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.
yacy_search_server/source/de/anomic/plasma/plasmaSnippetCache.java

425 lines
18 KiB

// plasmaSnippetCache.java
// -----------------------
// part of YaCy
// (C) by Michael Peter Christen; mc@anomic.de
// first published on http://www.anomic.de
// Frankfurt, Germany, 2005
// last major change: 07.06.2005
//
// 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
//
// Using this software in any meaning (reading, learning, copying, compiling,
// running) means that you agree that the Author(s) is (are) not responsible
// for cost, loss of data or any harm that may be caused directly or indirectly
// by usage of this softare or this documentation. The usage of this software
// is on your own risk. The installation and usage (starting/running) of this
// software may allow other people or application to access your computer and
// any attached devices and is highly dependent on the configuration of the
// software which must be done by the user of the software; the author(s) is
// (are) also not responsible for proper configuration and usage of the
// software, even if provoked by documentation provided together with
// the software.
//
// Any changes to this file according to the GPL as documented in the file
// gpl.txt aside this file in the shipment you received can be done to the
// lines that follows this copyright notice here, but changes must not be
// done inside the copyright notive above. A re-distribution must contain
// the intact and unchanged copyright notice.
// Contributions and changes to the program code must be marked as such.
package de.anomic.plasma;
import java.io.IOException;
import java.net.URL;
import java.util.Enumeration;
import java.util.HashMap;
import java.util.HashSet;
import java.util.Iterator;
import java.util.Set;
import de.anomic.http.httpHeader;
import de.anomic.kelondro.kelondroMScoreCluster;
import de.anomic.server.logging.serverLog;
import de.anomic.yacy.yacySearch;
import de.anomic.htmlFilter.htmlFilterContentScraper;
public class plasmaSnippetCache {
private static final int maxCache = 500;
public static final int SOURCE_CACHE = 0;
public static final int SOURCE_FILE = 1;
public static final int SOURCE_WEB = 2;
public static final int ERROR_NO_HASH_GIVEN = 11;
public static final int ERROR_SOURCE_LOADING = 12;
public static final int ERROR_RESOURCE_LOADING = 13;
public static final int ERROR_PARSER_FAILED = 14;
public static final int ERROR_PARSER_NO_LINES = 15;
public static final int ERROR_NO_MATCH = 16;
private int snippetsScoreCounter;
private kelondroMScoreCluster snippetsScore;
private HashMap snippetsCache;
private plasmaHTCache cacheManager;
private plasmaParser parser;
private serverLog log;
private plasmaSwitchboard sb;
public plasmaSnippetCache(
plasmaSwitchboard theSb,
plasmaHTCache cacheManager,
plasmaParser parser,
serverLog log
) {
this.cacheManager = cacheManager;
this.parser = parser;
this.log = log;
this.sb = theSb;
this.snippetsScoreCounter = 0;
this.snippetsScore = new kelondroMScoreCluster();
this.snippetsCache = new HashMap();
}
public class result {
public String line;
public String error;
public int source;
public result(String line, int source, String errortext) {
this.line = line;
this.source = source;
this.error = errortext;
}
public String toString() {
return line;
}
}
public boolean existsInCache(URL url, Set queryhashes) {
String hashes = yacySearch.set2string(queryhashes);
return retrieveFromCache(hashes, plasmaURL.urlHash(url)) != null;
}
public result retrieve(URL url, Set queryhashes, boolean fetchOnline, int snippetMaxLength) {
// heise = "0OQUNU3JSs05"
if (queryhashes.size() == 0) {
//System.out.println("found no queryhashes for URL retrieve " + url);
return new result(null, ERROR_NO_HASH_GIVEN, "no query hashes given");
}
String urlhash = plasmaURL.urlHash(url);
// try to get snippet from snippetCache
int source = SOURCE_CACHE;
String wordhashes = yacySearch.set2string(queryhashes);
String line = retrieveFromCache(wordhashes, urlhash);
if (line != null) {
//System.out.println("found snippet for URL " + url + " in cache: " + line);
return new result(line, source, null);
}
// if the snippet is not in the cache, we can try to get it from the htcache
byte[] resource = null;
try {
resource = cacheManager.loadResource(url);
if ((fetchOnline) && (resource == null)) {
loadResourceFromWeb(url, 5000);
resource = cacheManager.loadResource(url);
source = SOURCE_WEB;
}
} catch (IOException e) {
return new result(null, ERROR_SOURCE_LOADING, "error loading resource from web: " + e.getMessage());
}
if (resource == null) {
//System.out.println("cannot load document for URL " + url);
return new result(null, ERROR_RESOURCE_LOADING, "error loading resource from web, cacheManager returned NULL");
}
plasmaParserDocument document = parseDocument(url, resource);
if (document == null) return new result(null, ERROR_PARSER_FAILED, "parser error/failed"); // cannot be parsed
//System.out.println("loaded document for URL " + url);
String[] sentences = document.getSentences();
//System.out.println("----" + url.toString()); for (int l = 0; l < sentences.length; l++) System.out.println(sentences[l]);
if ((sentences == null) || (sentences.length == 0)) {
//System.out.println("found no sentences in url " + url);
return new result(null, ERROR_PARSER_NO_LINES, "parser returned no sentences");
}
// we have found a parseable non-empty file: use the lines
line = computeSnippet(sentences, queryhashes, 8 + 6 * queryhashes.size(), snippetMaxLength);
//System.out.println("loaded snippet for URL " + url + ": " + line);
if (line == null) return new result(null, ERROR_NO_MATCH, "no matching snippet found");
if (line.length() > snippetMaxLength) line = line.substring(0, snippetMaxLength);
// finally store this snippet in our own cache
storeToCache(wordhashes, urlhash, line);
return new result(line, source, null);
}
public void storeToCache(String wordhashes, String urlhash, String snippet) {
// generate key
String key = urlhash + wordhashes;
// do nothing if snippet is known
if (snippetsCache.containsKey(key)) return;
// learn new snippet
snippetsScore.addScore(key, snippetsScoreCounter++);
snippetsCache.put(key, snippet);
// care for counter
if (snippetsScoreCounter == java.lang.Integer.MAX_VALUE) {
snippetsScoreCounter = 0;
snippetsScore = new kelondroMScoreCluster();
snippetsCache = new HashMap();
}
// flush cache if cache is full
while (snippetsCache.size() > maxCache) {
key = (String) snippetsScore.getMinObject();
snippetsScore.deleteScore(key);
snippetsCache.remove(key);
}
}
private String retrieveFromCache(String wordhashes, String urlhash) {
// generate key
String key = urlhash + wordhashes;
return (String) snippetsCache.get(key);
}
private String computeSnippet(String[] sentences, Set queryhashes, int minLength, int maxLength) {
try {
if ((sentences == null) || (sentences.length == 0)) return null;
if ((queryhashes == null) || (queryhashes.size() == 0)) return null;
kelondroMScoreCluster hitTable = new kelondroMScoreCluster();
Iterator j;
HashMap hs;
String hash;
for (int i = 0; i < sentences.length; i++) {
//System.out.println("Sentence " + i + ": " + sentences[i]);
if (sentences[i].length() > minLength) {
hs = hashSentence(sentences[i]);
j = queryhashes.iterator();
while (j.hasNext()) {
hash = (String) j.next();
if (hs.containsKey(hash)) {
//System.out.println("hash " + hash + " appears in line " + i);
hitTable.incScore(new Integer(i));
}
}
}
}
int score = hitTable.getMaxScore(); // best number of hits
if (score <= 0) return null;
// we found (a) line(s) that have <score> hits.
// now find the shortest line of these hits
int shortLineIndex = -1;
int shortLineLength = Integer.MAX_VALUE;
for (int i = 0; i < sentences.length; i++) {
if ((hitTable.getScore(new Integer(i)) == score) &&
(sentences[i].length() < shortLineLength)) {
shortLineIndex = i;
shortLineLength = sentences[i].length();
}
}
// find a first result
String result = sentences[shortLineIndex];
// remove all hashes that appear in the result
hs = hashSentence(result);
j = queryhashes.iterator();
Integer pos;
Set remaininghashes = new HashSet();
int p, minpos = result.length(), maxpos = -1;
while (j.hasNext()) {
hash = (String) j.next();
pos = (Integer) hs.get(hash);
if (pos == null) {
remaininghashes.add(new String(hash));
} else {
p = pos.intValue();
if (p > maxpos) maxpos = p;
if (p < minpos) minpos = p;
}
}
// check result size
maxpos = maxpos + 10;
if (maxpos > result.length()) maxpos = result.length();
if (minpos < 0) minpos = 0;
// we have a result, but is it short enough?
if (maxpos - minpos + 10 > maxLength) {
// the string is too long, even if we cut at both ends
// so cut here in the middle of the string
int lenb = result.length();
result = result.substring(0, (minpos + 20 > result.length()) ? result.length() : minpos + 20).trim() +
" [..] " +
result.substring((maxpos + 26 > result.length()) ? result.length() : maxpos + 26).trim();
maxpos = maxpos + lenb - result.length() + 6;
}
if (maxpos > maxLength) {
// the string is too long, even if we cut it at the end
// so cut it here at both ends at once
int newlen = maxpos - minpos + 10;
int around = (maxLength - newlen) / 2;
result = "[..] " + result.substring(minpos - around, ((maxpos + around) > result.length()) ? result.length() : (maxpos + around)).trim() + " [..]";
minpos = around;
maxpos = result.length() - around - 5;
}
if (result.length() > maxLength) {
// trim result, 1st step (cut at right side)
result = result.substring(0, maxpos).trim() + " [..]";
}
if (result.length() > maxLength) {
// trim result, 2nd step (cut at left side)
result = "[..] " + result.substring(minpos).trim();
}
if (result.length() > maxLength) {
// trim result, 3rd step (cut in the middle)
result = result.substring(6, 20).trim() + " [..] " + result.substring(result.length() - 26, result.length() - 6).trim();
}
if (queryhashes.size() == 0) return result;
// the result has not all words in it.
// find another sentence that represents the missing other words
// and find recursively more sentences
maxLength = maxLength - result.length();
if (maxLength < 20) maxLength = 20;
String nextSnippet = computeSnippet(sentences, remaininghashes, minLength, maxLength);
return result + ((nextSnippet == null) ? "" : (" / " + nextSnippet));
} catch (IndexOutOfBoundsException e) {
log.logSevere("computeSnippet: error with string generation", e);
return "";
}
}
private HashMap hashSentence(String sentence) {
// generates a word-wordPos mapping
HashMap map = new HashMap();
Enumeration words = plasmaCondenser.wordTokenizer(sentence, 0);
int pos = 0;
String word;
while (words.hasMoreElements()) {
word = (String) words.nextElement();
map.put(plasmaWordIndexEntry.word2hash(word), new Integer(pos));
pos += word.length() + 1;
}
return map;
}
public plasmaParserDocument parseDocument(URL url, byte[] resource) {
if (resource == null) return null;
httpHeader header = null;
try {
header = this.cacheManager.getCachedResponse(plasmaURL.urlHash(url));
} catch (IOException e) {}
if (header == null) {
String filename = this.cacheManager.getCachePath(url).getName();
int p = filename.lastIndexOf('.');
if ( // if no extension is available
(p < 0) ||
// or the extension is supported by one of the parsers
((p >= 0) && (plasmaParser.supportedFileExtContains(filename.substring(p + 1))))
) {
String supposedMime = "text/html";
// if the mimeType Parser is installed we can set the mimeType to null to force
// a mimetype detection
if (plasmaParser.supportedMimeTypesContains("application/octet-stream")) {
supposedMime = null;
} else if (p != -1){
// otherwise we try to determine the mimeType per file Extension
supposedMime = plasmaParser.getMimeTypeByFileExt(filename.substring(p + 1));
}
return this.parser.parseSource(url, supposedMime, resource);
} else {
return null;
}
} else {
if (plasmaParser.supportedMimeTypesContains(header.mime())) {
return this.parser.parseSource(url, header.mime(), resource);
} else {
return null;
}
}
}
public byte[] getResource(URL url, boolean fetchOnline) {
// load the url as resource from the web
try {
//return httpc.singleGET(url, 5000, null, null, remoteProxyHost, remoteProxyPort);
byte[] resource = cacheManager.loadResource(url);
if ((fetchOnline) && (resource == null)) {
loadResourceFromWeb(url, 5000);
resource = cacheManager.loadResource(url);
}
return resource;
} catch (IOException e) {
return null;
}
}
public void loadResourceFromWeb(URL url, int socketTimeout) throws IOException {
plasmaCrawlWorker.load(
url,
"",
null,
null,
0,
null,
socketTimeout,
this.sb.remoteProxyConfig,
this.cacheManager,
this.log);
}
public void fetch(plasmaSearchResult acc, Set queryhashes, String urlmask, int fetchcount, long maxTime) {
// fetch snippets
int i = 0;
plasmaCrawlLURL.Entry urlentry;
String urlstring;
long limitTime = (maxTime < 0) ? Long.MAX_VALUE : System.currentTimeMillis() + maxTime;
while ((acc.hasMoreElements()) && (i < fetchcount) && (System.currentTimeMillis() < limitTime)) {
urlentry = acc.nextElement();
if (urlentry.url().getHost().endsWith(".yacyh")) continue;
urlstring = htmlFilterContentScraper.urlNormalform(urlentry.url());
if ((urlstring.matches(urlmask)) &&
(!(existsInCache(urlentry.url(), queryhashes)))) {
new Fetcher(urlentry.url(), queryhashes).start();
i++;
}
}
}
public class Fetcher extends Thread {
URL url;
Set queryhashes;
public Fetcher(URL url, Set queryhashes) {
if (url.getHost().endsWith(".yacyh")) return;
this.url = url;
this.queryhashes = queryhashes;
}
public void run() {
log.logFine("snippetFetcher: try to get URL " + url);
plasmaSnippetCache.result snippet = retrieve(url, queryhashes, true, 260);
if (snippet.line == null)
log.logFine("snippetFetcher: cannot get URL " + url + ". error(" + snippet.source + "): " + snippet.error);
else
log.logFine("snippetFetcher: got URL " + url + ", the snippet is '" + snippet.line + "', source=" + snippet.source);
}
}
}