Package com.findwise.utils.tika

Examples of com.findwise.utils.tika.InputStreamParser

70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
    public void process(LocalDocument doc) throws Exception {
        Map<String, Object> urls = FieldHelper.getFieldMatchingPattern(doc.getContentMap(),
                urlFieldPattern);
        UriParser uriParser = new UriParser();
        DocumentParserHelper documentParserHelper = new DocumentParserHelper(addMetaData, addLanguage);
        InputStreamParser inputStreamParser = new InputStreamParser(parser);
        for (String field : urls.keySet()) {
            Iterator<URL> it = uriParser.getUrlsFromObject(urls.get(field)).iterator();
            for (int i = 1; it.hasNext(); i++) {
                String num = (i > 1) ? "" + i : "";
                URL url = it.next();
                URLConnection connection = createConnection(url);
                final InputStream inputStream = connection.getInputStream();
                try {
                    String prefix = field + num + "_";
                    ParsedData parsedData = inputStreamParser.parse(inputStream);
                    documentParserHelper.addParsedDataToDocument(parsedData, doc, prefix);
                } finally {
                    inputStream.close();
                }
            }
View Full Code Here
154
155
156
157
158
159
160
161
162
163
164
        return (size <= maxSizeInBytes);
    }

    private void enrichDocumentWithFileContents(LocalDocument doc, InputStream stream) throws IOException,
            SAXException, TikaException {
        InputStreamParser inputStreamParser = new InputStreamParser();
        ParsedData parsedData = inputStreamParser.parse(stream);

        addTextToDocument(doc, parsedData.getContent());
        addMetadataToDocument(doc, parsedData.getMetadata());
    }
View Full Code Here
35
36
37
38
39
40
41
42
43
44
        List<String> files = doc.getFileNames();
        DocumentParserHelper documentParserHelper = new DocumentParserHelper(addMetaData, addLanguage);
        for (String fileName : files) {
            DocumentFile<Local> df = doc.getFile(fileName);
            String prefix = fileName.replace('.', '_') + "_";
            InputStreamParser inputStreamParser = new InputStreamParser(parser);
            ParsedData parsedData = inputStreamParser.parse(df.getStream());
            documentParserHelper.addParsedDataToDocument(parsedData, doc, prefix);
        }
    }
View Full Code Here
TOP

Related Classes of com.findwise.utils.tika.InputStreamParser

  • com.findwise.hydra.stage.tika.SimpleFetchingTikaStage
  • com.findwise.hydra.stage.tika.TextExtractionStage
  • com.findwise.hydra.stage.tika.TikaStage
  • org.apache.tika.metadata.Metadata
  • org.apache.tika.parser.AutoDetectParser
  • org.apache.tika.parser.ParseContext
  • org.apache.tika.sax.BodyContentHandler
  • java.io.StringWriter
    • 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.