/** Returns true if mediaType falls withing the given range (pattern), false otherwise */
private boolean isMediaTypeMatch(MediaType mediaType, MediaType rangePattern) {
String WILDCARD = "*";
String rangePatternType = rangePattern.getType();
String rangePatternSubtype = rangePattern.getSubtype();
return (rangePatternType.equals(WILDCARD) || rangePatternType.equals(mediaType.getType()))
&& (rangePatternSubtype.equals(WILDCARD)
|| rangePatternSubtype.equals(mediaType.getSubtype()));
}
@Override
public void parse(
InputStream stream, ContentHandler handler, Metadata metadata, ParseContext context)
throws IOException, SAXException, TikaException {
// Is it a supported image?
String filename = metadata.get(Metadata.RESOURCE_NAME_KEY);
String type = metadata.get(Metadata.CONTENT_TYPE);
boolean accept = false;
if (type != null) {
for (MediaType mt : types) {
if (mt.toString().equals(type)) {
accept = true;
}
}
}
if (filename != null) {
for (MediaType mt : types) {
String ext = "." + mt.getSubtype();
if (filename.endsWith(ext)) {
accept = true;
}
}
}
if (!accept) return;
handleImage(stream, filename, type);
}
public static String getMimeTypeFromContentType(String contentType) {
String result = "";
MediaType mt = MediaType.parse(contentType);
if (mt != null) {
result = mt.getType() + "/" + mt.getSubtype();
}
return result;
}
public static void main(String[] args) throws IOException {
DemoWebMiningOptions options = new DemoWebMiningOptions();
CmdLineParser parser = new CmdLineParser(options);
try {
parser.parseArgument(args);
} catch (CmdLineException e) {
System.err.println(e.getMessage());
printUsageAndExit(parser);
}
// Build and run the flow.
try {
Path workingDirPath = new Path(options.getWorkingDir());
JobConf conf = new JobConf();
FileSystem fs = workingDirPath.getFileSystem(conf);
setupWorkingDir(fs, workingDirPath, CrawlConfig.SEED_URLS_FILENAME);
Path latestDirPath = CrawlDirUtils.findLatestLoopDir(fs, workingDirPath);
if (latestDirPath == null) {
error("No previous cycle output dirs exist in " + workingDirPath, parser);
}
Path crawlDbPath = new Path(latestDirPath, CrawlConfig.CRAWLDB_SUBDIR_NAME);
UserAgent userAgent =
new UserAgent(options.getAgentName(), CrawlConfig.EMAIL_ADDRESS, CrawlConfig.WEB_ADDRESS);
FetcherPolicy fetcherPolicy = new FetcherPolicy();
fetcherPolicy.setCrawlDelay(CrawlConfig.DEFAULT_CRAWL_DELAY);
fetcherPolicy.setMaxContentSize(CrawlConfig.MAX_CONTENT_SIZE);
fetcherPolicy.setFetcherMode(FetcherMode.EFFICIENT);
// We only care about mime types that the Tika HTML parser can handle,
// so restrict it to the same.
Set<String> validMimeTypes = new HashSet<String>();
Set<MediaType> supportedTypes = new HtmlParser().getSupportedTypes(new ParseContext());
for (MediaType supportedType : supportedTypes) {
validMimeTypes.add(
String.format("%s/%s", supportedType.getType(), supportedType.getSubtype()));
}
fetcherPolicy.setValidMimeTypes(validMimeTypes);
// Let's limit our crawl to two loops
for (int curLoop = 1; curLoop <= 2; curLoop++) {
Path curLoopDirPath = CrawlDirUtils.makeLoopDir(fs, workingDirPath, curLoop);
Flow flow =
DemoWebMiningWorkflow.createWebMiningWorkflow(
crawlDbPath, curLoopDirPath, fetcherPolicy, userAgent, options, curLoop == 1);
flow.complete();
// Update crawlDbPath to point to the latest crawl db
crawlDbPath = new Path(curLoopDirPath, CrawlConfig.CRAWLDB_SUBDIR_NAME);
}
} catch (Exception e) {
System.err.println("Exception running job: " + e.getMessage());
e.printStackTrace(System.err);
System.exit(-1);
}
}
