Examples with DefaultHandler org.xml.sax.helpers.DefaultHandler used on opensource projects

Example 61 with DefaultHandler

use of org.xml.sax.helpers.DefaultHandler in project tika by apache.

the class WebPParserTest method testSimple.

/*
        Two photos in test-documents (testWebp_Alpha_Lossy.webp and testWebp_Alpha_Lossless.webp)
        are in the public domain.  These files were retrieved from:
        https://github.com/drewnoakes/metadata-extractor-images/tree/master/webp
        These photos are also available here:
        https://developers.google.com/speed/webp/gallery2#webp_links
        Credits for the photo:
        "Free Stock Photo in High Resolution - Yellow Rose 3 - Flowers"
        Image Author: Jon Sullivan
     */
@Test
public void testSimple() throws Exception {
    Metadata metadata = new Metadata();
    InputStream stream = getClass().getResourceAsStream("/test-documents/testWebp_Alpha_Lossy.webp");
    parser.parse(stream, new DefaultHandler(), metadata, new ParseContext());
    assertEquals("301", metadata.get("Image Height"));
    assertEquals("400", metadata.get("Image Width"));
    assertEquals("true", metadata.get("Has Alpha"));
    assertEquals("false", metadata.get("Is Animation"));
    assertEquals("image/webp", metadata.get(Metadata.CONTENT_TYPE));
    IOUtils.closeQuietly(stream);
    metadata = new Metadata();
    stream = getClass().getResourceAsStream("/test-documents/testWebp_Alpha_Lossless.webp");
    parser.parse(stream, new DefaultHandler(), metadata, new ParseContext());
    //unfortunately, there isn't much metadata in lossless
    assertEquals("image/webp", metadata.get(Metadata.CONTENT_TYPE));
}

Example 62 with DefaultHandler

use of org.xml.sax.helpers.DefaultHandler in project tika by apache.

the class ICNSParserTest method testICNS.

/**
     * Tests a file with multiple icons and masks
     */
@Test
public void testICNS() throws Exception {
    Metadata metadata = new Metadata();
    metadata.set(Metadata.CONTENT_TYPE, "image/icns");
    metadata.set("Icons count", "2");
    metadata.set("Icons details", "16x16 (24 bpp), 32x32 (24 bpp)");
    metadata.set("Masked icon count", "2");
    metadata.set("Masked icon details", "16x16 (8 bpp), 32x32 (8 bpp)");
    InputStream stream = getClass().getResourceAsStream("/test-documents/testICNS.icns");
    parser.parse(stream, new DefaultHandler(), metadata, new ParseContext());
}

Example 63 with DefaultHandler

use of org.xml.sax.helpers.DefaultHandler in project tika by apache.

the class ImageParserTest method testBMP.

@Test
public void testBMP() throws Exception {
    Metadata metadata = new Metadata();
    metadata.set(Metadata.CONTENT_TYPE, "image/bmp");
    InputStream stream = getClass().getResourceAsStream("/test-documents/testBMP.bmp");
    parser.parse(stream, new DefaultHandler(), metadata, new ParseContext());
    assertEquals("75", metadata.get("height"));
    assertEquals("100", metadata.get("width"));
    assertEquals("8 8 8", metadata.get("Data BitsPerSample"));
    assertEquals("1.0", metadata.get("Dimension PixelAspectRatio"));
    //TODO: figure out why we're getting 0.35273367 in Ubuntu, but not Windows
    //assertEquals("0", metadata.get("Dimension VerticalPhysicalPixelSpacing"));
    //assertEquals("0", metadata.get("Dimension HorizontalPhysicalPixelSpacing"));
    assertEquals("BI_RGB", metadata.get("Compression CompressionTypeName"));
    assertEquals("image/bmp", metadata.get("Content-Type"));
    assertEquals("100", metadata.get(Metadata.IMAGE_WIDTH));
    assertEquals("75", metadata.get(Metadata.IMAGE_LENGTH));
    assertEquals("8 8 8", metadata.get(Metadata.BITS_PER_SAMPLE));
}

Example 64 with DefaultHandler

use of org.xml.sax.helpers.DefaultHandler in project tika by apache.

the class ImageParserTest method testGIF.

@Test
public void testGIF() throws Exception {
    Metadata metadata = new Metadata();
    metadata.set(Metadata.CONTENT_TYPE, "image/gif");
    InputStream stream = getClass().getResourceAsStream("/test-documents/testGIF.gif");
    parser.parse(stream, new DefaultHandler(), metadata, new ParseContext());
    assertEquals("75", metadata.get("height"));
    assertEquals("100", metadata.get("width"));
    assertEquals("true", metadata.get("Compression Lossless"));
    assertEquals("Normal", metadata.get("Dimension ImageOrientation"));
    assertEquals("lzw", metadata.get("Compression CompressionTypeName"));
    assertEquals("0", metadata.get("Dimension HorizontalPixelOffset"));
    assertEquals("imageLeftPosition=0, imageTopPosition=0, imageWidth=100, imageHeight=75, interlaceFlag=false", metadata.get("ImageDescriptor"));
    assertEquals("Index", metadata.get("Data SampleFormat"));
    assertEquals("3", metadata.get("Chroma NumChannels"));
    assertEquals("1", metadata.get("Compression NumProgressiveScans"));
    assertEquals("RGB", metadata.get("Chroma ColorSpaceType"));
    assertEquals("Licensed to the Apache Software Foundation (ASF) under one or more contributor license agreements.  See the NOTICE file distributed with this work for additional information regarding copyright ownership.", metadata.get("CommentExtensions CommentExtension"));
    assertEquals("value=Licensed to the Apache Software Foundation (ASF) under one or more contributor license agreements.  See the NOTICE file distributed with this work for additional information regarding copyright ownership., encoding=ISO-8859-1, compression=none", metadata.get("Text TextEntry"));
    assertEquals("true", metadata.get("Chroma BlackIsZero"));
    assertEquals("disposalMethod=none, userInputFlag=false, transparentColorFlag=false, delayTime=0, transparentColorIndex=0", metadata.get("GraphicControlExtension"));
    assertEquals("0", metadata.get("Dimension VerticalPixelOffset"));
    assertEquals("image/gif", metadata.get("Content-Type"));
    assertEquals("100", metadata.get(Metadata.IMAGE_WIDTH));
    assertEquals("75", metadata.get(Metadata.IMAGE_LENGTH));
    assertEquals("Licensed to the Apache Software Foundation (ASF) under one or more contributor license agreements.  See the NOTICE file distributed with this work for additional information regarding copyright ownership.", metadata.get(TikaCoreProperties.COMMENTS));
}

Example 65 with DefaultHandler

use of org.xml.sax.helpers.DefaultHandler in project tika by apache.

the class HtmlParserTest method assertScriptLink.

private void assertScriptLink(String html, String url) throws Exception {
    // IdentityHtmlMapper is needed to extract <script> tags
    ParseContext context = new ParseContext();
    context.set(HtmlMapper.class, IdentityHtmlMapper.INSTANCE);
    Metadata metadata = new Metadata();
    metadata.set(Metadata.CONTENT_TYPE, "text/html");
    final List<String> links = new ArrayList<String>();
    new HtmlParser().parse(new ByteArrayInputStream(html.getBytes(UTF_8)), new DefaultHandler() {

        @Override
        public void startElement(String u, String l, String name, Attributes atts) {
            if (name.equals("script") && atts.getValue("", "src") != null) {
                links.add(atts.getValue("", "src"));
            }
        }
    }, metadata, context);
    assertEquals(1, links.size());
    assertEquals(url, links.get(0));
}