Package org.apache.tika.exception

Examples of org.apache.tika.exception.TikaException

48
49
50
51
52
53
54
55
56
57
58
                new XHTMLContentHandler(handler, metadata);
            xhtml.startDocument();
            xhtml.element("p", sd.getText(0, sd.getLength()));
            xhtml.endDocument();
        } catch (BadLocationException e) {
            throw new TikaException("Error parsing an RTF document", e);
        } catch (InternalError e) {
            throw new TikaException(
                    "Internal error parsing an RTF document, see TIKA-282", e);
        }
    }
View Full Code Here
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        DataInputStream datainput = new DataInputStream(stream);
        if (!checkSignature(datainput)) {
            throw new TikaException("FLV signature not detected");
        }

        // header
        int version = datainput.readUnsignedByte();
        if (version != 1) {
            // should be 1, perhaps this is not flv?
            throw new TikaException("Unpexpected FLV version: " + version);
        }

        int typeFlags = datainput.readUnsignedByte();

        long len = readUInt32(datainput);
        if (len != 9) {
            // we only know about format with header of 9 bytes
            throw new TikaException("Unpexpected FLV header length: " + len);
        }

        long sizePrev = readUInt32(datainput);
        if (sizePrev != 0) {
            // should be 0, perhaps this is not flv?
            throw new TikaException(
                    "Unpexpected FLV first previous block size: " + sizePrev);
        }

        metadata.set(Metadata.CONTENT_TYPE, "video/x-flv");
        metadata.set("hasVideo", Boolean.toString((typeFlags & MASK_VIDEO) != 0));
View Full Code Here
51
52
53
54
55
56
57
58
59
60
61
                    metadata.set("height", Integer.toString(reader.getHeight(0)));
                    metadata.set("width", Integer.toString(reader.getWidth(0)));
                    reader.dispose();
                }
            } catch (IIOException e) {
                throw new TikaException(type + " parse error", e);
            }
        }

        XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
        xhtml.startDocument();
View Full Code Here
156
157
158
159
160
161
162
163
164
165
166
167
                parse(new DocumentSummaryInformation(properties), metadata);
            }
        } catch (FileNotFoundException e) {
            // entry does not exist, just skip it
        } catch (NoPropertySetStreamException e) {
            throw new TikaException("Not a HPSF document", e);
        } catch (UnexpectedPropertySetTypeException e) {
            throw new TikaException("Unexpected HPSF document", e);
        } catch (MarkUnsupportedException e) {
            throw new TikaException("Invalid DocumentInputStream", e);
        }
    }
View Full Code Here
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
145
146
            }
        }

        String encoding = metadata.get(Metadata.CONTENT_ENCODING);
        if (encoding == null) {
            throw new TikaException(
                    "Text encoding could not be detected and no encoding"
                    + " hint is available in document metadata");
        }

        // TIKA-341: Only stomp on content-type after we're done trying to
        // use it to guess at the charset.
        metadata.set(Metadata.CONTENT_TYPE, "text/plain");

        try {
            Reader reader =
                new BufferedReader(new InputStreamReader(stream, encoding));

            // TIKA-240: Drop the BOM when extracting plain text
            reader.mark(1);
            int bom = reader.read();
            if (bom != '\ufeff') { // zero-width no-break space
                reader.reset();
            }

            XHTMLContentHandler xhtml =
                new XHTMLContentHandler(handler, metadata);
            xhtml.startDocument();

            xhtml.startElement("p");
            char[] buffer = new char[4096];
            int n = reader.read(buffer);
            while (n != -1) {
                xhtml.characters(buffer, 0, n);
                n = reader.read(buffer);
            }
            xhtml.endElement("p");

            xhtml.endDocument();
        } catch (UnsupportedEncodingException e) {
            throw new TikaException(
                    "Unsupported text encoding: " + encoding, e);
        }
    }
View Full Code Here
64
65
66
67
68
69
70
71
            SAXParser parser = factory.newSAXParser();
            parser.parse(
                    new CloseShieldInputStream(stream),
                    new OfflineContentHandler(xhtml));
        } catch (ParserConfigurationException e) {
            throw new TikaException("XML parser configuration error", e);
        }
    }
View Full Code Here
48
49
50
51
52
53
54
55
56
57
58
59
                    (POIXMLTextExtractor) ExtractorFactory.createExtractor(stream),
                    locale);
            extractor.getMetadataExtractor().extract(metadata);
            extractor.getXHTML(handler, metadata);
        } catch (InvalidFormatException e) {
            throw new TikaException("Error creating OOXML extractor", e);
        } catch (OpenXML4JException e) {
            throw new TikaException("Error creating OOXML extractor", e);
        } catch (XmlException e) {
            throw new TikaException("Error creating OOXML extractor", e);
        }
    }
View Full Code Here
329
330
331
332
333
334
335
336
            parser.parse(
                    new CloseShieldInputStream(stream),
                    new OfflineContentHandler(
                            new NSNormalizerContentHandler(dh)));
        } catch (ParserConfigurationException e) {
            throw new TikaException("XML parser configuration error", e);
        }
    }
View Full Code Here
155
156
157
158
159
160
161
162
163
164
165
   
        detector.setText(stream);
   
        CharsetMatch match = detector.detect();
        if (match == null) {
            throw new TikaException("Unable to detect character encoding");
        }
       
        metadata.set(Metadata.CONTENT_ENCODING, match.getName());
        String language = match.getLanguage();
        if (language != null) {
View Full Code Here
130
131
132
133
134
135
136
137
138
    public static List<Parser> getParsersFromZip(File zip, TikaConfig config)
            throws TikaException, FileNotFoundException {
        String zipMimeType = config.getMimeRepository().getMimeType(zip)
        .getName();
        if (!zipMimeType.equalsIgnoreCase("application/zip")) {
            throw new TikaException("The file you are using is note a zip file");
        }
        return getParsersFromZip(new FileInputStream(zip), config);
    }
View Full Code Here
TOP

Related Classes of org.apache.tika.exception.TikaException

  • de.innosystec.unrar.tika.RARParser
  • org.apache.jackrabbit.core.query.lucene.JackrabbitParser
  • org.apache.jackrabbit.core.query.pdf.PDF2XHTML
  • org.apache.nutch.parse.tika.TikaConfig
  • org.apache.tika.config.TikaConfig
  • org.apache.tika.detect.AutoDetectReader
  • org.apache.tika.embedder.ExternalEmbedder
  • org.apache.tika.extractor.ParserContainerExtractor
  • org.apache.tika.fork.ForkClient
  • org.apache.tika.fork.ForkParser
    • 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.