@Override
public String push(final Request entry, CrawlProfile profile, final RobotsTxt robots)
throws IOException, SpaceExceededException {
assert entry != null;
final byte[] hash = entry.url().hash();
synchronized (this) {
// double-check
if (this.has(hash)) return "double occurrence in urlFileIndex";
// increase dom counter
if (profile != null) {
int maxPages = profile.domMaxPages();
if (maxPages != Integer.MAX_VALUE && maxPages > 0) {
String host = entry.url().getHost();
profile.domInc(host);
}
}
// add to index
Index depthStack = getStack(entry.depth());
final int s = depthStack.size();
depthStack.put(entry.toRow());
assert s < depthStack.size()
: "hash = " + ASCII.String(hash) + ", s = " + s + ", size = " + depthStack.size();
assert depthStack.has(hash) : "hash = " + ASCII.String(hash);
}
return null;
}
public static serverObjects respond(
final RequestHeader header, final serverObjects post, final serverSwitch env) {
// return variable that accumulates replacements
final Switchboard sb = (Switchboard) env;
// clean up all search events
SearchEventCache.cleanupEvents(true);
sb.index.clearCaches(); // every time the ranking is changed we need to remove old orderings
// inital values for AJAX Elements (without JavaScript)
final serverObjects prop = new serverObjects();
prop.put("rejected", 0);
Segment segment = sb.index;
Fulltext fulltext = segment.fulltext();
String localSolr = "/solr/select?core=collection1&q=*:*&start=0&rows=3";
String remoteSolr =
env.getConfig(SwitchboardConstants.FEDERATED_SERVICE_SOLR_INDEXING_URL, localSolr);
if (!remoteSolr.endsWith("/")) remoteSolr = remoteSolr + "/";
prop.put(
"urlpublictextSolrURL",
fulltext.connectedLocalSolr()
? localSolr
: remoteSolr + "collection1/select?&q=*:*&start=0&rows=3");
prop.putNum("urlpublictextSize", fulltext.collectionSize());
prop.putNum("urlpublictextSegmentCount", fulltext.getDefaultConnector().getSegmentCount());
prop.put(
"webgraphSolrURL",
fulltext.connectedLocalSolr()
? localSolr.replace("collection1", "webgraph")
: remoteSolr + "webgraph/select?&q=*:*&start=0&rows=3");
prop.putNum("webgraphSize", fulltext.useWebgraph() ? fulltext.webgraphSize() : 0);
prop.putNum(
"webgraphSegmentCount",
fulltext.useWebgraph() ? fulltext.getWebgraphConnector().getSegmentCount() : 0);
prop.putNum("citationSize", segment.citationCount());
prop.putNum("citationSegmentCount", segment.citationSegmentCount());
prop.putNum("rwipublictextSize", segment.RWICount());
prop.putNum("rwipublictextSegmentCount", segment.RWISegmentCount());
prop.put("list", "0");
prop.put("loaderSize", 0);
prop.put("loaderMax", 0);
prop.put("list-loader", 0);
int coreCrawlJobSize = sb.crawlQueues.coreCrawlJobSize();
int limitCrawlJobSize = sb.crawlQueues.limitCrawlJobSize();
int remoteTriggeredCrawlJobSize = sb.crawlQueues.remoteTriggeredCrawlJobSize();
int noloadCrawlJobSize = sb.crawlQueues.noloadCrawlJobSize();
int allsize =
coreCrawlJobSize + limitCrawlJobSize + remoteTriggeredCrawlJobSize + noloadCrawlJobSize;
prop.put("localCrawlSize", coreCrawlJobSize);
prop.put("localCrawlState", "");
prop.put("limitCrawlSize", limitCrawlJobSize);
prop.put("limitCrawlState", "");
prop.put("remoteCrawlSize", remoteTriggeredCrawlJobSize);
prop.put("remoteCrawlState", "");
prop.put("noloadCrawlSize", noloadCrawlJobSize);
prop.put("noloadCrawlState", "");
prop.put("terminate-button", allsize == 0 ? 0 : 1);
prop.put("list-remote", 0);
prop.put("forwardToCrawlStart", "0");
prop.put("info", "0");
boolean debug = (post != null && post.containsKey("debug"));
if (post != null) {
String c = post.toString();
if (c.length() < 1000) ConcurrentLog.info("Crawl Start", c);
}
if (post != null && post.containsKey("queues_terminate_all")) {
// terminate crawls individually
sb.crawlQueues.noticeURL.clear();
for (final byte[] h : sb.crawler.getActive()) {
CrawlProfile p = sb.crawler.getActive(h);
if (CrawlSwitchboard.DEFAULT_PROFILES.contains(p.name())) continue;
if (p != null) sb.crawler.putPassive(h, p);
sb.crawler.removeActive(h);
sb.crawler.removePassive(h);
try {
sb.crawlQueues.noticeURL.removeByProfileHandle(p.handle(), 10000);
} catch (SpaceExceededException e) {
}
}
// clear stacks
for (StackType stackType : StackType.values()) sb.crawlQueues.noticeURL.clear(stackType);
try {
sb.cleanProfiles();
} catch (final InterruptedException e) {
/* ignore this */
}
// remove pause
sb.continueCrawlJob(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL);
sb.setConfig(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL + "_isPaused_cause", "");
sb.continueCrawlJob(SwitchboardConstants.CRAWLJOB_REMOTE_TRIGGERED_CRAWL);
sb.setConfig(SwitchboardConstants.CRAWLJOB_REMOTE_TRIGGERED_CRAWL + "_isPaused_cause", "");
prop.put("terminate-button", 0);
}
if (post != null && post.containsKey("continue")) {
// continue queue
final String queue = post.get("continue", "");
if ("localcrawler".equals(queue)) {
sb.continueCrawlJob(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL);
sb.setConfig(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL + "_isPaused_cause", "");
} else if ("remotecrawler".equals(queue)) {
sb.continueCrawlJob(SwitchboardConstants.CRAWLJOB_REMOTE_TRIGGERED_CRAWL);
sb.setConfig(SwitchboardConstants.CRAWLJOB_REMOTE_TRIGGERED_CRAWL + "_isPaused_cause", "");
}
}
if (post != null && post.containsKey("pause")) {
// pause queue
final String queue = post.get("pause", "");
if ("localcrawler".equals(queue)) {
sb.pauseCrawlJob(
SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL,
"user request in Crawler_p from " + header.refererHost());
} else if ("remotecrawler".equals(queue)) {
sb.pauseCrawlJob(
SwitchboardConstants.CRAWLJOB_REMOTE_TRIGGERED_CRAWL,
"user request in Crawler_p from " + header.refererHost());
}
}
String queuemessage =
sb.getConfig(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL + "_isPaused_cause", "");
if (queuemessage.length() == 0) {
prop.put("info-queue", 0);
} else {
prop.put("info-queue", 1);
prop.putHTML("info-queue_message", "pause reason: " + queuemessage);
}
if (post != null && post.containsKey("terminate"))
try {
final String handle = post.get("handle", "");
// termination of a crawl: shift the crawl from active to passive
final CrawlProfile p = sb.crawler.getActive(handle.getBytes());
if (p != null) sb.crawler.putPassive(handle.getBytes(), p);
// delete all entries from the crawl queue that are deleted here
sb.crawler.removeActive(handle.getBytes());
sb.crawler.removePassive(handle.getBytes());
sb.crawlQueues.noticeURL.removeByProfileHandle(handle, 10000);
} catch (final SpaceExceededException e) {
ConcurrentLog.logException(e);
}
if (post != null && post.containsKey("crawlingstart")) {
// init crawl
if (sb.peers == null) {
prop.put("info", "3");
} else {
// remove crawlingFileContent before we record the call
String crawlingFileName = post.get("crawlingFile");
final File crawlingFile;
if (crawlingFileName == null || crawlingFileName.isEmpty()) {
crawlingFile = null;
} else {
if (crawlingFileName.startsWith("file://"))
crawlingFileName = crawlingFileName.substring(7);
crawlingFile = new File(crawlingFileName);
}
if (crawlingFile != null && crawlingFile.exists()) {
post.remove("crawlingFile$file");
}
// prepare some filter that are adjusted in case that this is wanted
boolean storeHTCache = "on".equals(post.get("storeHTCache", "off"));
String newcrawlingMustMatch = post.get("mustmatch", CrawlProfile.MATCH_ALL_STRING);
String newcrawlingMustNotMatch = post.get("mustnotmatch", CrawlProfile.MATCH_NEVER_STRING);
if (newcrawlingMustMatch.length() < 2)
newcrawlingMustMatch =
CrawlProfile
.MATCH_ALL_STRING; // avoid that all urls are filtered out if bad value was
// submitted
boolean fullDomain =
"domain".equals(post.get("range", "wide")); // special property in simple crawl start
boolean subPath =
"subpath".equals(post.get("range", "wide")); // special property in simple crawl start
final boolean restrictedcrawl =
fullDomain || subPath || !CrawlProfile.MATCH_ALL_STRING.equals(newcrawlingMustMatch);
final boolean deleteage = restrictedcrawl && "age".equals(post.get("deleteold", "off"));
Date deleteageDate = null;
if (deleteage) {
deleteageDate =
timeParser(
true,
post.getInt("deleteIfOlderNumber", -1),
post.get("deleteIfOlderUnit", "year")); // year, month, day, hour
}
final boolean deleteold =
(deleteage && deleteageDate != null)
|| (restrictedcrawl && post.getBoolean("deleteold"));
final String sitemapURLStr = post.get("sitemapURL", "");
String crawlingStart0 = post.get("crawlingURL", "").trim(); // the crawljob start url
String[] rootURLs0 =
crawlingStart0.indexOf('\n') > 0 || crawlingStart0.indexOf('\r') > 0
? crawlingStart0.split("[\\r\\n]+")
: crawlingStart0.split(Pattern.quote("|"));
Set<DigestURL> rootURLs = new HashSet<DigestURL>();
String crawlName = "";
if (crawlingFile == null)
for (String crawlingStart : rootURLs0) {
if (crawlingStart == null || crawlingStart.length() == 0) continue;
// add the prefix http:// if necessary
int pos = crawlingStart.indexOf("://", 0);
if (pos == -1) {
if (crawlingStart.startsWith("ftp")) crawlingStart = "ftp://" + crawlingStart;
else crawlingStart = "http://" + crawlingStart;
}
try {
DigestURL crawlingStartURL = new DigestURL(crawlingStart);
rootURLs.add(crawlingStartURL);
crawlName +=
((crawlingStartURL.getHost() == null)
? crawlingStartURL.toNormalform(true)
: crawlingStartURL.getHost())
+ ',';
if (crawlingStartURL != null
&& (crawlingStartURL.isFile() || crawlingStartURL.isSMB())) storeHTCache = false;
} catch (final MalformedURLException e) {
ConcurrentLog.logException(e);
}
}
else {
crawlName = crawlingFile.getName();
}
if (crawlName.endsWith(",")) crawlName = crawlName.substring(0, crawlName.length() - 1);
if (crawlName.length() > 64) {
crawlName =
"crawl_for_"
+ rootURLs.size()
+ "_start_points_"
+ Integer.toHexString(crawlName.hashCode());
int p = crawlName.lastIndexOf(',');
if (p >= 8) crawlName = crawlName.substring(0, p);
}
if (crawlName.length() == 0 && sitemapURLStr.length() > 0)
crawlName = "sitemap loader for " + sitemapURLStr;
// in case that a root url has a file protocol, then the site filter does not work, patch
// that:
if (fullDomain) {
for (DigestURL u : rootURLs)
if (u.isFile()) {
fullDomain = false;
subPath = true;
break;
}
}
// delete old robots entries
for (DigestURL ru : rootURLs) {
sb.robots.delete(ru);
try {
if (ru.getHost() != null) { // might be null for file://
Cache.delete(RobotsTxt.robotsURL(RobotsTxt.getHostPort(ru)).hash());
}
} catch (IOException e) {
}
}
try {
sb.robots.clear();
} catch (IOException e) {
} // to be safe: clear all.
// set the crawl filter
String ipMustMatch = post.get("ipMustmatch", CrawlProfile.MATCH_ALL_STRING);
final String ipMustNotMatch = post.get("ipMustnotmatch", CrawlProfile.MATCH_NEVER_STRING);
if (ipMustMatch.length() < 2) ipMustMatch = CrawlProfile.MATCH_ALL_STRING;
final String countryMustMatch =
post.getBoolean("countryMustMatchSwitch") ? post.get("countryMustMatchList", "") : "";
sb.setConfig("crawlingIPMustMatch", ipMustMatch);
sb.setConfig("crawlingIPMustNotMatch", ipMustNotMatch);
if (countryMustMatch.length() > 0)
sb.setConfig("crawlingCountryMustMatch", countryMustMatch);
String crawlerNoDepthLimitMatch =
post.get("crawlingDepthExtension", CrawlProfile.MATCH_NEVER_STRING);
final String indexUrlMustMatch = post.get("indexmustmatch", CrawlProfile.MATCH_ALL_STRING);
final String indexUrlMustNotMatch =
post.get("indexmustnotmatch", CrawlProfile.MATCH_NEVER_STRING);
final String indexContentMustMatch =
post.get("indexcontentmustmatch", CrawlProfile.MATCH_ALL_STRING);
final String indexContentMustNotMatch =
post.get("indexcontentmustnotmatch", CrawlProfile.MATCH_NEVER_STRING);
final boolean crawlOrder = post.get("crawlOrder", "off").equals("on");
env.setConfig("crawlOrder", crawlOrder);
if (crawlOrder)
crawlerNoDepthLimitMatch =
CrawlProfile.MATCH_NEVER_STRING; // without limitation the crawl order does not work
int newcrawlingdepth = post.getInt("crawlingDepth", 8);
env.setConfig("crawlingDepth", Integer.toString(newcrawlingdepth));
if ((crawlOrder) && (newcrawlingdepth > 8)) newcrawlingdepth = 8;
boolean directDocByURL =
"on"
.equals(
post.get(
"directDocByURL",
"off")); // catch also all linked media documents without loading them
env.setConfig("crawlingDirectDocByURL", directDocByURL);
final String collection = post.get("collection", "user");
env.setConfig("collection", collection);
// recrawl
final String recrawl = post.get("recrawl", "nodoubles"); // nodoubles, reload, scheduler
Date crawlingIfOlder = null;
if ("reload".equals(recrawl)) {
crawlingIfOlder =
timeParser(
true,
post.getInt("reloadIfOlderNumber", -1),
post.get("reloadIfOlderUnit", "year")); // year, month, day, hour
}
env.setConfig(
"crawlingIfOlder",
crawlingIfOlder == null ? Long.MAX_VALUE : crawlingIfOlder.getTime());
// store this call as api call
sb.tables.recordAPICall(
post,
"Crawler_p.html",
WorkTables.TABLE_API_TYPE_CRAWLER,
"crawl start for "
+ ((rootURLs.size() == 0)
? post.get("crawlingFile", "")
: rootURLs.iterator().next().toNormalform(true)));
final boolean crawlingDomMaxCheck = "on".equals(post.get("crawlingDomMaxCheck", "off"));
final int crawlingDomMaxPages =
(crawlingDomMaxCheck) ? post.getInt("crawlingDomMaxPages", -1) : -1;
env.setConfig("crawlingDomMaxPages", Integer.toString(crawlingDomMaxPages));
boolean followFrames = "on".equals(post.get("followFrames", "false"));
env.setConfig("followFrames", followFrames);
boolean obeyHtmlRobotsNoindex = "on".equals(post.get("obeyHtmlRobotsNoindex", "false"));
env.setConfig("obeyHtmlRobotsNoindex", obeyHtmlRobotsNoindex);
boolean obeyHtmlRobotsNofollow = "on".equals(post.get("obeyHtmlRobotsNofollow", "false"));
env.setConfig("obeyHtmlRobotsNofollow", obeyHtmlRobotsNofollow);
final boolean indexText = "on".equals(post.get("indexText", "false"));
env.setConfig("indexText", indexText);
final boolean indexMedia = "on".equals(post.get("indexMedia", "false"));
env.setConfig("indexMedia", indexMedia);
env.setConfig("storeHTCache", storeHTCache);
String defaultAgentName =
sb.isIntranetMode()
? ClientIdentification.yacyIntranetCrawlerAgentName
: ClientIdentification.yacyInternetCrawlerAgentName;
String agentName = post.get("agentName", defaultAgentName);
ClientIdentification.Agent agent = ClientIdentification.getAgent(agentName);
if (agent == null) agent = ClientIdentification.getAgent(defaultAgentName);
CacheStrategy cachePolicy = CacheStrategy.parse(post.get("cachePolicy", "iffresh"));
if (cachePolicy == null) cachePolicy = CacheStrategy.IFFRESH;
String crawlingMode = post.get("crawlingMode", "url");
if ("file".equals(crawlingMode) && post.containsKey("crawlingFile")) {
newcrawlingMustNotMatch = CrawlProfile.MATCH_NEVER_STRING;
directDocByURL = false;
}
if ("sitemap".equals(crawlingMode)) {
newcrawlingMustMatch = CrawlProfile.MATCH_ALL_STRING;
newcrawlingMustNotMatch = CrawlProfile.MATCH_NEVER_STRING;
newcrawlingdepth = 0;
directDocByURL = false;
}
if ("sitelist".equals(crawlingMode)) {
newcrawlingMustNotMatch = CrawlProfile.MATCH_NEVER_STRING;
Set<DigestURL> newRootURLs = new HashSet<DigestURL>();
for (DigestURL sitelistURL : rootURLs) {
// download document
Document scraper;
try {
scraper =
sb.loader.loadDocument(
sitelistURL, CacheStrategy.IFFRESH, BlacklistType.CRAWLER, agent);
// get links and generate filter
for (DigestURL u : scraper.getHyperlinks().keySet()) {
newRootURLs.add(u);
}
} catch (final IOException e) {
ConcurrentLog.logException(e);
}
}
rootURLs = newRootURLs;
crawlingMode = "url";
if ((fullDomain || subPath) && newcrawlingdepth > 0)
newcrawlingMustMatch =
CrawlProfile
.MATCH_ALL_STRING; // to prevent that there is a restriction on the original
// urls
}
// delete all error urls for that domain
// and all urls for that host from the crawl queue
Set<String> hosthashes = new HashSet<String>();
boolean anysmbftporpdf = false;
for (DigestURL u : rootURLs) {
sb.index.fulltext().remove(u.hash());
hosthashes.add(u.hosthash());
if ("smb.ftp".indexOf(u.getProtocol()) >= 0
|| "pdf".equals(MultiProtocolURL.getFileExtension(u.getFileName())))
anysmbftporpdf = true;
}
sb.crawlQueues.removeHosts(hosthashes);
sb.index.fulltext().commit(true);
boolean crawlingQ =
anysmbftporpdf
|| "on".equals(post.get("crawlingQ", "off"))
|| "sitemap".equals(crawlingMode);
env.setConfig("crawlingQ", crawlingQ);
// compute mustmatch filter according to rootURLs
if ((fullDomain || subPath) && newcrawlingdepth > 0) {
String siteFilter = ".*";
if (fullDomain) {
siteFilter = CrawlProfile.siteFilter(rootURLs);
if (deleteold) {
sb.index.fulltext().deleteStaleDomainHashes(hosthashes, deleteageDate);
}
} else if (subPath) {
siteFilter = CrawlProfile.subpathFilter(rootURLs);
if (deleteold) {
for (DigestURL u : rootURLs) {
String basepath = u.toNormalform(true);
if (!basepath.endsWith("/")) {
int p = basepath.lastIndexOf("/");
if (p > 0) basepath = basepath.substring(0, p + 1);
}
int count = sb.index.fulltext().remove(basepath, deleteageDate);
if (count > 0)
ConcurrentLog.info(
"Crawler_p", "deleted " + count + " documents for host " + u.getHost());
}
}
}
if (CrawlProfile.MATCH_ALL_STRING.equals(newcrawlingMustMatch)) {
newcrawlingMustMatch = siteFilter;
} else if (!CrawlProfile.MATCH_ALL_STRING.equals(siteFilter)) {
// combine both
newcrawlingMustMatch = "(" + newcrawlingMustMatch + ")|(" + siteFilter + ")";
}
}
// check if the crawl filter works correctly
try {
Pattern mmp = Pattern.compile(newcrawlingMustMatch);
for (DigestURL u : rootURLs) {
assert mmp.matcher(u.toNormalform(true)).matches()
: "pattern " + mmp.toString() + " does not match url " + u.toNormalform(true);
}
} catch (final PatternSyntaxException e) {
prop.put("info", "4"); // crawlfilter does not match url
prop.putHTML("info_newcrawlingfilter", newcrawlingMustMatch);
prop.putHTML("info_error", e.getMessage());
}
boolean hasCrawlstartDataOK = !crawlName.isEmpty();
if (hasCrawlstartDataOK) {
// check crawlurl was given in sitecrawl
if ("url".equals(crawlingMode) && rootURLs.size() == 0) hasCrawlstartDataOK = false;
}
String snapshotsMaxDepthString = post.get("snapshotsMaxDepth", "-1");
int snapshotsMaxDepth = Integer.parseInt(snapshotsMaxDepthString);
boolean snapshotsLoadImage = post.getBoolean("snapshotsLoadImage");
boolean snapshotsReplaceOld = post.getBoolean("snapshotsReplaceOld");
String snapshotsMustnotmatch = post.get("snapshotsMustnotmatch", "");
// get vocabulary scraper info
JSONObject vocabulary_scraper =
new JSONObject(); // key = vocabulary_name, value = properties with key = type (i.e.
// 'class') and value = keyword in context
for (String key : post.keySet()) {
if (key.startsWith("vocabulary_")) {
if (key.endsWith("_class")) {
String vocabulary = key.substring(11, key.length() - 6);
String value = post.get(key);
if (value != null && value.length() > 0) {
JSONObject props;
try {
props = vocabulary_scraper.getJSONObject(vocabulary);
} catch (JSONException e) {
props = new JSONObject();
vocabulary_scraper.put(vocabulary, props);
}
props.put("class", value);
}
}
}
}
int timezoneOffset = post.getInt("timezoneOffset", 0);
// in case that we crawl from a file, load that file and re-compute mustmatch pattern
List<AnchorURL> hyperlinks_from_file = null;
if ("file".equals(crawlingMode)
&& post.containsKey("crawlingFile")
&& crawlingFile != null) {
final String crawlingFileContent = post.get("crawlingFile$file", "");
try {
// check if the crawl filter works correctly
final ContentScraper scraper =
new ContentScraper(
new DigestURL(crawlingFile), 10000000, new VocabularyScraper(), timezoneOffset);
final Writer writer = new TransformerWriter(null, null, scraper, null, false);
if (crawlingFile != null && crawlingFile.exists()) {
FileUtils.copy(new FileInputStream(crawlingFile), writer);
} else {
FileUtils.copy(crawlingFileContent, writer);
}
writer.close();
// get links and generate filter
hyperlinks_from_file = scraper.getAnchors();
if (newcrawlingdepth > 0) {
if (fullDomain) {
newcrawlingMustMatch = CrawlProfile.siteFilter(hyperlinks_from_file);
} else if (subPath) {
newcrawlingMustMatch = CrawlProfile.subpathFilter(hyperlinks_from_file);
}
}
} catch (final Exception e) {
// mist
prop.put("info", "7"); // Error with file
prop.putHTML("info_crawlingStart", crawlingFileName);
prop.putHTML("info_error", e.getMessage());
ConcurrentLog.logException(e);
}
sb.continueCrawlJob(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL);
}
// prepare a new crawling profile
final CrawlProfile profile;
byte[] handle;
if (hasCrawlstartDataOK) {
profile =
new CrawlProfile(
crawlName,
newcrawlingMustMatch,
newcrawlingMustNotMatch,
ipMustMatch,
ipMustNotMatch,
countryMustMatch,
crawlerNoDepthLimitMatch,
indexUrlMustMatch,
indexUrlMustNotMatch,
indexContentMustMatch,
indexContentMustNotMatch,
newcrawlingdepth,
directDocByURL,
crawlingIfOlder,
crawlingDomMaxPages,
crawlingQ,
followFrames,
obeyHtmlRobotsNoindex,
obeyHtmlRobotsNofollow,
indexText,
indexMedia,
storeHTCache,
crawlOrder,
snapshotsMaxDepth,
snapshotsLoadImage,
snapshotsReplaceOld,
snapshotsMustnotmatch,
cachePolicy,
collection,
agentName,
new VocabularyScraper(vocabulary_scraper),
timezoneOffset);
handle = ASCII.getBytes(profile.handle());
// before we fire up a new crawl, we make sure that another crawl with the same name is
// not running
sb.crawler.removeActive(handle);
sb.crawler.removePassive(handle);
try {
sb.crawlQueues.noticeURL.removeByProfileHandle(profile.handle(), 10000);
} catch (final SpaceExceededException e1) {
}
} else {
profile = null;
handle = null;
}
// start the crawl
if ("url".equals(crawlingMode)) {
if (rootURLs.size() == 0) {
prop.put("info", "5"); // Crawling failed
prop.putHTML("info_crawlingURL", "(no url given)");
prop.putHTML("info_reasonString", "you must submit at least one crawl url");
} else {
// stack requests
sb.crawler.putActive(handle, profile);
final Set<DigestURL> successurls = new HashSet<DigestURL>();
final Map<DigestURL, String> failurls = new HashMap<DigestURL, String>();
sb.stackURLs(rootURLs, profile, successurls, failurls);
if (failurls.size() == 0) {
// liftoff!
prop.put("info", "8");
prop.putHTML("info_crawlingURL", post.get("crawlingURL"));
// generate a YaCyNews if the global flag was set
if (!sb.isRobinsonMode() && crawlOrder) {
final Map<String, String> m =
new HashMap<String, String>(profile); // must be cloned
m.remove("specificDepth");
m.remove("indexText");
m.remove("indexMedia");
m.remove("remoteIndexing");
m.remove("xsstopw");
m.remove("xpstopw");
m.remove("xdstopw");
m.remove("storeTXCache");
m.remove("storeHTCache");
m.remove("generalFilter");
m.remove("specificFilter");
m.put("intention", post.get("intention", "").replace(',', '/'));
sb.peers.newsPool.publishMyNews(
sb.peers.mySeed(), NewsPool.CATEGORY_CRAWL_START, m);
}
} else {
StringBuilder fr = new StringBuilder();
for (Map.Entry<DigestURL, String> failure : failurls.entrySet()) {
sb.crawlQueues.errorURL.push(
failure.getKey(),
0,
null,
FailCategory.FINAL_LOAD_CONTEXT,
failure.getValue(),
-1);
fr.append(failure.getValue()).append('/');
}
prop.put("info", "5"); // Crawling failed
prop.putHTML("info_crawlingURL", (post.get("crawlingURL")));
prop.putHTML("info_reasonString", fr.toString());
}
if (successurls.size() > 0)
sb.continueCrawlJob(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL);
}
} else if ("sitemap".equals(crawlingMode)) {
try {
final DigestURL sitemapURL =
sitemapURLStr.indexOf("//") > 0
? new DigestURL(sitemapURLStr)
: new DigestURL(
rootURLs.iterator().next(),
sitemapURLStr); // fix for relative paths which should not exist but are
// used anyway
sb.crawler.putActive(handle, profile);
final SitemapImporter importer = new SitemapImporter(sb, sitemapURL, profile);
importer.start();
sb.continueCrawlJob(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL);
} catch (final Exception e) {
// mist
prop.put("info", "6"); // Error with url
prop.putHTML("info_crawlingStart", sitemapURLStr);
prop.putHTML("info_error", e.getMessage());
ConcurrentLog.logException(e);
}
} else if ("file".equals(crawlingMode)) {
if (post.containsKey("crawlingFile")
&& crawlingFile != null
&& hyperlinks_from_file != null) {
try {
if (newcrawlingdepth > 0) {
if (fullDomain) {
newcrawlingMustMatch = CrawlProfile.siteFilter(hyperlinks_from_file);
} else if (subPath) {
newcrawlingMustMatch = CrawlProfile.subpathFilter(hyperlinks_from_file);
}
}
sb.crawler.putActive(handle, profile);
sb.crawlStacker.enqueueEntriesAsynchronous(
sb.peers.mySeed().hash.getBytes(),
profile.handle(),
hyperlinks_from_file,
profile.timezoneOffset());
} catch (final PatternSyntaxException e) {
prop.put("info", "4"); // crawlfilter does not match url
prop.putHTML("info_newcrawlingfilter", newcrawlingMustMatch);
prop.putHTML("info_error", e.getMessage());
} catch (final Exception e) {
// mist
prop.put("info", "7"); // Error with file
prop.putHTML("info_crawlingStart", crawlingFileName);
prop.putHTML("info_error", e.getMessage());
ConcurrentLog.logException(e);
}
sb.continueCrawlJob(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL);
}
}
}
}
/*
* <input id="customPPM" name="customPPM" type="number" min="10" max="30000" style="width:46px" value="#[customPPMdefault]#" />PPM
<input id="latencyFactor" name="latencyFactor" type="number" min="0.1" max="3.0" step="0.1" style="width:32px" value="#[latencyFactorDefault]#" />LF
<input id="MaxSameHostInQueue" name="MaxSameHostInQueue" type="number" min="1" max="30" style="width:32px" value="#[MaxSameHostInQueueDefault]#" />MH
<input type="submit" name="crawlingPerformance" value="set" />
(<a href="/Crawler_p.html?crawlingPerformance=minimum">min</a>/<a href="/Crawler_p.html?crawlingPerformance=maximum">max</a>)
</td>
*/
if (post != null && post.containsKey("crawlingPerformance")) {
final String crawlingPerformance = post.get("crawlingPerformance", "custom");
final long LCbusySleep1 =
sb.getConfigLong(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL_BUSYSLEEP, 1000L);
int wantedPPM = (LCbusySleep1 == 0) ? 30000 : (int) (60000L / LCbusySleep1);
try {
wantedPPM = post.getInt("customPPM", wantedPPM);
} catch (final NumberFormatException e) {
}
if ("minimum".equals(crawlingPerformance.toLowerCase())) wantedPPM = 10;
if ("maximum".equals(crawlingPerformance.toLowerCase())) wantedPPM = 30000;
int wPPM = wantedPPM;
if (wPPM <= 0) {
wPPM = 1;
}
if (wPPM >= 30000) {
wPPM = 30000;
}
final int newBusySleep = 60000 / wPPM; // for wantedPPM = 10: 6000; for wantedPPM = 1000: 60
final float loadprereq =
wantedPPM <= 10 ? 1.0f : wantedPPM <= 100 ? 2.0f : wantedPPM >= 1000 ? 8.0f : 3.0f;
BusyThread thread;
thread = sb.getThread(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL);
if (thread != null) {
sb.setConfig(
SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL_BUSYSLEEP, thread.setBusySleep(newBusySleep));
sb.setConfig(
SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL_LOADPREREQ,
thread.setLoadPreReqisite(loadprereq));
thread.setLoadPreReqisite(loadprereq);
thread.setIdleSleep(2000);
}
float latencyFactor = post.getFloat("latencyFactor", 0.5f);
int MaxSameHostInQueue = post.getInt("MaxSameHostInQueue", 20);
env.setConfig(SwitchboardConstants.CRAWLER_LATENCY_FACTOR, latencyFactor);
env.setConfig(SwitchboardConstants.CRAWLER_MAX_SAME_HOST_IN_QUEUE, MaxSameHostInQueue);
}
// performance settings
final long LCbusySleep =
env.getConfigLong(SwitchboardConstants.CRAWLJOB_LOCAL_CRAWL_BUSYSLEEP, 1000L);
final int LCppm = (int) (60000L / Math.max(1, LCbusySleep));
prop.put("customPPMdefault", Integer.toString(LCppm));
prop.put(
"latencyFactorDefault",
env.getConfigFloat(SwitchboardConstants.CRAWLER_LATENCY_FACTOR, 0.5f));
prop.put(
"MaxSameHostInQueueDefault",
env.getConfigInt(SwitchboardConstants.CRAWLER_MAX_SAME_HOST_IN_QUEUE, 20));
// generate crawl profile table
int count = 0;
boolean dark = true;
final int domlistlength = (post == null) ? 160 : post.getInt("domlistlength", 160);
CrawlProfile profile;
// put active crawls into list
String hosts = "";
for (final byte[] h : sb.crawler.getActive()) {
profile = sb.crawler.getActive(h);
if (CrawlSwitchboard.DEFAULT_PROFILES.contains(profile.name())) continue;
profile.putProfileEntry("crawlProfilesShow_list_", prop, true, dark, count, domlistlength);
prop.put("crawlProfilesShow_list_" + count + "_debug", debug ? 1 : 0);
if (debug) {
RowHandleSet urlhashes = sb.crawler.getURLHashes(h);
prop.put(
"crawlProfilesShow_list_" + count + "_debug_count",
urlhashes == null ? "unknown" : Integer.toString(urlhashes.size()));
}
hosts = hosts + "," + profile.name();
dark = !dark;
count++;
}
prop.put("crawlProfilesShow_debug", debug ? 1 : 0);
prop.put("crawlProfilesShow_list", count);
prop.put("crawlProfilesShow_count", count);
prop.put("crawlProfilesShow", count == 0 ? 0 : 1);
prop.put("crawlProfilesShow_linkstructure", 0);
if (post != null) { // handle config button to display graphic
if (post.get("hidewebstructuregraph") != null)
sb.setConfig(SwitchboardConstants.DECORATION_GRAFICS_LINKSTRUCTURE, false);
if (post.get("showwebstructuregraph") != null)
sb.setConfig(SwitchboardConstants.DECORATION_GRAFICS_LINKSTRUCTURE, true);
}
if (count > 0
&& sb.getConfigBool(SwitchboardConstants.DECORATION_GRAFICS_LINKSTRUCTURE, true)) {
// collect the host names for 'wide' crawls which can be visualized
boolean showLinkstructure = hosts.length() > 0 && !hosts.contains("file:");
if (showLinkstructure) {
StringBuilder q = new StringBuilder();
hosts = hosts.substring(1);
q.append(CollectionSchema.host_s.getSolrFieldName())
.append(':')
.append(hosts)
.append(" OR ")
.append(CollectionSchema.host_s.getSolrFieldName())
.append(':')
.append("www.")
.append(hosts);
try {
prop.put(
"crawlProfilesShow_linkstructure",
count == 1
&& sb.index.fulltext().getDefaultConnector().getCountByQuery(q.toString()) > 0
? 1
: 2);
prop.put("crawlProfilesShow_linkstructure_hosts", hosts);
} catch (IOException e) {
}
}
}
// return rewrite properties
return prop;
}
@Override
public Request pop(boolean delay, CrawlSwitchboard cs, RobotsTxt robots) throws IOException {
// returns a crawl entry from the stack and ensures minimum delta times
long sleeptime = 0;
Request crawlEntry = null;
CrawlProfile profileEntry = null;
synchronized (this) {
mainloop:
while (true) {
Index depthStack = getLowestStack();
if (depthStack == null) return null;
Row.Entry rowEntry = null;
while (depthStack.size() > 0) {
rowEntry = depthStack.removeOne();
if (rowEntry != null) break;
}
if (rowEntry == null) continue mainloop;
crawlEntry = new Request(rowEntry);
// check blacklist (again) because the user may have created blacklist entries after the
// queue has been filled
if (Switchboard.urlBlacklist.isListed(BlacklistType.CRAWLER, crawlEntry.url())) {
if (log.isFine()) log.fine("URL '" + crawlEntry.url() + "' is in blacklist.");
continue mainloop;
}
// at this point we must check if the crawlEntry has relevance because the crawl profile
// still exists
// if not: return null. A calling method must handle the null value and try again
profileEntry = cs.get(UTF8.getBytes(crawlEntry.profileHandle()));
if (profileEntry == null) {
if (log.isFine()) log.fine("no profile entry for handle " + crawlEntry.profileHandle());
continue mainloop;
}
// depending on the caching policy we need sleep time to avoid DoS-like situations
sleeptime = Latency.getDomainSleepTime(robots, profileEntry, crawlEntry.url());
break;
}
}
if (crawlEntry == null) return null;
ClientIdentification.Agent agent =
profileEntry == null
? ClientIdentification.yacyInternetCrawlerAgent
: profileEntry.getAgent();
long robotsTime = Latency.getRobotsTime(robots, crawlEntry.url(), agent);
Latency.updateAfterSelection(crawlEntry.url(), profileEntry == null ? 0 : robotsTime);
if (delay && sleeptime > 0) {
// force a busy waiting here
// in best case, this should never happen if the balancer works properly
// this is only to protection against the worst case, where the crawler could
// behave in a DoS-manner
if (log.isInfo())
log.info(
"forcing crawl-delay of "
+ sleeptime
+ " milliseconds for "
+ crawlEntry.url().getHost()
+ ": "
+ Latency.waitingRemainingExplain(crawlEntry.url(), robots, agent));
long loops = sleeptime / 1000;
long rest = sleeptime % 1000;
if (loops < 3) {
rest = rest + 1000 * loops;
loops = 0;
}
Thread.currentThread()
.setName(
"Balancer waiting for "
+ crawlEntry.url().getHost()
+ ": "
+ sleeptime
+ " milliseconds");
synchronized (this) {
// must be synchronized here to avoid 'takeover' moves from other threads which then idle
// the same time which would not be enough
if (rest > 0) {
try {
this.wait(rest);
} catch (final InterruptedException e) {
}
}
for (int i = 0; i < loops; i++) {
if (log.isInfo())
log.info(
"waiting for "
+ crawlEntry.url().getHost()
+ ": "
+ (loops - i)
+ " seconds remaining...");
try {
this.wait(1000);
} catch (final InterruptedException e) {
}
}
}
Latency.updateAfterSelection(crawlEntry.url(), robotsTime);
}
return crawlEntry;
}