Examples with AnalyzedTokenReadings org.languagetool.AnalyzedTokenReadings used on opensource projects

Example 81 with AnalyzedTokenReadings

use of org.languagetool.AnalyzedTokenReadings in project languagetool by languagetool-org.

the class MatchState method filterReadings.

public final AnalyzedTokenReadings filterReadings() {
    List<AnalyzedToken> l = new ArrayList<>();
    if (formattedToken != null) {
        if (match.isStaticLemma()) {
            matchedToken.leaveReading(new AnalyzedToken(matchedToken.getToken(), match.getPosTag(), formattedToken.getToken()));
            formattedToken = matchedToken;
        }
        String token = formattedToken.getToken();
        Pattern regexMatch = match.getRegexMatch();
        String regexReplace = match.getRegexReplace();
        if (regexMatch != null && regexReplace != null) {
            /* only replace if it is something to replace */
            token = regexMatch.matcher(token).replaceAll(regexReplace);
        }
        token = convertCase(token, token, null);
        String posTag = match.getPosTag();
        if (posTag != null) {
            int numRead = formattedToken.getReadingsLength();
            if (match.isPostagRegexp()) {
                Pattern pPosRegexMatch = match.getPosRegexMatch();
                String posTagReplace = match.getPosTagReplace();
                String targetPosTag;
                for (int i = 0; i < numRead; i++) {
                    String testTag = formattedToken.getAnalyzedToken(i).getPOSTag();
                    if (testTag != null && pPosRegexMatch.matcher(testTag).matches()) {
                        targetPosTag = testTag;
                        if (posTagReplace != null) {
                            targetPosTag = pPosRegexMatch.matcher(targetPosTag).replaceAll(posTagReplace);
                        }
                        l.add(new AnalyzedToken(token, targetPosTag, formattedToken.getAnalyzedToken(i).getLemma()));
                        l.get(l.size() - 1).setWhitespaceBefore(formattedToken.isWhitespaceBefore());
                    }
                }
                if (l.isEmpty()) {
                    l.addAll(getNewToken(numRead, token));
                }
            } else {
                l.addAll(getNewToken(numRead, token));
            }
            String lemma = formattedToken.getAnalyzedToken(0).getLemma();
            if (formattedToken.isSentenceEnd()) {
                l.add(new AnalyzedToken(formattedToken.getToken(), SENTENCE_END_TAGNAME, lemma));
            }
            if (formattedToken.isParagraphEnd()) {
                l.add(new AnalyzedToken(formattedToken.getToken(), PARAGRAPH_END_TAGNAME, lemma));
            }
        }
    }
    if (l.isEmpty()) {
        return formattedToken;
    }
    final AnalyzedTokenReadings anTkRead = new AnalyzedTokenReadings(l.toArray(new AnalyzedToken[l.size()]), formattedToken.getStartPos());
    anTkRead.setWhitespaceBefore(formattedToken.isWhitespaceBefore());
    if (!formattedToken.getChunkTags().isEmpty()) {
        anTkRead.setChunkTags(formattedToken.getChunkTags());
    }
    if (formattedToken.isImmunized()) {
        anTkRead.immunize();
    }
    return anTkRead;
}

Example 82 with AnalyzedTokenReadings

use of org.languagetool.AnalyzedTokenReadings in project languagetool by languagetool-org.

the class AdvancedWordRepeatRule method match.

/*
   * Tests if any word form is repeated in the sentence.
   */
@Override
public final RuleMatch[] match(AnalyzedSentence sentence) {
    List<RuleMatch> ruleMatches = new ArrayList<>();
    AnalyzedTokenReadings[] tokens = sentence.getTokensWithoutWhitespace();
    boolean repetition = false;
    Set<String> inflectedWords = new TreeSet<>();
    String prevLemma;
    int curToken = 0;
    // start from real token, 0 = SENT_START
    for (int i = 1; i < tokens.length; i++) {
        String token = tokens[i].getToken();
        // avoid "..." etc. to be matched:
        boolean isWord = true;
        boolean hasLemma = true;
        if (token.length() < 2) {
            isWord = false;
        }
        for (AnalyzedToken analyzedToken : tokens[i]) {
            String posTag = analyzedToken.getPOSTag();
            if (posTag != null) {
                if (StringTools.isEmpty(posTag)) {
                    isWord = false;
                    break;
                }
                String lemma = analyzedToken.getLemma();
                if (lemma == null) {
                    hasLemma = false;
                    break;
                }
                if (getExcludedWordsPattern().contains(lemma)) {
                    isWord = false;
                    break;
                }
                Matcher m2 = getExcludedPos().matcher(posTag);
                if (m2.matches()) {
                    isWord = false;
                    break;
                }
            } else {
                hasLemma = false;
            }
        }
        Matcher m1 = getExcludedNonWordsPattern().matcher(tokens[i].getToken());
        if (isWord && m1.matches()) {
            isWord = false;
        }
        prevLemma = "";
        if (isWord) {
            boolean notSentEnd = false;
            for (AnalyzedToken analyzedToken : tokens[i]) {
                String pos = analyzedToken.getPOSTag();
                if (pos != null) {
                    notSentEnd |= JLanguageTool.SENTENCE_END_TAGNAME.equals(pos);
                }
                if (hasLemma) {
                    String curLemma = analyzedToken.getLemma();
                    if (!prevLemma.equals(curLemma) && !notSentEnd) {
                        if (inflectedWords.contains(curLemma) && curToken != i) {
                            repetition = true;
                        } else {
                            inflectedWords.add(analyzedToken.getLemma());
                            curToken = i;
                        }
                    }
                    prevLemma = curLemma;
                } else {
                    if (inflectedWords.contains(tokens[i].getToken()) && !notSentEnd) {
                        repetition = true;
                    } else {
                        inflectedWords.add(tokens[i].getToken());
                    }
                }
            }
        }
        if (repetition) {
            int pos = tokens[i].getStartPos();
            RuleMatch ruleMatch = new RuleMatch(this, pos, pos + token.length(), getMessage(), getShortMessage());
            ruleMatches.add(ruleMatch);
            repetition = false;
        }
    }
    return toRuleMatchArray(ruleMatches);
}

Example 83 with AnalyzedTokenReadings

use of org.languagetool.AnalyzedTokenReadings in project languagetool by languagetool-org.

the class DemoRule method match.

// This is the method with the error detection logic that you need to implement:
@Override
public RuleMatch[] match(AnalyzedSentence sentence) throws IOException {
    List<RuleMatch> ruleMatches = new ArrayList<>();
    // Let's get all the tokens (i.e. words) of this sentence, but not the spaces:
    AnalyzedTokenReadings[] tokens = sentence.getTokensWithoutWhitespace();
    // be a special token that indicates the start of a sentence:
    for (AnalyzedTokenReadings token : tokens) {
        // the original word from the input text
        System.out.println("Token: " + token.getToken());
        // so we iterate over the readings:
        for (AnalyzedToken analyzedToken : token.getReadings()) {
            System.out.println("  Lemma: " + analyzedToken.getLemma());
            System.out.println("  POS: " + analyzedToken.getPOSTag());
        }
        // then show to the user:
        if (token.getToken().equals("demo")) {
            RuleMatch ruleMatch = new RuleMatch(this, token.getStartPos(), token.getEndPos(), "The demo rule thinks this looks wrong");
            // the user will see this as a suggested correction
            ruleMatch.setSuggestedReplacement("blablah");
            ruleMatches.add(ruleMatch);
        }
    }
    return toRuleMatchArray(ruleMatches);
}

Example 84 with AnalyzedTokenReadings

use of org.languagetool.AnalyzedTokenReadings in project languagetool by languagetool-org.

the class PatternRuleTest method testBadSentences.

private void testBadSentences(JLanguageTool languageTool, JLanguageTool allRulesLanguageTool, Language lang, Map<String, AbstractPatternRule> complexRules, AbstractPatternRule rule) throws IOException {
    List<IncorrectExample> badSentences = rule.getIncorrectExamples();
    if (badSentences.size() == 0) {
        fail("No incorrect examples found for rule " + rule.getFullId());
    }
    // necessary for XML Pattern rules containing <or>
    List<AbstractPatternRule> rules = allRulesLanguageTool.getPatternRulesByIdAndSubId(rule.getId(), rule.getSubId());
    for (IncorrectExample origBadExample : badSentences) {
        // enable indentation use
        String origBadSentence = origBadExample.getExample().replaceAll("[\\n\\t]+", "");
        List<String> expectedCorrections = origBadExample.getCorrections();
        int expectedMatchStart = origBadSentence.indexOf("<marker>");
        int expectedMatchEnd = origBadSentence.indexOf("</marker>") - "<marker>".length();
        if (expectedMatchStart == -1 || expectedMatchEnd == -1) {
            fail(lang + ": No error position markup ('<marker>...</marker>') in bad example in rule " + rule.getFullId());
        }
        String badSentence = cleanXML(origBadSentence);
        assertTrue(badSentence.trim().length() > 0);
        // necessary for XML Pattern rules containing <or>
        List<RuleMatch> matches = new ArrayList<>();
        for (Rule auxRule : rules) {
            matches.addAll(getMatches(auxRule, badSentence, languageTool));
        }
        if (rule instanceof RegexPatternRule || rule instanceof PatternRule && !((PatternRule) rule).isWithComplexPhrase()) {
            if (matches.size() != 1) {
                AnalyzedSentence analyzedSentence = languageTool.getAnalyzedSentence(badSentence);
                StringBuilder sb = new StringBuilder("Analyzed token readings:");
                for (AnalyzedTokenReadings atr : analyzedSentence.getTokens()) {
                    sb.append(" ").append(atr);
                }
                String info = "";
                if (rule instanceof RegexPatternRule) {
                    info = "\nRegexp: " + ((RegexPatternRule) rule).getPattern().toString();
                }
                fail(lang + " rule " + rule.getFullId() + ":\n\"" + badSentence + "\"\n" + "Errors expected: 1\n" + "Errors found   : " + matches.size() + "\n" + "Message: " + rule.getMessage() + "\n" + sb + "\nMatches: " + matches + info);
            }
            assertEquals(lang + ": Incorrect match position markup (start) for rule " + rule.getFullId() + ", sentence: " + badSentence, expectedMatchStart, matches.get(0).getFromPos());
            assertEquals(lang + ": Incorrect match position markup (end) for rule " + rule.getFullId() + ", sentence: " + badSentence, expectedMatchEnd, matches.get(0).getToPos());
            // make sure suggestion is what we expect it to be
            assertSuggestions(badSentence, lang, expectedCorrections, rule, matches);
            // make sure the suggested correction doesn't produce an error:
            if (matches.get(0).getSuggestedReplacements().size() > 0) {
                int fromPos = matches.get(0).getFromPos();
                int toPos = matches.get(0).getToPos();
                for (String replacement : matches.get(0).getSuggestedReplacements()) {
                    String fixedSentence = badSentence.substring(0, fromPos) + replacement + badSentence.substring(toPos);
                    matches = getMatches(rule, fixedSentence, languageTool);
                    if (matches.size() > 0) {
                        fail("Incorrect input:\n" + "  " + badSentence + "\nCorrected sentence:\n" + "  " + fixedSentence + "\nBy Rule:\n" + "  " + rule.getFullId() + "\nThe correction triggered an error itself:\n" + "  " + matches.get(0) + "\n");
                    }
                }
            }
        } else {
            // for multiple rules created with complex phrases
            matches = getMatches(rule, badSentence, languageTool);
            if (matches.size() == 0 && !complexRules.containsKey(rule.getId() + badSentence)) {
                complexRules.put(rule.getId() + badSentence, rule);
            }
            if (matches.size() != 0) {
                complexRules.put(rule.getId() + badSentence, null);
                assertTrue(lang + ": Did expect one error in: \"" + badSentence + "\" (Rule: " + rule.getFullId() + "), got " + matches.size(), matches.size() == 1);
                assertEquals(lang + ": Incorrect match position markup (start) for rule " + rule.getFullId(), expectedMatchStart, matches.get(0).getFromPos());
                assertEquals(lang + ": Incorrect match position markup (end) for rule " + rule.getFullId(), expectedMatchEnd, matches.get(0).getToPos());
                assertSuggestions(badSentence, lang, expectedCorrections, rule, matches);
                assertSuggestionsDoNotCreateErrors(badSentence, languageTool, rule, matches);
            }
        }
    // check for overlapping rules
    /*matches = getMatches(rule, badSentence, languageTool);
      List<RuleMatch> matchesAllRules = allRulesLanguageTool.check(badSentence);
      for (RuleMatch match : matchesAllRules) {
        if (!match.getRule().getId().equals(rule.getId()) && !matches.isEmpty()
            && rangeIsOverlapping(matches.get(0).getFromPos(), matches.get(0).getToPos(), match.getFromPos(), match.getToPos()))
          System.err.println("WARN: " + lang.getShortCode() + ": '" + badSentence + "' in "
                  + rule.getId() + " also matched " + match.getRule().getId());
      }*/
    }
}

Example 85 with AnalyzedTokenReadings

use of org.languagetool.AnalyzedTokenReadings in project languagetool by languagetool-org.

the class ContextBuilder method getContext.

public List<String> getContext(AnalyzedTokenReadings[] tokens, int pos, int contextSize) {
    List<String> l = new ArrayList<>();
    int i = 0;
    for (AnalyzedTokenReadings token : tokens) {
        if (i == pos) {
            l.addAll(getLeftContext(tokens, pos, contextSize));
            l.add(token.getToken());
            l.addAll(getRightContext(tokens, pos, contextSize));
            break;
        }
        i++;
    }
    return l;
}