Examples of TemporaryResources

Examples of org.apache.tika.io.TemporaryResources

88
89
90
91
92
93
94
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
            handler.characters(chars, 0, chars.length);
            handler.endElement(XHTML, "h1", "h1");
        }

        // Use the delegate parser to parse this entry
        TemporaryResources tmp = new TemporaryResources();
        try {
            final TikaInputStream newStream = TikaInputStream.get(new CloseShieldInputStream(stream), tmp);
            if (stream instanceof TikaInputStream) {
                final Object container = ((TikaInputStream) stream).getOpenContainer();

                // TODO: we can't let ZipPackage through,
                // becase of POI bug 51949.  This is less
                // efficient because the inner parser will
                // have to re-open the zip archive again.
                // Once we upgrade to POI 3.8 beta 5 we can
                // remove this:
                if ((container != null && !(container.getClass().getSimpleName().equals("ZipPackage")))) {
                    newStream.setOpenContainer(container);
                }
            }
            DELEGATING_PARSER.parse(
                                    newStream,
                                    new EmbeddedContentHandler(new BodyContentHandler(handler)),
                                    metadata, context);
        } catch (TikaException e) {
            // TODO: can we log a warning somehow?
            // Could not parse the entry, just skip the content
        } finally {
            tmp.close();
        }

        if(outputHtml) {
           handler.endElement(XHTML, "div", "div");
        }
View Full Code Here

Examples of org.apache.tika.io.TemporaryResources

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
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
      
        PDDocument pdfDocument = null;
        TemporaryResources tmp = new TemporaryResources();

        try {
            // PDFBox can process entirely in memory, or can use a temp file
            //  for unpacked / processed resources
            // Decide which to do based on if we're reading from a file or not already
            TikaInputStream tstream = TikaInputStream.cast(stream);
            if (tstream != null && tstream.hasFile()) {
               // File based, take that as a cue to use a temporary file
               RandomAccess scratchFile = new RandomAccessFile(tmp.createTemporaryFile(), "rw");
               pdfDocument = PDDocument.load(new CloseShieldInputStream(stream), scratchFile, true);
            } else {
               // Go for the normal, stream based in-memory parsing
               pdfDocument = PDDocument.load(new CloseShieldInputStream(stream), true);
            }
          
            if (pdfDocument.isEncrypted()) {
                String password = null;
               
                // Did they supply a new style Password Provider?
                PasswordProvider passwordProvider = context.get(PasswordProvider.class);
                if (passwordProvider != null) {
                   password = passwordProvider.getPassword(metadata);
                }
               
                // Fall back on the old style metadata if set
                if (password == null && metadata.get(PASSWORD) != null) {
                   password = metadata.get(PASSWORD);
                }
               
                // If no password is given, use an empty string as the default
                if (password == null) {
                   password = "";
                }
              
                try {
                    pdfDocument.decrypt(password);
                } catch (Exception e) {
                    // Ignore
                }
            }
            metadata.set(Metadata.CONTENT_TYPE, "application/pdf");
            extractMetadata(pdfDocument, metadata);
            PDF2XHTML.process(pdfDocument, handler, metadata,
                              extractAnnotationText, enableAutoSpace,
                              suppressDuplicateOverlappingText, sortByPosition);

            extractEmbeddedDocuments(context, pdfDocument, handler);
        } finally {
            if (pdfDocument != null) {
               pdfDocument.close();
            }
            tmp.dispose();
        }
    }
View Full Code Here

Examples of org.apache.tika.io.TemporaryResources

232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
    public void parse(
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        Parser parser = getParser(metadata);
        TemporaryResources tmp = new TemporaryResources();
        try {
            TikaInputStream taggedStream = TikaInputStream.get(stream, tmp);
            TaggedContentHandler taggedHandler = new TaggedContentHandler(handler);
            try {
                parser.parse(taggedStream, taggedHandler, metadata, context);
            } catch (RuntimeException e) {
                throw new TikaException(
                        "Unexpected RuntimeException from " + parser, e);
            } catch (IOException e) {
                taggedStream.throwIfCauseOf(e);
                throw new TikaException(
                        "TIKA-198: Illegal IOException from " + parser, e);
            } catch (SAXException e) {
                taggedHandler.throwIfCauseOf(e);
                throw new TikaException(
                        "TIKA-237: Illegal SAXException from " + parser, e);
            }
        } finally {
            tmp.dispose();
        }
    }
View Full Code Here

Examples of org.apache.tika.io.TemporaryResources

62
63
64
65
66
67
68
69
70
71
72
73
74
    public void parse(
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        TemporaryResources tmp = new TemporaryResources();
        try {
            TikaInputStream tis = TikaInputStream.get(stream, tmp);
            parse(tis, handler, metadata, context);
        } finally {
            tmp.dispose();
        }
    }
View Full Code Here

Examples of org.apache.tika.io.TemporaryResources

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
    public void parse(
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        TemporaryResources tmp = new TemporaryResources();
        try {
            TikaInputStream tis = TikaInputStream.get(stream, tmp);

            // Automatically detect the MIME type of the document
            MediaType type = detector.detect(tis, metadata);
            metadata.set(Metadata.CONTENT_TYPE, type.toString());

            // TIKA-216: Zip bomb prevention
            SecureContentHandler sch = new SecureContentHandler(handler, tis);
            try {
                // Parse the document
                super.parse(tis, sch, metadata, context);
            } catch (SAXException e) {
                // Convert zip bomb exceptions to TikaExceptions
                sch.throwIfCauseOf(e);
                throw e;
            }
        } finally {
            tmp.dispose();
        }
    }
View Full Code Here

Examples of org.apache.tika.io.TemporaryResources

136
137
138
139
140
141
142
143
144
145
146
147
148
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        XHTMLContentHandler xhtml =
            new XHTMLContentHandler(handler, metadata);

        TemporaryResources tmp = new TemporaryResources();
        try {
            parse(TikaInputStream.get(stream, tmp),
                    xhtml, metadata, tmp);
        } finally {
            tmp.dispose();
        }
    }
View Full Code Here

Examples of org.apache.tika.io.TemporaryResources

48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
    public void parse(
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        TemporaryResources tmp = new TemporaryResources();
        try {
            TikaInputStream tis = TikaInputStream.get(stream, tmp);
            new ImageMetadataExtractor(metadata).parseJpeg(tis.getFile());
            new JempboxExtractor(metadata).parse(tis);
        } finally {
            tmp.dispose();
        }

        XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
        xhtml.startDocument();
        xhtml.endDocument();
View Full Code Here

Examples of org.apache.tika.io.TemporaryResources

92
93
94
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
145
146
147
148
149
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
      
        PDDocument pdfDocument = null;
        TemporaryResources tmp = new TemporaryResources();

        try {
            // PDFBox can process entirely in memory, or can use a temp file
            //  for unpacked / processed resources
            // Decide which to do based on if we're reading from a file or not already
            TikaInputStream tstream = TikaInputStream.cast(stream);
            if (tstream != null && tstream.hasFile()) {
               // File based, take that as a cue to use a temporary file
               RandomAccess scratchFile = new RandomAccessFile(tmp.createTemporaryFile(), "rw");
               pdfDocument = PDDocument.load(new CloseShieldInputStream(stream), scratchFile, true);
            } else {
               // Go for the normal, stream based in-memory parsing
               pdfDocument = PDDocument.load(new CloseShieldInputStream(stream), true);
            }
          
            if (pdfDocument.isEncrypted()) {
                String password = null;
               
                // Did they supply a new style Password Provider?
                PasswordProvider passwordProvider = context.get(PasswordProvider.class);
                if (passwordProvider != null) {
                   password = passwordProvider.getPassword(metadata);
                }
               
                // Fall back on the old style metadata if set
                if (password == null && metadata.get(PASSWORD) != null) {
                   password = metadata.get(PASSWORD);
                }
               
                // If no password is given, use an empty string as the default
                if (password == null) {
                   password = "";
                }
              
                try {
                    pdfDocument.decrypt(password);
                } catch (Exception e) {
                    // Ignore
                }
            }
            metadata.set(Metadata.CONTENT_TYPE, "application/pdf");
            extractMetadata(pdfDocument, metadata);
            PDF2XHTML.process(pdfDocument, handler, metadata,
                              extractAnnotationText, enableAutoSpace,
                              suppressDuplicateOverlappingText, sortByPosition);
        } finally {
            if (pdfDocument != null) {
               pdfDocument.close();
            }
            tmp.dispose();
        }
    }
View Full Code Here

Examples of org.apache.tika.io.TemporaryResources

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
        public void parse(
                InputStream stream, ContentHandler ignored,
                Metadata metadata, ParseContext context)
                throws IOException, SAXException, TikaException {
            TemporaryResources tmp = new TemporaryResources();
            try {
                TikaInputStream tis = TikaInputStream.get(stream, tmp);

                // Figure out what we have to process
                String filename = metadata.get(Metadata.RESOURCE_NAME_KEY);
                MediaType type = detector.detect(tis, metadata);

                if (extractor == null) {
                    // Let the handler process the embedded resource
                    handler.handle(filename, type, tis);
                } else {
                    // Use a temporary file to process the stream twice
                    File file = tis.getFile();

                    // Let the handler process the embedded resource
                    InputStream input = TikaInputStream.get(file);
                    try {
                        handler.handle(filename, type, input);
                    } finally {
                        input.close();
                    }

                    // Recurse
                    extractor.extract(tis, extractor, handler);
                }
            } finally {
                tmp.dispose();
            }
        }
View Full Code Here

Examples of org.apache.tika.io.TemporaryResources

88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
            handler.characters(chars, 0, chars.length);
            handler.endElement(XHTML, "h1", "h1");
        }

        // Use the delegate parser to parse this entry
        TemporaryResources tmp = new TemporaryResources();
        try {
            final TikaInputStream newStream = TikaInputStream.get(new CloseShieldInputStream(stream), tmp);
            if (stream instanceof TikaInputStream) {
                final Object container = ((TikaInputStream) stream).getOpenContainer();
                if (container != null) {
                    newStream.setOpenContainer(container);
                }
            }
            DELEGATING_PARSER.parse(
                                    newStream,
                                    new EmbeddedContentHandler(new BodyContentHandler(handler)),
                                    metadata, context);
        } catch (TikaException e) {
            // TODO: can we log a warning somehow?
            // Could not parse the entry, just skip the content
        } finally {
            tmp.close();
        }

        if(outputHtml) {
           handler.endElement(XHTML, "div", "div");
        }
View Full Code Here
TOP
Copyright © 2018 www.massapi.com. 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.