field of the added document.
131 <li>At search, as a consequence of
132 {@link org.apache.lucene.queryParser.QueryParser#parse(java.lang.String) QueryParser.parse(queryText)},
133 the QueryParser may invoke the Analyzer in effect.
134 Note that for some queries analysis does not take place, e.g. wildcard queries.
137 However an application might invoke Analysis of any text for testing or for any other purpose, something like:
138 <PRE class="prettyprint">
139 Analyzer analyzer = new StandardAnalyzer(); // or any other analyzer
140 TokenStream ts = analyzer.tokenStream("myfield",new StringReader("some text goes here"));
141 while (ts.incrementToken()) {
142 System.out.println("token: "+ts));
146 <h2>Indexing Analysis vs. Search Analysis</h2>
148 Selecting the "correct" analyzer is crucial
149 for search quality, and can also affect indexing and search performance.
150 The "correct" analyzer differs between applications.
151 Lucene java's wiki page
152 <a href="http://wiki.apache.org/lucene-java/AnalysisParalysis">AnalysisParalysis</a>
153 provides some data on "analyzing your analyzer".
154 Here are some rules of thumb:
156 <li>Test test test... (did we say test?)</li>
157 <li>Beware of over analysis – might hurt indexing performance.</li>
158 <li>Start with same analyzer for indexing and search, otherwise searches would not find what they are supposed to...</li>
159 <li>In some cases a different analyzer is required for indexing and search, for instance:
161 <li>Certain searches require more stop words to be filtered. (I.e. more than those that were filtered at indexing.)</li>
162 <li>Query expansion by synonyms, acronyms, auto spell correction, etc.</li>
164 This might sometimes require a modified analyzer – see the next section on how to do that.
168 <h2>Implementing your own Analyzer</h2>
169 <p>Creating your own Analyzer is straightforward. It usually involves either wrapping an existing Tokenizer and set of TokenFilters to create a new Analyzer
170 or creating both the Analyzer and a Tokenizer or TokenFilter. Before pursuing this approach, you may find it worthwhile
171 to explore the contrib/analyzers library and/or ask on the java-user@lucene.apache.org mailing list first to see if what you need already exists.
172 If you are still committed to creating your own Analyzer or TokenStream derivation (Tokenizer or TokenFilter) have a look at
173 the source code of any one of the many samples located in this package.
176 The following sections discuss some aspects of implementing your own analyzer.
178 <h3>Field Section Boundaries</h3>
180 When {@link org.apache.lucene.document.Document#add(org.apache.lucene.document.Fieldable) document.add(field)}
181 is called multiple times for the same field name, we could say that each such call creates a new
182 section for that field in that document.
183 In fact, a separate call to
184 {@link org.apache.lucene.analysis.Analyzer#tokenStream(java.lang.String, java.io.Reader) tokenStream(field,reader)}
185 would take place for each of these so called "sections".
186 However, the default Analyzer behavior is to treat all these sections as one large section.
187 This allows phrase search and proximity search to seamlessly cross
188 boundaries between these "sections".
189 In other words, if a certain field "f" is added like this:
190 <PRE class="prettyprint">
191 document.add(new Field("f","first ends",...);
192 document.add(new Field("f","starts two",...);
193 indexWriter.addDocument(document);
195 Then, a phrase search for "ends starts" would find that document.
196 Where desired, this behavior can be modified by introducing a "position gap" between consecutive field "sections",
198 {@link org.apache.lucene.analysis.Analyzer#getPositionIncrementGap(java.lang.String) Analyzer.getPositionIncrementGap(fieldName)}:
199 <PRE class="prettyprint">
200 Analyzer myAnalyzer = new StandardAnalyzer() {
201 public int getPositionIncrementGap(String fieldName) {
207 <h3>Token Position Increments</h3>
209 By default, all tokens created by Analyzers and Tokenizers have a
210 {@link org.apache.lucene.analysis.tokenattributes.PositionIncrementAttribute#getPositionIncrement() position increment} of one.
211 This means that the position stored for that token in the index would be one more than
212 that of the previous token.
213 Recall that phrase and proximity searches rely on position info.
216 If the selected analyzer filters the stop words "is" and "the", then for a document
217 containing the string "blue is the sky", only the tokens "blue", "sky" are indexed,
218 with position("sky") = 1 + position("blue"). Now, a phrase query "blue is the sky"
219 would find that document, because the same analyzer filters the same stop words from
220 that query. But also the phrase query "blue sky" would find that document.
223 If this behavior does not fit the application needs,
224 a modified analyzer can be used, that would increment further the positions of
225 tokens following a removed stop word, using
226 {@link org.apache.lucene.analysis.tokenattributes.PositionIncrementAttribute#setPositionIncrement(int)}.
227 This can be done with something like:
228 <PRE class="prettyprint">
229 public TokenStream tokenStream(final String fieldName, Reader reader) {
230 final TokenStream ts = someAnalyzer.tokenStream(fieldName, reader);
231 TokenStream res = new TokenStream() {
232 TermAttribute termAtt = addAttribute(TermAttribute.class);
233 PositionIncrementAttribute posIncrAtt = addAttribute(PositionIncrementAttribute.class);
235 public boolean incrementToken() throws IOException {
236 int extraIncrement = 0;
238 boolean hasNext = ts.incrementToken();
240 if (stopWords.contains(termAtt.term())) {
241 extraIncrement++; // filter this word
244 if (extraIncrement>0) {
245 posIncrAtt.setPositionIncrement(posIncrAtt.getPositionIncrement()+extraIncrement);
255 Now, with this modified analyzer, the phrase query "blue sky" would find that document.
256 But note that this is yet not a perfect solution, because any phrase query "blue w1 w2 sky"
257 where both w1 and w2 are stop words would match that document.
260 Few more use cases for modifying position increments are:
262 <li>Inhibiting phrase and proximity matches in sentence boundaries – for this, a tokenizer that
263 identifies a new sentence can add 1 to the position increment of the first token of the new sentence.</li>
264 <li>Injecting synonyms – here, synonyms of a token should be added after that token,
265 and their position increment should be set to 0.
266 As result, all synonyms of a token would be considered to appear in exactly the
267 same position as that token, and so would they be seen by phrase and proximity searches.</li>
270 <h2>New TokenStream API</h2>
272 With Lucene 2.9 we introduce a new TokenStream API. The old API used to produce Tokens. A Token
273 has getter and setter methods for different properties like positionIncrement and termText.
274 While this approach was sufficient for the default indexing format, it is not versatile enough for
275 Flexible Indexing, a term which summarizes the effort of making the Lucene indexer pluggable and extensible for custom
279 A fully customizable indexer means that users will be able to store custom data structures on disk. Therefore an API
280 is necessary that can transport custom types of data from the documents to the indexer.
282 <h3>Attribute and AttributeSource</h3>
283 Lucene 2.9 therefore introduces a new pair of classes called {@link org.apache.lucene.util.Attribute} and
284 {@link org.apache.lucene.util.AttributeSource}. An Attribute serves as a
285 particular piece of information about a text token. For example, {@link org.apache.lucene.analysis.tokenattributes.TermAttribute}
286 contains the term text of a token, and {@link org.apache.lucene.analysis.tokenattributes.OffsetAttribute} contains the start and end character offsets of a token.
287 An AttributeSource is a collection of Attributes with a restriction: there may be only one instance of each attribute type. TokenStream now extends AttributeSource, which
288 means that one can add Attributes to a TokenStream. Since TokenFilter extends TokenStream, all filters are also
291 Lucene now provides six Attributes out of the box, which replace the variables the Token class has:
293 <li>{@link org.apache.lucene.analysis.tokenattributes.TermAttribute}<p>The term text of a token.</p></li>
294 <li>{@link org.apache.lucene.analysis.tokenattributes.OffsetAttribute}<p>The start and end offset of token in characters.</p></li>
295 <li>{@link org.apache.lucene.analysis.tokenattributes.PositionIncrementAttribute}<p>See above for detailed information about position increment.</p></li>
296 <li>{@link org.apache.lucene.analysis.tokenattributes.PayloadAttribute}<p>The payload that a Token can optionally have.</p></li>
297 <li>{@link org.apache.lucene.analysis.tokenattributes.TypeAttribute}<p>The type of the token. Default is 'word'.</p></li>
298 <li>{@link org.apache.lucene.analysis.tokenattributes.FlagsAttribute}<p>Optional flags a token can have.</p></li>
301 <h3>Using the new TokenStream API</h3>
302 There are a few important things to know in order to use the new API efficiently which are summarized here. You may want
303 to walk through the example below first and come back to this section afterwards.
305 Please keep in mind that an AttributeSource can only have one instance of a particular Attribute. Furthermore, if
306 a chain of a TokenStream and multiple TokenFilters is used, then all TokenFilters in that chain share the Attributes
307 with the TokenStream.
311 Attribute instances are reused for all tokens of a document. Thus, a TokenStream/-Filter needs to update
312 the appropriate Attribute(s) in incrementToken(). The consumer, commonly the Lucene indexer, consumes the data in the
313 Attributes and then calls incrementToken() again until it returns false, which indicates that the end of the stream
314 was reached. This means that in each call of incrementToken() a TokenStream/-Filter can safely overwrite the data in
315 the Attribute instances.
319 For performance reasons a TokenStream/-Filter should add/get Attributes during instantiation; i.e., create an attribute in the
320 constructor and store references to it in an instance variable. Using an instance variable instead of calling addAttribute()/getAttribute()
321 in incrementToken() will avoid attribute lookups for every token in the document.
325 All methods in AttributeSource are idempotent, which means calling them multiple times always yields the same
326 result. This is especially important to know for addAttribute(). The method takes the <b>type</b> (<code>Class</code>)
327 of an Attribute as an argument and returns an <b>instance</b>. If an Attribute of the same type was previously added, then
328 the already existing instance is returned, otherwise a new instance is created and returned. Therefore TokenStreams/-Filters
329 can safely call addAttribute() with the same Attribute type multiple times. Even consumers of TokenStreams should
330 normally call addAttribute() instead of getAttribute(), because it would not fail if the TokenStream does not have this
331 Attribute (getAttribute() would throw an IllegalArgumentException, if the Attribute is missing). More advanced code
332 could simply check with hasAttribute(), if a TokenStream has it, and may conditionally leave out processing for
336 In this example we will create a WhiteSpaceTokenizer and use a LengthFilter to suppress all words that only
337 have two or less characters. The LengthFilter is part of the Lucene core and its implementation will be explained
338 here to illustrate the usage of the new TokenStream API.<br>
339 Then we will develop a custom Attribute, a PartOfSpeechAttribute, and add another filter to the chain which
340 utilizes the new custom attribute, and call it PartOfSpeechTaggingFilter.
341 <h4>Whitespace tokenization</h4>
342 <pre class="prettyprint">
343 public class MyAnalyzer extends Analyzer {
345 public TokenStream tokenStream(String fieldName, Reader reader) {
346 TokenStream stream = new WhitespaceTokenizer(reader);
350 public static void main(String[] args) throws IOException {
352 final String text = "This is a demo of the new TokenStream API";
354 MyAnalyzer analyzer = new MyAnalyzer();
355 TokenStream stream = analyzer.tokenStream("field", new StringReader(text));
357 // get the TermAttribute from the TokenStream
358 TermAttribute termAtt = stream.addAttribute(TermAttribute.class);
362 // print all tokens until stream is exhausted
363 while (stream.incrementToken()) {
364 System.out.println(termAtt.term());
372 In this easy example a simple white space tokenization is performed. In main() a loop consumes the stream and
373 prints the term text of the tokens by accessing the TermAttribute that the WhitespaceTokenizer provides.
386 <h4>Adding a LengthFilter</h4>
387 We want to suppress all tokens that have 2 or less characters. We can do that easily by adding a LengthFilter
388 to the chain. Only the tokenStream() method in our analyzer needs to be changed:
389 <pre class="prettyprint">
390 public TokenStream tokenStream(String fieldName, Reader reader) {
391 TokenStream stream = new WhitespaceTokenizer(reader);
392 stream = new LengthFilter(stream, 3, Integer.MAX_VALUE);
396 Note how now only words with 3 or more characters are contained in the output:
405 Now let's take a look how the LengthFilter is implemented (it is part of Lucene's core):
406 <pre class="prettyprint">
407 public final class LengthFilter extends TokenFilter {
412 private TermAttribute termAtt;
415 * Build a filter that removes words that are too long or too
416 * short from the text.
418 public LengthFilter(TokenStream in, int min, int max)
423 termAtt = addAttribute(TermAttribute.class);
427 * Returns the next input Token whose term() is the right len
429 public final boolean incrementToken() throws IOException
431 assert termAtt != null;
432 // return the first non-stop word found
433 while (input.incrementToken()) {
434 int len = termAtt.termLength();
435 if (len >= min && len <= max) {
438 // note: else we ignore it but should we index each part of it?
440 // reached EOS -- return null
445 The TermAttribute is added in the constructor and stored in the instance variable <code>termAtt</code>.
446 Remember that there can only be a single instance of TermAttribute in the chain, so in our example the
447 <code>addAttribute()</code> call in LengthFilter returns the TermAttribute that the WhitespaceTokenizer already added. The tokens
448 are retrieved from the input stream in the <code>incrementToken()</code> method. By looking at the term text
449 in the TermAttribute the length of the term can be determined and too short or too long tokens are skipped.
450 Note how <code>incrementToken()</code> can efficiently access the instance variable; no attribute lookup
451 is neccessary. The same is true for the consumer, which can simply use local references to the Attributes.
453 <h4>Adding a custom Attribute</h4>
454 Now we're going to implement our own custom Attribute for part-of-speech tagging and call it consequently
455 <code>PartOfSpeechAttribute</code>. First we need to define the interface of the new Attribute:
456 <pre class="prettyprint">
457 public interface PartOfSpeechAttribute extends Attribute {
458 public static enum PartOfSpeech {
459 Noun, Verb, Adjective, Adverb, Pronoun, Preposition, Conjunction, Article, Unknown
462 public void setPartOfSpeech(PartOfSpeech pos);
464 public PartOfSpeech getPartOfSpeech();
468 Now we also need to write the implementing class. The name of that class is important here: By default, Lucene
469 checks if there is a class with the name of the Attribute with the postfix 'Impl'. In this example, we would
470 consequently call the implementing class <code>PartOfSpeechAttributeImpl</code>. <br/>
471 This should be the usual behavior. However, there is also an expert-API that allows changing these naming conventions:
472 {@link org.apache.lucene.util.AttributeSource.AttributeFactory}. The factory accepts an Attribute interface as argument
473 and returns an actual instance. You can implement your own factory if you need to change the default behavior. <br/><br/>
475 Now here is the actual class that implements our new Attribute. Notice that the class has to extend
476 {@link org.apache.lucene.util.AttributeImpl}:
478 <pre class="prettyprint">
479 public final class PartOfSpeechAttributeImpl extends AttributeImpl
480 implements PartOfSpeechAttribute{
482 private PartOfSpeech pos = PartOfSpeech.Unknown;
484 public void setPartOfSpeech(PartOfSpeech pos) {
488 public PartOfSpeech getPartOfSpeech() {
492 public void clear() {
493 pos = PartOfSpeech.Unknown;
496 public void copyTo(AttributeImpl target) {
497 ((PartOfSpeechAttributeImpl) target).pos = pos;
500 public boolean equals(Object other) {
505 if (other instanceof PartOfSpeechAttributeImpl) {
506 return pos == ((PartOfSpeechAttributeImpl) other).pos;
512 public int hashCode() {
513 return pos.ordinal();
517 This is a simple Attribute implementation has only a single variable that stores the part-of-speech of a token. It extends the
518 new <code>AttributeImpl</code> class and therefore implements its abstract methods <code>clear(), copyTo(), equals(), hashCode()</code>.
519 Now we need a TokenFilter that can set this new PartOfSpeechAttribute for each token. In this example we show a very naive filter
520 that tags every word with a leading upper-case letter as a 'Noun' and all other words as 'Unknown'.
521 <pre class="prettyprint">
522 public static class PartOfSpeechTaggingFilter extends TokenFilter {
523 PartOfSpeechAttribute posAtt;
524 TermAttribute termAtt;
526 protected PartOfSpeechTaggingFilter(TokenStream input) {
528 posAtt = addAttribute(PartOfSpeechAttribute.class);
529 termAtt = addAttribute(TermAttribute.class);
532 public boolean incrementToken() throws IOException {
533 if (!input.incrementToken()) {return false;}
534 posAtt.setPartOfSpeech(determinePOS(termAtt.termBuffer(), 0, termAtt.termLength()));
538 // determine the part of speech for the given term
539 protected PartOfSpeech determinePOS(char[] term, int offset, int length) {
540 // naive implementation that tags every uppercased word as noun
541 if (length > 0 && Character.isUpperCase(term[0])) {
542 return PartOfSpeech.Noun;
544 return PartOfSpeech.Unknown;
548 Just like the LengthFilter, this new filter accesses the attributes it needs in the constructor and
549 stores references in instance variables. Notice how you only need to pass in the interface of the new
550 Attribute and instantiating the correct class is automatically been taken care of.
551 Now we need to add the filter to the chain:
552 <pre class="prettyprint">
553 public TokenStream tokenStream(String fieldName, Reader reader) {
554 TokenStream stream = new WhitespaceTokenizer(reader);
555 stream = new LengthFilter(stream, 3, Integer.MAX_VALUE);
556 stream = new PartOfSpeechTaggingFilter(stream);
560 Now let's look at the output:
569 Apparently it hasn't changed, which shows that adding a custom attribute to a TokenStream/Filter chain does not
570 affect any existing consumers, simply because they don't know the new Attribute. Now let's change the consumer
571 to make use of the new PartOfSpeechAttribute and print it out:
572 <pre class="prettyprint">
573 public static void main(String[] args) throws IOException {
575 final String text = "This is a demo of the new TokenStream API";
577 MyAnalyzer analyzer = new MyAnalyzer();
578 TokenStream stream = analyzer.tokenStream("field", new StringReader(text));
580 // get the TermAttribute from the TokenStream
581 TermAttribute termAtt = stream.addAttribute(TermAttribute.class);
583 // get the PartOfSpeechAttribute from the TokenStream
584 PartOfSpeechAttribute posAtt = stream.addAttribute(PartOfSpeechAttribute.class);
588 // print all tokens until stream is exhausted
589 while (stream.incrementToken()) {
590 System.out.println(termAtt.term() + ": " + posAtt.getPartOfSpeech());
597 The change that was made is to get the PartOfSpeechAttribute from the TokenStream and print out its contents in
598 the while loop that consumes the stream. Here is the new output:
607 Each word is now followed by its assigned PartOfSpeech tag. Of course this is a naive
608 part-of-speech tagging. The word 'This' should not even be tagged as noun; it is only spelled capitalized because it
609 is the first word of a sentence. Actually this is a good opportunity for an excerise. To practice the usage of the new
610 API the reader could now write an Attribute and TokenFilter that can specify for each word if it was the first token
611 of a sentence or not. Then the PartOfSpeechTaggingFilter can make use of this knowledge and only tag capitalized words
612 as nouns if not the first word of a sentence (we know, this is still not a correct behavior, but hey, it's a good exercise).
613 As a small hint, this is how the new Attribute class could begin:
614 <pre class="prettyprint">
615 public class FirstTokenOfSentenceAttributeImpl extends Attribute
616 implements FirstTokenOfSentenceAttribute {
618 private boolean firstToken;
620 public void setFirstToken(boolean firstToken) {
621 this.firstToken = firstToken;
624 public boolean getFirstToken() {
628 public void clear() {