boxtec
/
yacy_search_server
mirror of https://github.com/yacy/yacy_search_server
4
1
Fork 0
Code Issues Projects Releases Wiki Activity

removed snippet pattern filter - it was not used

Browse Source
pull/1/head
Michael Peter Christen 13 years ago
parent d3964253ae
commit 241dd8410a
9 changed files with 11 additions and 62 deletions
Show Stats Download Patch File Download Diff File

3
htroot/api/ymarks/manage_tags.java
Unescape View File

@ -4,7 +4,6 @@ import java.util.regex.Pattern;
import net.yacy.cora.protocol.RequestHeader;
import net.yacy.kelondro.blob.Tables.Row;
import net.yacy.kelondro.index.RowSpaceExceededException;
import net.yacy.kelondro.logging.Log;
import net.yacy.search.Switchboard;
import de.anomic.data.UserDB;
@ -75,8 +74,6 @@ public class manage_tags {
sb.tables.bookmarks.replaceTags(row_iter, bmk_user, tags, replace);
prop.put("status", 1);
} catch (final IOException e) {
Log.logException(e);
} catch (final RowSpaceExceededException e) {
Log.logException(e);
}
} else {

4
htroot/yacy/search.java
Unescape View File

@ -35,7 +35,6 @@ import java.util.Iterator;
import java.util.Map;
import java.util.TreeMap;
import java.util.TreeSet;
import java.util.regex.Pattern;
import net.yacy.cora.document.ASCII;
import net.yacy.cora.document.Classification;
@ -120,7 +119,6 @@ public final class search {
final String modifier = post.get("modifier", "").trim();
final String contentdom = post.get("contentdom", "all");
final String filter = post.get("filter", ".*"); // a filter on the url
final Pattern snippetPattern = Pattern.compile(post.get("snippet", ".*")); // a filter on the snippet
String sitehash = post.get("sitehash", ""); if (sitehash.length() == 0) sitehash = null;
String authorhash = post.get("authorhash", ""); if (authorhash.length() == 0) authorhash = null;
String language = post.get("language", "");
@ -228,7 +226,6 @@ public final class search {
abstractSet,
new HandleSet(WordReferenceRow.urlEntryRow.primaryKeyLength, WordReferenceRow.urlEntryRow.objectOrder, 0),
null,
snippetPattern,
null,
modifier,
maxdist,
@ -291,7 +288,6 @@ public final class search {
queryhashes,
excludehashes,
null,
snippetPattern,
null,
modifier,
maxdist,

2
htroot/yacysearch.java
Unescape View File

@ -713,7 +713,6 @@ public class yacysearch {
// do the search
final HandleSet queryHashes = Word.words2hashesHandles(query[0]);
final Pattern snippetPattern = QueryParams.stringSearchPattern(originalquerystring);
// check filters
try {
@ -740,7 +739,6 @@ public class yacysearch {
queryHashes,
Word.words2hashesHandles(query[1]),
Word.words2hashesHandles(query[2]),
snippetPattern,
tenant,
modifier.toString().trim(),
maxDistance,

2
source/de/anomic/data/ymark/YMarkTables.java
Unescape View File

@ -238,7 +238,7 @@ public class YMarkTables {
}
}
public void replaceTags(final Iterator<Row> rowIterator, final String bmk_user, final String tagString, final String replaceString) throws IOException, RowSpaceExceededException {
public void replaceTags(final Iterator<Row> rowIterator, final String bmk_user, final String tagString, final String replaceString) throws IOException {
final HashSet<String> remove = YMarkUtil.keysStringToSet(YMarkUtil.cleanTagsString(tagString.toLowerCase()));
final StringBuilder t = new StringBuilder(200);
HashSet<String> tags;

5
source/net/yacy/peers/Protocol.java
Unescape View File

@ -589,7 +589,6 @@ public final class Protocol
final String urlhashes,
final Pattern prefer,
final Pattern filter,
final Pattern snippet,
final String modifier,
final String language,
final String sitehash,
@ -635,7 +634,6 @@ public final class Protocol
urlhashes,
prefer,
filter,
snippet,
modifier,
language,
sitehash,
@ -893,7 +891,6 @@ public final class Protocol
final String urlhashes,
final Pattern prefer,
final Pattern filter,
final Pattern snippet,
final String modifier,
final String language,
final String sitehash,
@ -946,7 +943,6 @@ public final class Protocol
parts.put("urls", UTF8.StringBody(urlhashes));
parts.put("prefer", UTF8.StringBody(prefer.pattern()));
parts.put("filter", UTF8.StringBody(filter.pattern()));
parts.put("snippet", UTF8.StringBody(snippet.pattern()));
parts.put("modifier", UTF8.StringBody(modifier));
parts.put("language", UTF8.StringBody(language));
parts.put("sitehash", UTF8.StringBody(sitehash));
@ -1518,7 +1514,6 @@ public final class Protocol
"", // urlhashes,
QueryParams.matchnothing_pattern, // prefer,
QueryParams.catchall_pattern, // filter,
QueryParams.catchall_pattern, // snippet,
"", // modifier
"", // language,
"", // sitehash,

12
source/net/yacy/peers/RemoteSearch.java
Unescape View File

@ -59,7 +59,7 @@ public class RemoteSearch extends Thread {
private final int count, maxDistance;
private final long time;
final private RankingProfile rankingProfile;
final private Pattern prefer, filter, snippet;
final private Pattern prefer, filter;
final private QueryParams.Modifier modifier;
final private String language;
final private Bitfield constraint;
@ -70,7 +70,6 @@ public class RemoteSearch extends Thread {
final String urlhashes, // this is the field that is filled during a secondary search to restrict to specific urls that are to be retrieved
final Pattern prefer,
final Pattern filter,
final Pattern snippet,
final QueryParams.Modifier modifier,
final String language,
final String sitehash, final String authorhash, final String contentdom,
@ -92,7 +91,6 @@ public class RemoteSearch extends Thread {
this.urlhashes = urlhashes;
this.prefer = prefer;
this.filter = filter;
this.snippet = snippet;
this.modifier = modifier;
this.language = language;
this.sitehash = sitehash;
@ -121,7 +119,7 @@ public class RemoteSearch extends Thread {
this.urls = Protocol.search(
this.peers.mySeed(),
this.wordhashes, this.excludehashes, this.urlhashes,
this.prefer, this.filter, this.snippet, this.modifier.getModifier(),
this.prefer, this.filter, this.modifier.getModifier(),
this.language, this.sitehash, this.authorhash, this.contentdom,
this.count, this.time, this.maxDistance, this.global, this.partitions,
this.targetPeer, this.indexSegment, this.containerCache, this.secondarySearchSuperviser,
@ -163,7 +161,7 @@ public class RemoteSearch extends Thread {
public static void primaryRemoteSearches(
final List<RemoteSearch> searchThreads,
final String wordhashes, final String excludehashes,
final Pattern prefer, final Pattern filter, final Pattern snippet,
final Pattern prefer, final Pattern filter,
final QueryParams.Modifier modifier,
final String language,
final String sitehash,
@ -202,7 +200,7 @@ public class RemoteSearch extends Thread {
if (targetPeers[i] == null || targetPeers[i].hash == null) continue;
try {
RemoteSearch rs = new RemoteSearch(
wordhashes, excludehashes, "", prefer, filter, snippet, modifier,
wordhashes, excludehashes, "", prefer, filter, modifier,
language, sitehash, authorhash, contentdom,
count, time, maxDist, true, targets, targetPeers[i],
indexSegment, peers, containerCache, secondarySearchSuperviser, blacklist, rankingProfile, constraint);
@ -237,7 +235,7 @@ public class RemoteSearch extends Thread {
StringBuilder whs = new StringBuilder(24);
for (String s: wordhashes) whs.append(s);
final RemoteSearch searchThread = new RemoteSearch(
whs.toString(), "", urlhashes, QueryParams.matchnothing_pattern, QueryParams.catchall_pattern, QueryParams.catchall_pattern, new QueryParams.Modifier(""), "", "", "", "all", 20, time, 9999, true, 0, targetPeer,
whs.toString(), "", urlhashes, QueryParams.matchnothing_pattern, QueryParams.catchall_pattern, new QueryParams.Modifier(""), "", "", "", "all", 20, time, 9999, true, 0, targetPeer,
indexSegment, peers, containerCache, null, blacklist, rankingProfile, constraint);
searchThread.start();
return searchThread;

31
source/net/yacy/search/query/QueryParams.java
Unescape View File

@ -35,7 +35,6 @@ import java.util.Iterator;
import java.util.Map;
import java.util.Set;
import java.util.SortedSet;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
import java.util.regex.PatternSyntaxException;
@ -108,7 +107,6 @@ public final class QueryParams {
public final String queryString;
public HandleSet fullqueryHashes, queryHashes, excludeHashes;
public Pattern snippetMatcher;
public final int itemsPerPage;
public int offset;
public final Pattern urlMask, prefer;
@ -167,7 +165,6 @@ public final class QueryParams {
this.excludeHashes = Word.words2hashesHandles(cq[1]);
this.fullqueryHashes = Word.words2hashesHandles(cq[2]);
}
this.snippetMatcher = QueryParams.catchall_pattern;
this.ranking = ranking;
this.tenant = null;
this.modifier = new Modifier("");
@ -210,7 +207,6 @@ public final class QueryParams {
final String queryString, final HandleSet queryHashes,
final HandleSet excludeHashes,
final HandleSet fullqueryHashes,
final Pattern snippetMatcher,
final String tenant,
final String modifier,
final int maxDistance, final String prefer, final ContentDomain contentdom,
@ -237,7 +233,6 @@ public final class QueryParams {
this.queryHashes = queryHashes;
this.excludeHashes = excludeHashes;
this.fullqueryHashes = fullqueryHashes;
this.snippetMatcher = snippetMatcher;
this.tenant = (tenant != null && tenant.length() == 0) ? null : tenant;
this.modifier = new Modifier(modifier == null ? "" : modifier);
this.ranking = ranking;
@ -607,30 +602,4 @@ public final class QueryParams {
return sb;
}
private static Pattern StringMatchPattern = Pattern.compile(".*?(\".*?\").*");
/**
* calculate a pattern to match with a string search
* @param query
* @return
*/
public static Pattern stringSearchPattern(String query) {
final StringBuilder p = new StringBuilder(query.length());
p.append("(?iu)");
int seqc = 0;
while (query.length() > 0) {
final Matcher m = StringMatchPattern.matcher(query);
if (!m.matches()) break;
p.append(".*?").append(query.substring(m.start(1) + 1, m.end(1) - 1));
query = query.substring(m.end(1));
seqc++;
}
if (seqc == 0) return QueryParams.catchall_pattern;
p.append(".*");
return Pattern.compile(p.toString());
}
public static void main(final String[] args) {
final Pattern p = stringSearchPattern("die \"peer-to-peer Suchmaschine\" ohne Zensur als \"freie Software\" runterladen");
System.out.println(p.toString());
}
}

1
source/net/yacy/search/query/SearchEvent.java
Unescape View File

@ -172,7 +172,6 @@ public final class SearchEvent
QueryParams.hashSet2hashString(SearchEvent.this.query.excludeHashes),
SearchEvent.this.query.prefer,
SearchEvent.this.query.urlMask,
SearchEvent.this.query.snippetMatcher,
SearchEvent.this.query.modifier,
SearchEvent.this.query.targetlang == null ? "" : SearchEvent.this.query.targetlang,
SearchEvent.this.query.sitehash == null ? "" : SearchEvent.this.query.sitehash,

9
source/net/yacy/search/query/SnippetProcess.java
Unescape View File

@ -29,7 +29,6 @@ package net.yacy.search.query;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
import java.util.regex.Pattern;
import net.yacy.cora.document.ASCII;
import net.yacy.cora.document.Classification;
@ -365,7 +364,7 @@ public class SnippetProcess {
(this.rankingProcess.feedingIsFinished() && this.rankingProcess.sizeQueue() == 0)) {
break;
}
worker = new Worker(i, this.query.maxtime, this.query.snippetCacheStrategy, this.query.snippetMatcher, neededResults);
worker = new Worker(this.query.maxtime, this.query.snippetCacheStrategy, neededResults);
worker.start();
this.workerThreads[i] = worker;
if (this.rankingProcess.expectMoreRemoteReferences()) {
@ -387,7 +386,7 @@ public class SnippetProcess {
break;
}
if (this.workerThreads[i] == null || !this.workerThreads[i].isAlive()) {
worker = new Worker(i, this.query.maxtime, this.query.snippetCacheStrategy, this.query.snippetMatcher, neededResults);
worker = new Worker(this.query.maxtime, this.query.snippetCacheStrategy, neededResults);
worker.start();
this.workerThreads[i] = worker;
deployCount--;
@ -437,14 +436,12 @@ public class SnippetProcess {
private long lastLifeSign; // when the last time the run()-loop was executed
private final CacheStrategy cacheStrategy;
private final int neededResults;
private final Pattern snippetPattern;
private boolean shallrun;
private final SolrConnector solr;
public Worker(final int id, final long maxlifetime, final CacheStrategy cacheStrategy, final Pattern snippetPattern, final int neededResults) {
public Worker(final long maxlifetime, final CacheStrategy cacheStrategy, final int neededResults) {
this.cacheStrategy = cacheStrategy;
this.lastLifeSign = System.currentTimeMillis();
this.snippetPattern = snippetPattern;
this.timeout = System.currentTimeMillis() + Math.max(1000, maxlifetime);
this.neededResults = neededResults;
this.shallrun = true;

Write Preview
Loading…
Cancel
Save