Package org.apache.uima.annotator.regex

Examples of org.apache.uima.annotator.regex.ConceptFileParser

98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
      // default initialization for number format
      this.floatNumberFormat = NumberFormat.getNumberInstance();
      this.integerNumberFormat = NumberFormat.getIntegerInstance();

      // create a concept file parser object
      ConceptFileParser parser = new ConceptFileParser_impl();

      // get configuration parameter settings
      // get parameter ConceptFiles, default is an empty array
      String[] conceptFileNames = safeGetConfigParameterStringArrayValue(
            getContext(), REGEX_CONCEPTS_FILES, new String[] {});

      // get UIMA datapath and tokenize it into its elements
      StringTokenizer tokenizer = new StringTokenizer(getContext()
            .getDataPath(), PATH_SEPARATOR);
      ArrayList<File> datapathElements = new ArrayList<File>();
      while (tokenizer.hasMoreTokens()) {
         // add datapath elements to the 'datapathElements' array list
         datapathElements.add(new File(tokenizer.nextToken()));
      }

      // try to resolve the concept file names
      ArrayList<Concept> concepts = new ArrayList<Concept>();
      for (int i = 0; i < conceptFileNames.length; i++) {
         // try to resolve the relative file name with classpath or datapath
         ConceptFile file = resolveRelativeFilePath(conceptFileNames[i],
               datapathElements);

         // if the current concept file wasn't found, throw an exception
         if (file == null) {
            throw new RegexAnnotatorConfigException(
                  "regex_annotator_resource_not_found",
                  new Object[] { conceptFileNames[i] });
         } else {
            // log concept file path
            this.logger.logrb(Level.CONFIG, "RegExAnnotator", "initialize",
                  MESSAGE_DIGEST, "regex_annotator_rule_set_file",
                  new Object[] { file.getFilePath() });

            // parse concept file to internal objects
            Concept[] currentConcepts = parser.parseConceptFile(file
                  .getFilePath(), file.getStream());
            // add all concepts to the concepts list
            for (int c = 0; c < currentConcepts.length; c++) {
               concepts.add(currentConcepts[c]);
            }
View Full Code Here
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
    // default initialization for number format
    this.floatNumberFormat = NumberFormat.getNumberInstance();
    this.integerNumberFormat = NumberFormat.getIntegerInstance();

    // create a concept file parser object
    ConceptFileParser parser = new ConceptFileParser_impl();

    // get configuration parameter settings
    // get parameter ConceptFiles, default is an empty array
    String[] conceptFileNames = safeGetConfigParameterStringArrayValue(getContext(),
        REGEX_CONCEPTS_FILES, new String[] {});

    // get UIMA datapath and tokenize it into its elements
    StringTokenizer tokenizer = new StringTokenizer(getContext().getDataPath(), PATH_SEPARATOR);
    ArrayList<File> datapathElements = new ArrayList<File>();
    while (tokenizer.hasMoreTokens()) {
      // add datapath elements to the 'datapathElements' array list
      datapathElements.add(new File(tokenizer.nextToken()));
    }

    // try to resolve the concept file names
    ArrayList<Concept> concepts = new ArrayList<Concept>();
    for (int i = 0; i < conceptFileNames.length; i++) {
      // try to resolve the relative file name with classpath or datapath
      String filename = conceptFileNames[i];
      List<ConceptFile> cfList = new ArrayList<ConceptFile>();
      if (containsWildcardChar(filename)) {
        resolveRelativeWildcardFilePath(filename, datapathElements, cfList);
      } else {
        ConceptFile file = resolveRelativeFilePath(filename, datapathElements);
        // if the current concept file wasn't found, throw an exception
        if (file == null) {
          throw new RegexAnnotatorConfigException("regex_annotator_resource_not_found",
              new Object[] { conceptFileNames[i] });
        }
        cfList.add(file);
        // log concept file path
        this.logger.logrb(Level.CONFIG, "RegExAnnotator", "initialize", MESSAGE_DIGEST,
            "regex_annotator_rule_set_file", new Object[] { file.getFilePath() });

      }

      for (ConceptFile file : cfList) {
        // parse concept file to internal objects
        Concept[] currentConcepts = parser.parseConceptFile(file.getFilePath(), file.getStream());
        try {
          file.getStream().close();
        } catch (IOException e) {
          this.logger.logrb(Level.WARNING, "RegExAnnotator", "initialize", MESSAGE_DIGEST,
              "regex_annotator_error_closing_input_stream", new Object[] { file.getFilePath(),
View Full Code Here
TOP

Related Classes of org.apache.uima.annotator.regex.ConceptFileParser

  • org.apache.uima.annotator.regex.impl.RegExAnnotator
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.