* Licensed to the Apache Software Foundation (ASF) under one or more

* contributor license agreements.  See the NOTICE file distributed with

* this work for additional information regarding copyright ownership.

* The ASF licenses this file to You under the Apache License, Version 2.0

* (the "License"); you may not use this file except in compliance with

* the License.  You may obtain a copy of the License at

*

*     http://www.apache.org/licenses/LICENSE-2.0

*

* Unless required by applicable law or agreed to in writing, software

* distributed under the License is distributed on an "AS IS" BASIS,

* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

* See the License for the specific language governing permissions and

* limitations under the License.

*/

 

  public void testRandomEdits() throws IOException {

    List<Input> keys = new ArrayList<Input>();

    int numTerms = atLeast(100);

    for (int i = 0; i < numTerms; i++) {

      keys.add(new Input("boo" + _TestUtil.randomSimpleString(random()), 1 + random().nextInt(100)));

    }

    keys.add(new Input("foo bar boo far", 12));

    MockAnalyzer analyzer = new MockAnalyzer(random(), MockTokenizer.KEYWORD, false);

    FuzzySuggester suggester = new FuzzySuggester(analyzer, analyzer, FuzzySuggester.EXACT_FIRST | FuzzySuggester.PRESERVE_SEP, 256, -1, true, FuzzySuggester.DEFAULT_MAX_EDITS, FuzzySuggester.DEFAULT_TRANSPOSITIONS,

                                                  0, FuzzySuggester.DEFAULT_MIN_FUZZY_LENGTH, FuzzySuggester.DEFAULT_UNICODE_AWARE);

    suggester.build(new InputArrayIterator(keys));

    int numIters = atLeast(10);

    for (int i = 0; i < numIters; i++) {

      String addRandomEdit = addRandomEdit("foo bar boo", FuzzySuggester.DEFAULT_NON_FUZZY_PREFIX);

      List<LookupResult> results = suggester.lookup(_TestUtil.stringToCharSequence(addRandomEdit, random()), false, 2);

      assertEquals(addRandomEdit, 1, results.size());

      assertEquals("foo bar boo far", results.get(0).key.toString());

      assertEquals(12, results.get(0).value, 0.01F); 

    }

  }

 

  public void testNonLatinRandomEdits() throws IOException {

    List<Input> keys = new ArrayList<Input>();

    int numTerms = atLeast(100);

    for (int i = 0; i < numTerms; i++) {

      keys.add(new Input("буу" + _TestUtil.randomSimpleString(random()), 1 + random().nextInt(100)));

    }

    keys.add(new Input("фуу бар буу фар", 12));

    MockAnalyzer analyzer = new MockAnalyzer(random(), MockTokenizer.KEYWORD, false);

    FuzzySuggester suggester = new FuzzySuggester(analyzer, analyzer, FuzzySuggester.EXACT_FIRST | FuzzySuggester.PRESERVE_SEP, 256, -1, true, FuzzySuggester.DEFAULT_MAX_EDITS, FuzzySuggester.DEFAULT_TRANSPOSITIONS,

        0, FuzzySuggester.DEFAULT_MIN_FUZZY_LENGTH, true);

    suggester.build(new InputArrayIterator(keys));

    int numIters = atLeast(10);

    for (int i = 0; i < numIters; i++) {

      String addRandomEdit = addRandomEdit("фуу бар буу", 0);

      List<LookupResult> results = suggester.lookup(_TestUtil.stringToCharSequence(addRandomEdit, random()), false, 2);

      assertEquals(addRandomEdit, 1, results.size());

      assertEquals("фуу бар буу фар", results.get(0).key.toString());

      assertEquals(12, results.get(0).value, 0.01F);

    }

  }

  /** this is basically the WFST test ported to KeywordAnalyzer. so it acts the same */

  public void testKeyword() throws Exception {

    Input keys[] = new Input[] {

        new Input("foo", 50),

        new Input("bar", 10),

        new Input("barbar", 12),

        new Input("barbara", 6)

    };

   

    FuzzySuggester suggester = new FuzzySuggester(new MockAnalyzer(random(), MockTokenizer.KEYWORD, false));

    suggester.build(new InputArrayIterator(keys));

   

    List<LookupResult> results = suggester.lookup(_TestUtil.stringToCharSequence("bariar", random()), false, 2);

    assertEquals(2, results.size());

    assertEquals("barbar", results.get(0).key.toString());

    assertEquals(12, results.get(0).value, 0.01F);

   

    results = suggester.lookup(_TestUtil.stringToCharSequence("barbr", random()), false, 2);

    assertEquals(2, results.size());

    assertEquals("barbar", results.get(0).key.toString());

    assertEquals(12, results.get(0).value, 0.01F);

   

    results = suggester.lookup(_TestUtil.stringToCharSequence("barbara", random()), false, 2);

    assertEquals(2, results.size());

    assertEquals("barbara", results.get(0).key.toString());

    assertEquals(6, results.get(0).value, 0.01F);

   

    results = suggester.lookup(_TestUtil.stringToCharSequence("barbar", random()), false, 2);

    assertEquals(2, results.size());

    assertEquals("barbar", results.get(0).key.toString());

    assertEquals(12, results.get(0).value, 0.01F);

    assertEquals("barbara", results.get(1).key.toString());

    assertEquals(6, results.get(1).value, 0.01F);

   

    results = suggester.lookup(_TestUtil.stringToCharSequence("barbaa", random()), false, 2);

    assertEquals(2, results.size());

    assertEquals("barbar", results.get(0).key.toString());

    assertEquals(12, results.get(0).value, 0.01F);

    assertEquals("barbara", results.get(1).key.toString());

    assertEquals(6, results.get(1).value, 0.01F);

   

    // top N of 2, but only foo is available

    results = suggester.lookup(_TestUtil.stringToCharSequence("f", random()), false, 2);

    assertEquals(1, results.size());

    assertEquals("foo", results.get(0).key.toString());

    assertEquals(50, results.get(0).value, 0.01F);

   

    // top N of 1 for 'bar': we return this even though

    // barbar is higher because exactFirst is enabled:

    results = suggester.lookup(_TestUtil.stringToCharSequence("bar", random()), false, 1);

    assertEquals(1, results.size());

    assertEquals("bar", results.get(0).key.toString());

    assertEquals(10, results.get(0).value, 0.01F);

   

    // top N Of 2 for 'b'

    results = suggester.lookup(_TestUtil.stringToCharSequence("b", random()), false, 2);

    assertEquals(2, results.size());

    assertEquals("barbar", results.get(0).key.toString());

    assertEquals(12, results.get(0).value, 0.01F);

    assertEquals("bar", results.get(1).key.toString());

    assertEquals(10, results.get(1).value, 0.01F);

   

    // top N of 3 for 'ba'

    results = suggester.lookup(_TestUtil.stringToCharSequence("ba", random()), false, 3);

    assertEquals(3, results.size());

    assertEquals("barbar", results.get(0).key.toString());

    assertEquals(12, results.get(0).value, 0.01F);

    assertEquals("bar", results.get(1).key.toString());

    assertEquals(10, results.get(1).value, 0.01F);

    assertEquals("barbara", results.get(2).key.toString());

    assertEquals(6, results.get(2).value, 0.01F);

  }

 

  /**

   * basic "standardanalyzer" test with stopword removal

   */

  public void testStandard() throws Exception {

    Input keys[] = new Input[] {

        new Input("the ghost of christmas past", 50),

    };

   

    Analyzer standard = new MockAnalyzer(random(), MockTokenizer.WHITESPACE, true, MockTokenFilter.ENGLISH_STOPSET);

    FuzzySuggester suggester = new FuzzySuggester(standard, standard, AnalyzingSuggester.EXACT_FIRST | AnalyzingSuggester.PRESERVE_SEP, 256, -1, false, FuzzySuggester.DEFAULT_MAX_EDITS, FuzzySuggester.DEFAULT_TRANSPOSITIONS,

        FuzzySuggester.DEFAULT_NON_FUZZY_PREFIX, FuzzySuggester.DEFAULT_MIN_FUZZY_LENGTH, FuzzySuggester.DEFAULT_UNICODE_AWARE);

    suggester.build(new InputArrayIterator(keys));

   

    List<LookupResult> results = suggester.lookup(_TestUtil.stringToCharSequence("the ghost of chris", random()), false, 1);

    assertEquals(1, results.size());

    assertEquals("the ghost of christmas past", results.get(0).key.toString());

    assertEquals(50, results.get(0).value, 0.01F);

    // omit the 'the' since its a stopword, its suggested anyway

    results = suggester.lookup(_TestUtil.stringToCharSequence("ghost of chris", random()), false, 1);

    assertEquals(1, results.size());

    assertEquals("the ghost of christmas past", results.get(0).key.toString());

    assertEquals(50, results.get(0).value, 0.01F);

    // omit the 'the' and 'of' since they are stopwords, its suggested anyway

    results = suggester.lookup(_TestUtil.stringToCharSequence("ghost chris", random()), false, 1);

    assertEquals(1, results.size());

    assertEquals("the ghost of christmas past", results.get(0).key.toString());

    assertEquals(50, results.get(0).value, 0.01F);

  }

  public void testNoSeps() throws Exception {

    Input[] keys = new Input[] {

      new Input("ab cd", 0),

      new Input("abcd", 1),

    };

    int options = 0;

    Analyzer a = new MockAnalyzer(random());

    FuzzySuggester suggester = new FuzzySuggester(a, a, options, 256, -1, true, 1, true, 1, 3, false);

    suggester.build(new InputArrayIterator(keys));

    // TODO: would be nice if "ab " would allow the test to

    // pass, and more generally if the analyzer can know

    // that the user's current query has ended at a word,

    // but, analyzers don't produce SEP tokens!

    List<LookupResult> r = suggester.lookup(_TestUtil.stringToCharSequence("ab c", random()), false, 2);

    assertEquals(2, r.size());

    // With no PRESERVE_SEPS specified, "ab c" should also

    // complete to "abcd", which has higher weight so should

    // appear first:

    assertEquals("abcd", r.get(0).key.toString());

  }

  public void testGraphDups() throws Exception {

    final Analyzer analyzer = new Analyzer() {

      @Override

      protected TokenStreamComponents createComponents(String fieldName, Reader reader) {

        Tokenizer tokenizer = new MockTokenizer(reader, MockTokenizer.SIMPLE, true);

       

        return new TokenStreamComponents(tokenizer) {

          int tokenStreamCounter = 0;

          final TokenStream[] tokenStreams = new TokenStream[] {

            new CannedTokenStream(new Token[] {

                token("wifi",1,1),

                token("hotspot",0,2),

                token("network",1,1),

                token("is",1,1),

                token("slow",1,1)

              }),

            new CannedTokenStream(new Token[] {

                token("wi",1,1),

                token("hotspot",0,3),

                token("fi",1,1),

                token("network",1,1),

                token("is",1,1),

                token("fast",1,1)

              }),

            new CannedTokenStream(new Token[] {

                token("wifi",1,1),

                token("hotspot",0,2),

                token("network",1,1)

              }),

          };

          @Override

          public TokenStream getTokenStream() {

            TokenStream result = tokenStreams[tokenStreamCounter];

            tokenStreamCounter++;

            return result;

          }

        

          @Override

          protected void setReader(final Reader reader) throws IOException {

          }

        };

      }

    };

    Input keys[] = new Input[] {

        new Input("wifi network is slow", 50),

        new Input("wi fi network is fast", 10),

    };

    FuzzySuggester suggester = new FuzzySuggester(analyzer);

    suggester.build(new InputArrayIterator(keys));

   

    List<LookupResult> results = suggester.lookup("wifi network", false, 10);

    if (VERBOSE) {

      System.out.println("Results: " + results);

    }

    assertEquals(2, results.size());

    assertEquals("wifi network is slow", results.get(0).key);

    assertEquals(50, results.get(0).value);

    assertEquals("wi fi network is fast", results.get(1).key);

    assertEquals(10, results.get(1).value);

  }

  public void testEmpty() throws Exception {

    FuzzySuggester suggester = new FuzzySuggester(new MockAnalyzer(random(), MockTokenizer.KEYWORD, false));

    suggester.build(new InputArrayIterator(new Input[0]));

    List<LookupResult> result = suggester.lookup("a", false, 20);

    assertTrue(result.isEmpty());

  }

  public void testInputPathRequired() throws Exception {

    //  SynonymMap.Builder b = new SynonymMap.Builder(false);

    //  b.add(new CharsRef("ab"), new CharsRef("ba"), true);

    //  final SynonymMap map = b.build();

    //  The Analyzer below mimics the functionality of the SynonymAnalyzer

    //  using the above map, so that the suggest module does not need a dependency on the

    //  synonym module

    final Analyzer analyzer = new Analyzer() {

      @Override

      protected TokenStreamComponents createComponents(String fieldName, Reader reader) {

        Tokenizer tokenizer = new MockTokenizer(reader, MockTokenizer.SIMPLE, true);

       

        return new TokenStreamComponents(tokenizer) {

          int tokenStreamCounter = 0;

          final TokenStream[] tokenStreams = new TokenStream[] {

            new CannedTokenStream(new Token[] {

                token("ab",1,1),

                token("ba",0,1),

                token("xc",1,1)

              }),

            new CannedTokenStream(new Token[] {

                token("ba",1,1),         

                token("xd",1,1)

              }),

            new CannedTokenStream(new Token[] {

                token("ab",1,1),

                token("ba",0,1),

                token("x",1,1)

              })

          };

          @Override

          public TokenStream getTokenStream() {

            TokenStream result = tokenStreams[tokenStreamCounter];

            tokenStreamCounter++;

            return result;

          }

        

          @Override

          protected void setReader(final Reader reader) throws IOException {

          }

        };

      }

    };

    Input keys[] = new Input[] {

        new Input("ab xc", 50),

        new Input("ba xd", 50),

    };

    FuzzySuggester suggester = new FuzzySuggester(analyzer);

    suggester.build(new InputArrayIterator(keys));

    List<LookupResult> results = suggester.lookup("ab x", false, 1);

    assertTrue(results.size() == 1);

  }

  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;

  }

  /*

  private void printTokens(final Analyzer analyzer, String input) throws IOException {

    System.out.println("Tokens for " + input);

    TokenStream ts = analyzer.tokenStream("", new StringReader(input));

    ts.reset();

    final TermToBytesRefAttribute termBytesAtt = ts.addAttribute(TermToBytesRefAttribute.class);

    final PositionIncrementAttribute posIncAtt = ts.addAttribute(PositionIncrementAttribute.class);

    final PositionLengthAttribute posLengthAtt = ts.addAttribute(PositionLengthAttribute.class);

   

    while(ts.incrementToken()) {

      termBytesAtt.fillBytesRef();

      System.out.println(String.format("%s,%s,%s", termBytesAtt.getBytesRef().utf8ToString(), posIncAtt.getPositionIncrement(), posLengthAtt.getPositionLength()));     

    }

    ts.end();

    ts.close();

  }

  */

  private final Analyzer getUnusualAnalyzer() {

    return new Analyzer() {

      @Override

      protected TokenStreamComponents createComponents(String fieldName, Reader reader) {

        Tokenizer tokenizer = new MockTokenizer(reader, MockTokenizer.SIMPLE, true);

       

        return new TokenStreamComponents(tokenizer) {

          int count;

          @Override

          public TokenStream getTokenStream() {

            // 4th time we are called, return tokens a b,

            // else just a:

            if (count++ != 3) {

              return new CannedTokenStream(new Token[] {

                  token("a", 1, 1),

                });

            } else {

              // After that "a b":

              return new CannedTokenStream(new Token[] {

                  token("a", 1, 1),

                  token("b", 1, 1),

                });

            }

          }

        

          @Override

          protected void setReader(final Reader reader) throws IOException {

          }

        };

      }

    };

  }

  public void testExactFirst() throws Exception {

    Analyzer a = getUnusualAnalyzer();

    FuzzySuggester suggester = new FuzzySuggester(a, a, AnalyzingSuggester.EXACT_FIRST | AnalyzingSuggester.PRESERVE_SEP, 256, -1, true, 1, true, 1, 3, false);

    suggester.build(new InputArrayIterator(new Input[] {

          new Input("x y", 1),

          new Input("x y z", 3),

          new Input("x", 2),

          new Input("z z z", 20),

        }));

    //System.out.println("ALL: " + suggester.lookup("x y", false, 6));

    for(int topN=1;topN<6;topN++) {

      List<LookupResult> results = suggester.lookup("x y", false, topN);

      //System.out.println("topN=" + topN + " " + results);

      assertEquals(Math.min(topN, 4), results.size());

      assertEquals("x y", results.get(0).key);

      assertEquals(1, results.get(0).value);

      if (topN > 1) {

        assertEquals("z z z", results.get(1).key);

        assertEquals(20, results.get(1).value);

        if (topN > 2) {

          assertEquals("x y z", results.get(2).key);

          assertEquals(3, results.get(2).value);

          if (topN > 3) {

            assertEquals("x", results.get(3).key);

            assertEquals(2, results.get(3).value);

          }

        }

      }

    }

  }

  public void testNonExactFirst() throws Exception {

    Analyzer a = getUnusualAnalyzer();

    FuzzySuggester suggester = new FuzzySuggester(a, a, AnalyzingSuggester.PRESERVE_SEP, 256, -1, true, 1, true, 1, 3, false);

    suggester.build(new InputArrayIterator(new Input[] {

          new Input("x y", 1),

          new Input("x y z", 3),

          new Input("x", 2),

          new Input("z z z", 20),

        }));

    for(int topN=1;topN<6;topN++) {

      List<LookupResult> results = suggester.lookup("p", false, topN);

      assertEquals(Math.min(topN, 4), results.size());

      assertEquals("z z z", results.get(0).key);

      assertEquals(20, results.get(0).value);

      if (topN > 1) {

        assertEquals("x y z", results.get(1).key);

        assertEquals(3, results.get(1).value);

        if (topN > 2) {

          assertEquals("x", results.get(2).key);

          assertEquals(2, results.get(2).value);

         

          if (topN > 3) {

            assertEquals("x y", results.get(3).key);

            assertEquals(1, results.get(3).value);

          }

        }

      }

    }

  }

 

  // Holds surface form separately:

  private static class TermFreqPayload2 implements Comparable<TermFreqPayload2> {

    public final String surfaceForm;

    public final String analyzedForm;

    public final long weight;

    public TermFreqPayload2(String surfaceForm, String analyzedForm, long weight) {

      this.surfaceForm = surfaceForm;

      this.analyzedForm = analyzedForm;

      this.weight = weight;

    }

    @Override

    public int compareTo(TermFreqPayload2 other) {

      int cmp = analyzedForm.compareTo(other.analyzedForm);

      if (cmp != 0) {

        return cmp;

      } else if (weight > other.weight) {

        return -1;

      } else if (weight < other.weight) {

        return 1;

      } else {

        assert false;

        return 0;

      }

    }

  }

  static boolean isStopChar(char ch, int numStopChars) {

    //System.out.println("IS? " + ch + ": " + (ch - 'a') + ": " + ((ch - 'a') < numStopChars));

    return (ch - 'a') < numStopChars;

  }

  // Like StopFilter:

  private static class TokenEater extends TokenFilter {

    private final PositionIncrementAttribute posIncrAtt = addAttribute(PositionIncrementAttribute.class);

    private final CharTermAttribute termAtt = addAttribute(CharTermAttribute.class);

    private final int numStopChars;

    private final boolean preserveHoles;

    private boolean first;

    public TokenEater(boolean preserveHoles, TokenStream in, int numStopChars) {

      super(in);

      this.preserveHoles = preserveHoles;

      this.numStopChars = numStopChars;

    }

    @Override

    public void reset() throws IOException {

      super.reset();

      first = true;

    }

    @Override

    public final boolean incrementToken() throws IOException {

      int skippedPositions = 0;

      while (input.incrementToken()) {

        if (termAtt.length() != 1 || !isStopChar(termAtt.charAt(0), numStopChars)) {

          int posInc = posIncrAtt.getPositionIncrement() + skippedPositions;

          if (first) {

            if (posInc == 0) {

              // first token having posinc=0 is illegal.

              posInc = 1;

            }

            first = false;

          }

          posIncrAtt.setPositionIncrement(posInc);

          //System.out.println("RETURN term=" + termAtt + " numStopChars=" + numStopChars);

          return true;

        }

        if (preserveHoles) {

          skippedPositions += posIncrAtt.getPositionIncrement();

        }

      }

      return false;

    }

  }

  private static class MockTokenEatingAnalyzer extends Analyzer {

    private int numStopChars;

    private boolean preserveHoles;

    public MockTokenEatingAnalyzer(int numStopChars, boolean preserveHoles) {

      this.preserveHoles = preserveHoles;

      this.numStopChars = numStopChars;

    }

    @Override

    public TokenStreamComponents createComponents(String fieldName, Reader reader) {

      MockTokenizer tokenizer = new MockTokenizer(reader, MockTokenizer.WHITESPACE, false, MockTokenizer.DEFAULT_MAX_TOKEN_LENGTH);

      tokenizer.setEnableChecks(true);

      TokenStream next;

      if (numStopChars != 0) {

        next = new TokenEater(preserveHoles, tokenizer, numStopChars);

      } else {

        next = tokenizer;

      }

      return new TokenStreamComponents(tokenizer, next);

    }

  }

  public void testRandom() throws Exception {

    int numQueries = atLeast(100);

   

    final List<TermFreqPayload2> slowCompletor = new ArrayList<TermFreqPayload2>();

    final TreeSet<String> allPrefixes = new TreeSet<String>();

    final Set<String> seen = new HashSet<String>();

   

    Input[] keys = new Input[numQueries];

    boolean preserveSep = random().nextBoolean();

    boolean unicodeAware = random().nextBoolean();

    final int numStopChars = random().nextInt(10);

    final boolean preserveHoles = random().nextBoolean();

    if (VERBOSE) {

      System.out.println("TEST: " + numQueries + " words; preserveSep=" + preserveSep + " ; unicodeAware=" + unicodeAware + " numStopChars=" + numStopChars + " preserveHoles=" + preserveHoles);

    }

   

    for (int i = 0; i < numQueries; i++) {

      int numTokens = _TestUtil.nextInt(random(), 1, 4);

      String key;

      String analyzedKey;

      while(true) {

        key = "";

        analyzedKey = "";

        boolean lastRemoved = false;

        for(int token=0;token < numTokens;token++) {

          String s;

          while (true) {

            // TODO: would be nice to fix this slowCompletor/comparator to

            // use full range, but we might lose some coverage too...

            s = _TestUtil.randomSimpleString(random());

            if (s.length() > 0) {

              if (token > 0) {

                key += " ";

              }

              if (preserveSep && analyzedKey.length() > 0 && (unicodeAware ? analyzedKey.codePointAt(analyzedKey.codePointCount(0, analyzedKey.length())-1) != ' ' : analyzedKey.charAt(analyzedKey.length()-1) != ' ')) {

                analyzedKey += " ";

              }

              key += s;

              if (s.length() == 1 && isStopChar(s.charAt(0), numStopChars)) {

                if (preserveSep && preserveHoles) {

                  analyzedKey += '\u0000';

                }

                lastRemoved = true;

              } else {

                analyzedKey += s;

                lastRemoved = false;

              }

              break;

            }

          }

        }

        analyzedKey = analyzedKey.replaceAll("(^| )\u0000$", "");

        if (preserveSep && lastRemoved) {

          analyzedKey += " ";

        }

        // Don't add same surface form more than once:

        if (!seen.contains(key)) {

          seen.add(key);

          break;

        }

      }

      for (int j = 1; j < key.length(); j++) {

        allPrefixes.add(key.substring(0, j));

      }

      // we can probably do Integer.MAX_VALUE here, but why worry.

      int weight = random().nextInt(1<<24);

      keys[i] = new Input(key, weight);

      slowCompletor.add(new TermFreqPayload2(key, analyzedKey, weight));

    }

    if (VERBOSE) {

      // Don't just sort original list, to avoid VERBOSE

      // altering the test:

      List<TermFreqPayload2> sorted = new ArrayList<TermFreqPayload2>(slowCompletor);

      Collections.sort(sorted);

      for(TermFreqPayload2 ent : sorted) {

        System.out.println("  surface='" + ent.surfaceForm + " analyzed='" + ent.analyzedForm + "' weight=" + ent.weight);

      }

    }

    Analyzer a = new MockTokenEatingAnalyzer(numStopChars, preserveHoles);

    FuzzySuggester suggester = new FuzzySuggester(a, a,

                                                  preserveSep ? AnalyzingSuggester.PRESERVE_SEP : 0, 256, -1, true, 1, false, 1, 3, unicodeAware);

    suggester.build(new InputArrayIterator(keys));

    for (String prefix : allPrefixes) {

      if (VERBOSE) {

        System.out.println("\nTEST: prefix=" + prefix);

      }

      final int topN = _TestUtil.nextInt(random(), 1, 10);

      List<LookupResult> r = suggester.lookup(_TestUtil.stringToCharSequence(prefix, random()), false, topN);

      // 2. go thru whole set to find suggestions:

      List<LookupResult> matches = new ArrayList<LookupResult>();

      // "Analyze" the key:

      String[] tokens = prefix.split(" ");

      StringBuilder builder = new StringBuilder();

      boolean lastRemoved = false;

      for(int i=0;i<tokens.length;i++) {

        String token = tokens[i];

        if (preserveSep && builder.length() > 0 && !builder.toString().endsWith(" ")) {

          builder.append(' ');

        }

        if (token.length() == 1 && isStopChar(token.charAt(0), numStopChars)) {

          if (preserveSep && preserveHoles) {

            builder.append("\u0000");

          }

          lastRemoved = true;

        } else {

          builder.append(token);

          lastRemoved = false;

        }

      }

      String analyzedKey = builder.toString();

      // Remove trailing sep/holes (TokenStream.end() does

      // not tell us any trailing holes, yet ... there is an

      // issue open for this):

      while (true) {

        String s = analyzedKey.replaceAll("(^| )\u0000$", "");

        s = s.replaceAll("\\s+$", "");

        if (s.equals(analyzedKey)) {

          break;

        }

        analyzedKey = s;

      }

      if (analyzedKey.length() == 0) {

        // Currently suggester can't suggest from the empty

        // string!  You get no results, not all results...

        continue;

      }

      if (preserveSep && (prefix.endsWith(" ") || lastRemoved)) {

        analyzedKey += " ";

      }

      if (VERBOSE) {

        System.out.println("  analyzed: " + analyzedKey);

      }

      TokenStreamToAutomaton tokenStreamToAutomaton = suggester.getTokenStreamToAutomaton();

      // NOTE: not great that we ask the suggester to give

      // us the "answer key" (ie maybe we have a bug in

      // suggester.toLevA ...) ... but testRandom2() fixes

      // this:

      Automaton automaton = suggester.convertAutomaton(suggester.toLevenshteinAutomata(suggester.toLookupAutomaton(analyzedKey)));

      assertTrue(automaton.isDeterministic());

      // TODO: could be faster... but its slowCompletor for a reason

      BytesRef spare = new BytesRef();

      for (TermFreqPayload2 e : slowCompletor) {

        spare.copyChars(e.analyzedForm);

        Set<IntsRef> finiteStrings = suggester.toFiniteStrings(spare, tokenStreamToAutomaton);

        for (IntsRef intsRef : finiteStrings) {

          State p = automaton.getInitialState();

          BytesRef ref = Util.toBytesRef(intsRef, spare);

          boolean added = false;

          for (int i = ref.offset; i < ref.length; i++) {

            State q = p.step(ref.bytes[i] & 0xff);

            if (q == null) {

              break;

            } else if (q.isAccept()) {

              matches.add(new LookupResult(e.surfaceForm, e.weight));

              added = true;

              break;

            }

            p = q;

          }

          if (!added && p.isAccept()) {

            matches.add(new LookupResult(e.surfaceForm, e.weight));

          }

        }

      }

      assertTrue(numStopChars > 0 || matches.size() > 0);

      if (matches.size() > 1) {

        Collections.sort(matches, new Comparator<LookupResult>() {

            @Override

            public int compare(LookupResult left, LookupResult right) {

              int cmp = Float.compare(right.value, left.value);

              if (cmp == 0) {

                return left.compareTo(right);

              } else {

                return cmp;

              }

            }

          });

      }

      if (matches.size() > topN) {

        matches = matches.subList(0, topN);

      }

      if (VERBOSE) {

        System.out.println("  expected:");

        for(LookupResult lr : matches) {

          System.out.println("    key=" + lr.key + " weight=" + lr.value);

        }

        System.out.println("  actual:");

        for(LookupResult lr : r) {

          System.out.println("    key=" + lr.key + " weight=" + lr.value);

        }

      }

     

      assertEquals(prefix + "  " + topN, matches.size(), r.size());

      for(int hit=0;hit<r.size();hit++) {

        //System.out.println("  check hit " + hit);

        assertEquals(prefix + "  " + topN, matches.get(hit).key.toString(), r.get(hit).key.toString());

        assertEquals(matches.get(hit).value, r.get(hit).value, 0f);

      }

    }

  }

  public void testMaxSurfaceFormsPerAnalyzedForm() throws Exception {

    Analyzer a = new MockAnalyzer(random());

    FuzzySuggester suggester = new FuzzySuggester(a, a, 0, 2, -1, true, 1, true, 1, 3, false);

    List<Input> keys = Arrays.asList(new Input[] {

        new Input("a", 40),

        new Input("a ", 50),

        new Input(" a", 60),

      });

    Collections.shuffle(keys, random());

    suggester.build(new InputArrayIterator(keys));

    List<LookupResult> results = suggester.lookup("a", false, 5);

    assertEquals(2, results.size());

    assertEquals(" a", results.get(0).key);

    assertEquals(60, results.get(0).value);

    assertEquals("a ", results.get(1).key);

    assertEquals(50, results.get(1).value);

  }

  public void testEditSeps() throws Exception {

    Analyzer a = new MockAnalyzer(random());

    FuzzySuggester suggester = new FuzzySuggester(a, a, FuzzySuggester.PRESERVE_SEP, 2, -1, true, 2, true, 1, 3, false);

    List<Input> keys = Arrays.asList(new Input[] {

        new Input("foo bar", 40),

        new Input("foo bar baz", 50),

        new Input("barbaz", 60),

        new Input("barbazfoo", 10),

      });

    Collections.shuffle(keys, random());

    suggester.build(new InputArrayIterator(keys));

    assertEquals("[foo bar baz/50, foo bar/40]", suggester.lookup("foobar", false, 5).toString());

    assertEquals("[foo bar baz/50]", suggester.lookup("foobarbaz", false, 5).toString());

    assertEquals("[barbaz/60, barbazfoo/10]", suggester.lookup("bar baz", false, 5).toString());

    assertEquals("[barbazfoo/10]", suggester.lookup("bar baz foo", false, 5).toString());

  }

 

  @SuppressWarnings("fallthrough")

  private static String addRandomEdit(String string, int prefixLength) {

    char[] input = string.toCharArray();

    StringBuilder builder = new StringBuilder();

    for (int i = 0; i < input.length; i++) {

      if (i >= prefixLength && random().nextBoolean() && i < input.length-1) {

        switch(random().nextInt(4)) {

          case 3:

            if (i < input.length-1) {

              // Transpose input[i] and input[1+i]:

              builder.append(input[i+1]);

              builder.append(input[i]);

              for(int j=i+2;j<input.length;j++) {

                builder.append(input[j]);

              }

              return builder.toString();

            }

            // NOTE: fall through to delete:

          case 2:

            // Delete input[i]

            for (int j = i+1; j < input.length; j++) {

              builder.append(input[j]); 

            }

            return builder.toString();

          case 1:

            // Insert input[i+1] twice

            if (i+1<input.length) {

              builder.append(input[i+1]);

              builder.append(input[i++]);

              i++;

            }

            for (int j = i; j < input.length; j++) {

              builder.append(input[j]);

            }

            return builder.toString();

          case 0:

            // Insert random byte.

            // NOTE: can only use ascii here so that, in

            // UTF8 byte space it's still a single

            // insertion:

            // bytes 0x1e and 0x1f are reserved

            int x = random().nextBoolean() ? random().nextInt(30) :  32 + random().nextInt(128 - 32);

            builder.append((char) x);

            for (int j = i; j < input.length; j++) {

              builder.append(input[j]); 

            }

            return builder.toString();

        }

      }

      builder.append(input[i]);

    }

    return builder.toString();

  }

  private String randomSimpleString(int maxLen) {

    final int len = _TestUtil.nextInt(random(), 1, maxLen);

    final char[] chars = new char[len];

    for(int j=0;j<len;j++) {

      chars[j] = (char) ('a' + random().nextInt(4));

    }

    return new String(chars);

  }

  public void testRandom2() throws Throwable {

    final int NUM = atLeast(200);

    final List<Input> answers = new ArrayList<Input>();

    final Set<String> seen = new HashSet<String>();

    for(int i=0;i<NUM;i++) {

      final String s = randomSimpleString(8);

      if (!seen.contains(s)) {

        answers.add(new Input(s, random().nextInt(1000)));

        seen.add(s);

      }

    }

    Collections.sort(answers, new Comparator<Input>() {

        @Override

        public int compare(Input a, Input b) {

          return a.term.compareTo(b.term);

        }

      });

    if (VERBOSE) {

      System.out.println("\nTEST: targets");

      for(Input tf : answers) {

        System.out.println("  " + tf.term.utf8ToString() + " freq=" + tf.v);

      }

    }

    Analyzer a = new MockAnalyzer(random(), MockTokenizer.KEYWORD, false);

    int maxEdits = random().nextBoolean() ? 1 : 2;

    int prefixLen = random().nextInt(4);

    boolean transpositions = random().nextBoolean();

    // TODO: test graph analyzers

    // TODO: test exactFirst / preserveSep permutations

    FuzzySuggester suggest = new FuzzySuggester(a, a, 0, 256, -1, true, maxEdits, transpositions, prefixLen, prefixLen, false);

    if (VERBOSE) {

      System.out.println("TEST: maxEdits=" + maxEdits + " prefixLen=" + prefixLen + " transpositions=" + transpositions + " num=" + NUM);

    }

    Collections.shuffle(answers, random());

    suggest.build(new InputArrayIterator(answers.toArray(new Input[answers.size()])));

    final int ITERS = atLeast(100);

    for(int iter=0;iter<ITERS;iter++) {

      final String frag = randomSimpleString(6);

      if (VERBOSE) {

        System.out.println("\nTEST: iter frag=" + frag);

      }

      final List<LookupResult> expected = slowFuzzyMatch(prefixLen, maxEdits, transpositions, answers, frag);

      if (VERBOSE) {

        System.out.println("  expected: " + expected.size());

        for(LookupResult c : expected) {

          System.out.println("    " + c);

        }

      }

      final List<LookupResult> actual = suggest.lookup(frag, false, NUM);

      if (VERBOSE) {

        System.out.println("  actual: " + actual.size());

        for(LookupResult c : actual) {

          System.out.println("    " + c);

        }

      }

      Collections.sort(actual, new CompareByCostThenAlpha());

      final int limit = Math.min(expected.size(), actual.size());

      for(int ans=0;ans<limit;ans++) {

        final LookupResult c0 = expected.get(ans);

        final LookupResult c1 = actual.get(ans);

        assertEquals("expected " + c0.key +

                     " but got " + c1.key,

                     0,

                     CHARSEQUENCE_COMPARATOR.compare(c0.key, c1.key));

        assertEquals(c0.value, c1.value);

      }

      assertEquals(expected.size(), actual.size());

    }

  }

  private List<LookupResult> slowFuzzyMatch(int prefixLen, int maxEdits, boolean allowTransposition, List<Input> answers, String frag) {

    final List<LookupResult> results = new ArrayList<LookupResult>();

    final int fragLen = frag.length();

    for(Input tf : answers) {

      //System.out.println("  check s=" + tf.term.utf8ToString());

      boolean prefixMatches = true;

      for(int i=0;i<prefixLen;i++) {

        if (i == fragLen) {

          // Prefix still matches:

          break;

        }

        if (i == tf.term.length || tf.term.bytes[i] != (byte) frag.charAt(i)) {

          prefixMatches = false;

          break;

        }

      }

      //System.out.println("    prefixMatches=" + prefixMatches);

      if (prefixMatches) {

        final int len = tf.term.length;

        if (len >= fragLen-maxEdits) {

          // OK it's possible:

          //System.out.println("    possible");

          int d;

          final String s = tf.term.utf8ToString();

          if (fragLen == prefixLen) {

            d = 0;

          } else if (false && len < fragLen) {

            d = getDistance(frag, s, allowTransposition);

          } else {

            //System.out.println("    try loop");

            d = maxEdits + 1;

            //for(int ed=-maxEdits;ed<=maxEdits;ed++) {

            for(int ed=-maxEdits;ed<=maxEdits;ed++) {

              if (s.length() < fragLen - ed) {

                continue;

              }

              String check = s.substring(0, fragLen-ed);

              d = getDistance(frag, check, allowTransposition);

              //System.out.println("    sub check s=" + check + " d=" + d);

              if (d <= maxEdits) {

                break;

              }

            }

          }

          if (d <= maxEdits) {

            results.add(new LookupResult(tf.term.utf8ToString(), tf.v));

          }

        }

      }

      Collections.sort(results, new CompareByCostThenAlpha());

    }

    return results;

  }

  private static class CharSequenceComparator implements Comparator<CharSequence> {

    @Override

    public int compare(CharSequence o1, CharSequence o2) {

      final int l1 = o1.length();

      final int l2 = o2.length();

     

      final int aStop = Math.min(l1, l2);

      for (int i = 0; i < aStop; i++) {

        int diff = o1.charAt(i) - o2.charAt(i);

        if (diff != 0) {

          return diff;

        }

      }

      // One is a prefix of the other, or, they are equal:

      return l1 - l2;

    }

  }

  private static final Comparator<CharSequence> CHARSEQUENCE_COMPARATOR = new CharSequenceComparator();

    @Override

    public int compare(LookupResult a, LookupResult b) {

      if (a.value > b.value) {

        return -1;

      } else if (a.value < b.value) {

        return 1;

      } else {

        final int c = CHARSEQUENCE_COMPARATOR.compare(a.key, b.key);

        assert c != 0: "term=" + a.key;

        return c;

      }

    }

  }

  // NOTE: copied from

  // modules/suggest/src/java/org/apache/lucene/search/spell/LuceneLevenshteinDistance.java

  // and tweaked to return the edit distance not the float

  // lucene measure

  /* Finds unicode (code point) Levenstein (edit) distance

   * between two strings, including transpositions. */

  public int getDistance(String target, String other, boolean allowTransposition) {

    IntsRef targetPoints;

    IntsRef otherPoints;

    int n;

    int d[][]; // cost array

   

    // NOTE: if we cared, we could 3*m space instead of m*n space, similar to

    // what LevenshteinDistance does, except cycling thru a ring of three

    // horizontal cost arrays... but this comparator is never actually used by

    // DirectSpellChecker, its only used for merging results from multiple shards

    // in "distributed spellcheck", and its inefficient in other ways too...

    // cheaper to do this up front once

    targetPoints = toIntsRef(target);