public void testWarcReading() throws Exception {
ArchiveReader archiveReader = ArchiveReaderFactory.get(TestInfo.WARC_FILE1);
Iterator<? extends ArchiveRecord> it = archiveReader.iterator();
assertTrue("Warc should contains records", it.hasNext());
while (it.hasNext()) {
ArchiveRecord next = it.next();
System.out.println("mimetype:" + next.getHeader().getMimetype());
System.out.println("url:" + next.getHeader().getUrl());
}
}
/** Explode the archive into its constituent elements */
public void explode() throws CacheException {
int goodEntries = 0;
int badEntries = 0;
int entriesBetweenSleep = 0;
ArchiveReader arcReader = null;
logger.debug(
(storeArchive ? "Storing" : "Fetching") + " WARC file: " + origUrl + " will explode");
try {
// Wrap it in an ArchiveReader
logger.debug3("About to wrap stream");
arcReader = wrapStream(fetchUrl, arcStream);
logger.debug3("wrapStream() returns " + (arcReader == null ? "null" : "non-null"));
// Explode it
if (arcReader == null) {
throw new CacheException.ExploderException("no WarcReader for " + origUrl);
}
ArchivalUnit au = crawlFacade.getAu();
logger.debug("Exploding " + fetchUrl);
// Iterate through the elements in the WARC file, except the first
Iterator<ArchiveRecord> iter = arcReader.iterator();
// Skip first record
if (iter.hasNext()) iter.next();
while (iter.hasNext()) {
helper.pokeWDog();
// check need to pause
handlePause(++entriesBetweenSleep);
// handle each element in the archive
ArchiveRecord element = iter.next();
// Each element is a URL to be cached in our AU
ArchiveRecordHeader elementHeader = element.getHeader();
String elementUrl = elementHeader.getUrl();
String elementMimeType = elementHeader.getMimetype();
long elementLength = elementHeader.getLength();
long elementDate;
try {
elementDate = ArchiveUtils.parse14DigitDate(elementHeader.getDate()).getTime();
} catch (ParseException e) {
elementDate = 0;
}
logger.debug2("WARC url " + elementUrl + " mime " + elementMimeType);
// add check to determine if this is a url which should be cached
if (au.shouldBeCached(elementUrl) && elementUrl.startsWith("http:")) {
ArchiveEntry ae =
new ArchiveEntry(
elementUrl,
elementLength,
elementDate,
element, // ArchiveRecord extends InputStream
this,
fetchUrl);
ae.setHeaderFields(makeCIProperties(elementHeader));
long bytesStored = elementLength;
logger.debug3("ArchiveEntry: " + ae.getName() + " bytes " + bytesStored);
try {
helper.process(ae);
} catch (PluginException ex) {
throw new CacheException.ExploderException("helper.process() threw", ex);
}
if (ae.getBaseUrl() != null) {
if (ae.getRestOfUrl() != null && ae.getHeaderFields() != null) {
storeEntry(ae);
handleAddText(ae);
goodEntries++;
// this needs to use the correct depth ? how
CrawlUrlData cud = new CrawlUrlData(elementUrl, 0);
crawlFacade.addToParseQueue(cud);
crawlFacade.getCrawlerStatus().addContentBytesFetched(bytesStored);
}
} else {
badEntries++;
logger.debug2("Can't map " + elementUrl + " from " + archiveUrl);
}
}
}
} catch (IOException ex) {
throw new CacheException.ExploderException(ex);
} finally {
if (arcReader != null) {
try {
arcReader.close();
} catch (IOException ex) {
throw new CacheException.ExploderException(ex);
}
}
IOUtil.safeClose(arcStream);
}
// report failed fetches
if (badEntries != 0) {
String msg = archiveUrl + ": " + badEntries + "/" + goodEntries + " bad entries";
throw new CacheException.UnretryableException(msg);
}
}
@Override
public void map(
Text key, WritableArchiveRecord value, OutputCollector<Text, Text> output, Reporter reporter)
throws IOException {
ArchiveRecord record = value.getRecord();
ArchiveRecordHeader header = record.getHeader();
// Logging for debug info:
log.debug(
"Processing @"
+ header.getOffset()
+ "+"
+ record.available()
+ ","
+ header.getLength()
+ ": "
+ header.getUrl());
for (String h : header.getHeaderFields().keySet()) {
log.debug("ArchiveHeader: " + h + " -> " + header.getHeaderValue(h));
}
try {
MDX mdx = new MDX();
Date crawl_date = ArchiveUtils.parse14DigitISODate(header.getDate(), null);
if (crawl_date != null) {
mdx.setTs(ArchiveUtils.get14DigitDate(crawl_date));
} else {
mdx.setTs(header.getDate());
}
mdx.setUrl(header.getUrl());
mdx.setHash(header.getDigest());
// Data from WARC record:
mdx.put("source-file", key.toString());
mdx.put("content-type", header.getMimetype());
mdx.put("content-length", "" + header.getContentLength());
mdx.put("length", "" + header.getLength());
mdx.put("source-offset", "" + header.getOffset());
mdx.put("record-identifier", header.getRecordIdentifier());
for (String k : header.getHeaderFieldKeys()) {
mdx.put("HEADER-" + k, "" + header.getHeaderValue(k));
}
// check record type and look for HTTP data:
Header[] httpHeaders = null;
if (record instanceof WARCRecord) {
mdx.setRecordType("warc." + header.getHeaderValue(HEADER_KEY_TYPE));
mdx.setHash("" + header.getHeaderValue(WARCConstants.HEADER_KEY_PAYLOAD_DIGEST));
// There are not always headers! The code should check first.
String statusLine = HttpParser.readLine(record, "UTF-8");
if (statusLine != null && statusLine.startsWith("HTTP")) {
String firstLine[] = statusLine.split(" ");
if (firstLine.length > 1) {
String statusCode = firstLine[1].trim();
mdx.put("status-code", statusCode);
try {
httpHeaders = HttpParser.parseHeaders(record, "UTF-8");
} catch (ProtocolException p) {
log.error(
"ProtocolException ["
+ statusCode
+ "]: "
+ header.getHeaderValue(WARCConstants.HEADER_KEY_FILENAME)
+ "@"
+ header.getHeaderValue(WARCConstants.ABSOLUTE_OFFSET_KEY),
p);
}
} else {
log.warn("Could not parse status line: " + statusLine);
}
} else {
log.warn(
"Invalid status line: "
+ header.getHeaderValue(WARCConstants.HEADER_KEY_FILENAME)
+ "@"
+ header.getHeaderValue(WARCConstants.ABSOLUTE_OFFSET_KEY));
}
} else if (record instanceof ARCRecord) {
mdx.setRecordType("arc");
ARCRecord arcr = (ARCRecord) record;
mdx.put("status-code", "" + arcr.getStatusCode());
httpHeaders = arcr.getHttpHeaders();
} else {
mdx.setRecordType("unknown");
}
// Add in http headers
if (httpHeaders != null) {
for (Header h : httpHeaders) {
mdx.put("HTTP-" + h.getName(), h.getValue());
}
}
// URL:
String uri = header.getUrl();
if (uri != null) {
UsableURI uuri = UsableURIFactory.getInstance(uri);
// Hosts:
if ("https".contains(uuri.getScheme())) {
mdx.put("host", uuri.getAuthority());
}
} else {
mdx.put("errors", "malformed-url");
}
// Year
String date = header.getDate();
if (date != null && date.length() > 4) {
mdx.put("year", date.substring(0, 4));
} else {
mdx.put("errors", "malformed-date");
}
// And collect:
String outKey = mdx.getHash();
if (outKey == null || outKey == "" || "null".equals(outKey)) {
outKey = mdx.getRecordType() + ":" + header.getMimetype();
} else {
outKey = mdx.getRecordType() + ":" + outKey;
}
output.collect(new Text(outKey), new Text(mdx.toString()));
} catch (JSONException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
/** Explode the archive into its constituent elements */
public void explode() throws CacheException {
CachedUrl cachedUrl = null;
int goodEntries = 0;
int badEntries = 0;
int ignoredEntries = 0;
int entriesBetweenSleep = 0;
ArchiveReader arcReader = null;
logger.info(
(storeArchive ? "Storing" : "Fetching") + " WARC file: " + origUrl + " will explode");
try {
if (storeArchive) {
UrlCacher uc = au.makeUrlCacher(new UrlData(arcStream, arcProps, fetchUrl));
BitSet bs = new BitSet();
bs.set(UrlCacher.DONT_CLOSE_INPUT_STREAM_FLAG);
uc.setFetchFlags(bs);
uc.storeContent();
archiveData.resetInputStream();
arcStream = archiveData.input;
}
// Wrap it in an ArchiveReader
logger.debug3("About to wrap stream");
arcReader = wrapStream(fetchUrl, arcStream);
logger.debug3("wrapStream() returns " + (arcReader == null ? "null" : "non-null"));
// Explode it
if (arcReader == null) {
throw new CacheException.ExploderException("no WarcReader for " + origUrl);
}
ArchivalUnit au = crawlFacade.getAu();
Set stemSet = new HashSet();
logger.debug("Exploding " + fetchUrl);
// Iterate through the elements in the WARC file, except the first
Iterator i = arcReader.iterator();
// Skip first record
for (i.next(); i.hasNext(); ) {
// XXX probably not necessary
helper.pokeWDog();
if ((++entriesBetweenSleep % sleepAfter) == 0) {
long pauseTime =
CurrentConfig.getTimeIntervalParam(PARAM_RETRY_PAUSE, DEFAULT_RETRY_PAUSE);
Deadline pause = Deadline.in(pauseTime);
logger.debug3("Sleeping for " + StringUtil.timeIntervalToString(pauseTime));
while (!pause.expired()) {
try {
pause.sleep();
} catch (InterruptedException ie) {
// no action
}
}
}
ArchiveRecord element = (ArchiveRecord) i.next();
// Each element is a URL to be cached in a suitable AU
ArchiveRecordHeader elementHeader = element.getHeader();
String elementUrl = elementHeader.getUrl();
String elementMimeType = elementHeader.getMimetype();
long elementLength = elementHeader.getLength();
logger.debug2("WARC url " + elementUrl + " mime " + elementMimeType);
if (elementUrl.startsWith("http:")) {
ArchiveEntry ae =
new ArchiveEntry(
elementUrl,
elementLength,
0, // XXX need to convert getDate string to long
element, // ArchiveRecord extends InputStream
this,
fetchUrl);
ae.setHeaderFields(makeCIProperties(elementHeader));
long bytesStored = elementLength;
logger.debug3("ArchiveEntry: " + ae.getName() + " bytes " + bytesStored);
try {
helper.process(ae);
} catch (PluginException ex) {
throw new CacheException.ExploderException("helper.process() threw", ex);
}
if (ae.getBaseUrl() != null) {
if (ae.getRestOfUrl() != null && ae.getHeaderFields() != null) {
storeEntry(ae);
handleAddText(ae);
goodEntries++;
crawlFacade.getCrawlerStatus().addContentBytesFetched(bytesStored);
} else {
ignoredEntries++;
}
} else {
badEntries++;
logger.debug2("Can't map " + elementUrl + " from " + archiveUrl);
}
}
}
} catch (IOException ex) {
throw new CacheException.ExploderException(ex);
} finally {
if (arcReader != null)
try {
arcReader.close();
arcReader = null;
} catch (IOException ex) {
throw new CacheException.ExploderException(ex);
}
if (cachedUrl != null) {
cachedUrl.release();
}
IOUtil.safeClose(arcStream);
}
if (badEntries == 0 && goodEntries > 0) {
// Make it look like a new crawl finished on each AU to which
// URLs were added.
for (Iterator it = touchedAus.iterator(); it.hasNext(); ) {
ArchivalUnit au = (ArchivalUnit) it.next();
logger.debug3(archiveUrl + " touching " + au.toString());
AuUtil.getDaemon(au).getNodeManager(au).newContentCrawlFinished();
}
} else {
ArchivalUnit au = crawlFacade.getAu();
String msg = archiveUrl + ": " + badEntries + "/" + goodEntries + " bad entries";
throw new CacheException.UnretryableException(msg);
}
}
