You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
282 lines
11 KiB
282 lines
11 KiB
// CrawlProfile.java
|
|
// ------------------------
|
|
// part of YaCy
|
|
// (C) by Michael Peter Christen; mc@yacy.net
|
|
// first published on http://www.anomic.de
|
|
// Frankfurt, Germany, 2004
|
|
// last major change: 31.08.2010
|
|
//
|
|
// This program is free software; you can redistribute it and/or modify
|
|
// it under the terms of the GNU General Public License as published by
|
|
// the Free Software Foundation; either version 2 of the License, or
|
|
// (at your option) any later version.
|
|
//
|
|
// This program is distributed in the hope that it will be useful,
|
|
// but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
// GNU General Public License for more details.
|
|
//
|
|
// You should have received a copy of the GNU General Public License
|
|
// along with this program; if not, write to the Free Software
|
|
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
|
|
|
package de.anomic.crawler;
|
|
|
|
import java.util.Map;
|
|
import java.util.concurrent.ConcurrentHashMap;
|
|
import java.util.regex.Pattern;
|
|
|
|
import net.yacy.cora.document.UTF8;
|
|
import net.yacy.kelondro.data.meta.DigestURI;
|
|
import net.yacy.kelondro.data.word.Word;
|
|
import net.yacy.kelondro.logging.Log;
|
|
import net.yacy.kelondro.order.Base64Order;
|
|
import net.yacy.kelondro.order.Digest;
|
|
|
|
public class CrawlProfile extends ConcurrentHashMap<String, String> implements Map<String, String> {
|
|
|
|
private static final long serialVersionUID = 5527325718810703504L;
|
|
|
|
public static final String MATCH_ALL = ".*";
|
|
public static final String MATCH_NEVER = "";
|
|
|
|
// this is a simple record structure that hold all properties of a single crawl start
|
|
public static final String HANDLE = "handle";
|
|
public static final String NAME = "name";
|
|
public static final String START_URL = "startURL";
|
|
public static final String FILTER_MUSTMATCH = "generalFilter";
|
|
public static final String FILTER_MUSTNOTMATCH = "nevermatch";
|
|
public static final String DEPTH = "generalDepth";
|
|
public static final String RECRAWL_IF_OLDER = "recrawlIfOlder";
|
|
public static final String DOM_MAX_PAGES = "domMaxPages";
|
|
public static final String CRAWLING_Q = "crawlingQ";
|
|
public static final String INDEX_TEXT = "indexText";
|
|
public static final String INDEX_MEDIA = "indexMedia";
|
|
public static final String STORE_HTCACHE = "storeHTCache";
|
|
public static final String REMOTE_INDEXING = "remoteIndexing";
|
|
public static final String XSSTOPW = "xsstopw";
|
|
public static final String XDSTOPW = "xdstopw";
|
|
public static final String XPSTOPW = "xpstopw";
|
|
public static final String CACHE_STRAGEGY = "cacheStrategy";
|
|
|
|
private Pattern mustmatch = null, mustnotmatch = null;
|
|
|
|
public CrawlProfile(
|
|
final String name,
|
|
final DigestURI startURL,
|
|
final String mustmatch,
|
|
final String mustnotmatch,
|
|
final int depth,
|
|
final long recrawlIfOlder /*date*/,
|
|
final int domMaxPages,
|
|
final boolean crawlingQ,
|
|
final boolean indexText,
|
|
final boolean indexMedia,
|
|
final boolean storeHTCache,
|
|
final boolean remoteIndexing,
|
|
final boolean xsstopw,
|
|
final boolean xdstopw,
|
|
final boolean xpstopw,
|
|
final CacheStrategy cacheStrategy) {
|
|
super(40);
|
|
if (name == null || name.length() == 0) throw new NullPointerException("name must not be null");
|
|
final String handle = (startURL == null) ? Base64Order.enhancedCoder.encode(Digest.encodeMD5Raw(name)).substring(0, Word.commonHashLength) : UTF8.String(startURL.hash());
|
|
put(HANDLE, handle);
|
|
put(NAME, name);
|
|
put(START_URL, (startURL == null) ? "" : startURL.toNormalform(true, false));
|
|
put(FILTER_MUSTMATCH, (mustmatch == null) ? CrawlProfile.MATCH_ALL : mustmatch);
|
|
put(FILTER_MUSTNOTMATCH, (mustnotmatch == null) ? CrawlProfile.MATCH_NEVER : mustnotmatch);
|
|
put(DEPTH, depth);
|
|
put(RECRAWL_IF_OLDER, recrawlIfOlder);
|
|
put(DOM_MAX_PAGES, domMaxPages);
|
|
put(CRAWLING_Q, crawlingQ); // crawling of urls with '?'
|
|
put(INDEX_TEXT, indexText);
|
|
put(INDEX_MEDIA, indexMedia);
|
|
put(STORE_HTCACHE, storeHTCache);
|
|
put(REMOTE_INDEXING, remoteIndexing);
|
|
put(XSSTOPW, xsstopw); // exclude static stop-words
|
|
put(XDSTOPW, xdstopw); // exclude dynamic stop-word
|
|
put(XPSTOPW, xpstopw); // exclude parent stop-words
|
|
put(CACHE_STRAGEGY, cacheStrategy.toString());
|
|
}
|
|
|
|
public CrawlProfile(Map<String, String> ext) {
|
|
super(ext == null ? 1 : ext.size());
|
|
if (ext != null) this.putAll(ext);
|
|
}
|
|
|
|
public void put(String key, boolean value) {
|
|
super.put(key, Boolean.toString(value));
|
|
}
|
|
|
|
public void put(String key, int value) {
|
|
super.put(key, Integer.toString(value));
|
|
}
|
|
|
|
public void put(String key, long value) {
|
|
super.put(key, Long.toString(value));
|
|
}
|
|
|
|
public String handle() {
|
|
final String r = get(HANDLE);
|
|
//if (r == null) return null;
|
|
return r;
|
|
}
|
|
public String name() {
|
|
final String r = get(NAME);
|
|
if (r == null) return "";
|
|
return r;
|
|
}
|
|
public String startURL() {
|
|
final String r = get(START_URL);
|
|
return r;
|
|
}
|
|
public Pattern mustMatchPattern() {
|
|
if (this.mustmatch == null) {
|
|
String r = get(FILTER_MUSTMATCH);
|
|
if (r == null) r = CrawlProfile.MATCH_ALL;
|
|
this.mustmatch = Pattern.compile(r);
|
|
}
|
|
return this.mustmatch;
|
|
}
|
|
public Pattern mustNotMatchPattern() {
|
|
if (this.mustnotmatch == null) {
|
|
String r = get(FILTER_MUSTNOTMATCH);
|
|
if (r == null) r = CrawlProfile.MATCH_NEVER;
|
|
this.mustnotmatch = Pattern.compile(r);
|
|
}
|
|
return this.mustnotmatch;
|
|
}
|
|
public int depth() {
|
|
final String r = get(DEPTH);
|
|
if (r == null) return 0;
|
|
try {
|
|
return Integer.parseInt(r);
|
|
} catch (final NumberFormatException e) {
|
|
Log.logException(e);
|
|
return 0;
|
|
}
|
|
}
|
|
public CacheStrategy cacheStrategy() {
|
|
final String r = get(CACHE_STRAGEGY);
|
|
if (r == null) return CacheStrategy.IFFRESH;
|
|
try {
|
|
return CacheStrategy.decode(Integer.parseInt(r));
|
|
} catch (final NumberFormatException e) {
|
|
Log.logException(e);
|
|
return CacheStrategy.IFFRESH;
|
|
}
|
|
}
|
|
public void setCacheStrategy(CacheStrategy newStrategy) {
|
|
put(CACHE_STRAGEGY, newStrategy.toString());
|
|
}
|
|
public long recrawlIfOlder() {
|
|
// returns a long (millis) that is the minimum age that
|
|
// an entry must have to be re-crawled
|
|
final String r = get(RECRAWL_IF_OLDER);
|
|
if (r == null) return 0L;
|
|
try {
|
|
final long l = Long.parseLong(r);
|
|
return (l < 0) ? 0L : l;
|
|
} catch (final NumberFormatException e) {
|
|
Log.logException(e);
|
|
return 0L;
|
|
}
|
|
}
|
|
public int domMaxPages() {
|
|
// this is the maximum number of pages that are crawled for a single domain
|
|
// if -1, this means no limit
|
|
final String r = get(DOM_MAX_PAGES);
|
|
if (r == null) return Integer.MAX_VALUE;
|
|
try {
|
|
final int i = Integer.parseInt(r);
|
|
if (i < 0) return Integer.MAX_VALUE;
|
|
return i;
|
|
} catch (final NumberFormatException e) {
|
|
Log.logException(e);
|
|
return Integer.MAX_VALUE;
|
|
}
|
|
}
|
|
public boolean crawlingQ() {
|
|
final String r = get(CRAWLING_Q);
|
|
if (r == null) return false;
|
|
return (r.equals(Boolean.TRUE.toString()));
|
|
}
|
|
public boolean indexText() {
|
|
final String r = get(INDEX_TEXT);
|
|
if (r == null) return true;
|
|
return (r.equals(Boolean.TRUE.toString()));
|
|
}
|
|
public boolean indexMedia() {
|
|
final String r = get(INDEX_MEDIA);
|
|
if (r == null) return true;
|
|
return (r.equals(Boolean.TRUE.toString()));
|
|
}
|
|
public boolean storeHTCache() {
|
|
final String r = get(STORE_HTCACHE);
|
|
if (r == null) return false;
|
|
return (r.equals(Boolean.TRUE.toString()));
|
|
}
|
|
public boolean remoteIndexing() {
|
|
final String r = get(REMOTE_INDEXING);
|
|
if (r == null) return false;
|
|
return (r.equals(Boolean.TRUE.toString()));
|
|
}
|
|
public boolean excludeStaticStopwords() {
|
|
final String r = get(XSSTOPW);
|
|
if (r == null) return false;
|
|
return (r.equals(Boolean.TRUE.toString()));
|
|
}
|
|
public boolean excludeDynamicStopwords() {
|
|
final String r = get(XDSTOPW);
|
|
if (r == null) return false;
|
|
return (r.equals(Boolean.TRUE.toString()));
|
|
}
|
|
public boolean excludeParentStopwords() {
|
|
final String r = get(XPSTOPW);
|
|
if (r == null) return false;
|
|
return (r.equals(Boolean.TRUE.toString()));
|
|
}
|
|
|
|
public static enum CacheStrategy {
|
|
NOCACHE(0), // never use the cache, all content from fresh internet source
|
|
IFFRESH(1), // use the cache if the cache exists and is fresh using the proxy-fresh rules
|
|
IFEXIST(2), // use the cache if the cache exist. Do no check freshness. Otherwise use online source.
|
|
CACHEONLY(3); // never go online, use all content from cache. If no cache entry exist, consider content nevertheless as available
|
|
public int code;
|
|
private CacheStrategy(int code) {
|
|
this.code = code;
|
|
}
|
|
public String toString() {
|
|
return Integer.toString(this.code);
|
|
}
|
|
public static CacheStrategy decode(int code) {
|
|
for (CacheStrategy strategy: CacheStrategy.values()) if (strategy.code == code) return strategy;
|
|
return NOCACHE;
|
|
}
|
|
public static CacheStrategy parse(String name) {
|
|
if (name == null) return null;
|
|
if (name.equals("nocache")) return NOCACHE;
|
|
if (name.equals("iffresh")) return IFFRESH;
|
|
if (name.equals("ifexist")) return IFEXIST;
|
|
if (name.equals("cacheonly")) return CACHEONLY;
|
|
if (name.equals("true")) return IFFRESH;
|
|
if (name.equals("false")) return null; // if this cache strategy is assigned as query attribute, null means "do not create a snippet"
|
|
return null;
|
|
}
|
|
public String toName() {
|
|
return this.name().toLowerCase();
|
|
}
|
|
public boolean isAllowedToFetchOnline() {
|
|
return this.code < 3;
|
|
}
|
|
public boolean mustBeOffline() {
|
|
return this.code == 3;
|
|
}
|
|
}
|
|
|
|
public static long getRecrawlDate(final long oldTimeMinutes) {
|
|
return System.currentTimeMillis() - (60000L * oldTimeMinutes);
|
|
}
|
|
}
|