boxtec
/
yacy_search_server
mirror of https://github.com/yacy/yacy_search_server
4
1
Fork 0
Code Issues Projects Releases Wiki Activity
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.
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from '5f4a6892c1'
${ noResults }
yacy_search_server/source/net/yacy/search/schema/WebgraphConfiguration.java

435 lines
26 KiB
Raw Normal View History Unescape

introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
/**
* WebgraphConfiguration
* Copyright 2011 by Michael Peter Christen
* First released 14.04.2011 at http://yacy.net
*
* $LastChangedDate: 2011-04-14 22:05:04 +0200 (Do, 14 Apr 2011) $
* $LastChangedRevision: 7654 $
* $LastChangedBy: orbiter $
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library 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
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public License
* along with this program in the file lgpl21.txt
* If not, see <http://www.gnu.org/licenses/>.
*/
package net.yacy.search.schema;
import java.io.File;
import java.io.IOException;
import java.io.Serializable;
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
import java.util.ArrayList;
added all clickdepth computations for source and target paths in webstructure core
12 years ago
import java.util.Collection;
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
import java.util.Date;
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
import java.util.Iterator;
added clickdepth field writing for webgraph core (unfinished)
12 years ago
import java.util.LinkedHashSet;
import java.util.List;
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
import java.util.Map;
import java.util.Set;
added all clickdepth computations for source and target paths in webstructure core
12 years ago
import java.util.concurrent.BlockingQueue;
added a feature to set a collection for a crawl result based on a regular expression on th url: the collection attribut for a crawl start may be now either a token or a list of tokens, seperated by ',' where a token is either a string or a pair <string,pattern> where the string is separated to the pattern with a ':' and the string is assigned to the document as collection only if the pattern matches with the url.
12 years ago
import java.util.regex.Pattern;
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
Merge branch 'master' of ssh://git@gitorious.org/yacy/rc1.git Conflicts: htroot/IndexFederated_p.html source/net/yacy/cora/federate/solr/YaCySchema.java source/net/yacy/peers/Protocol.java source/net/yacy/search/Switchboard.java source/net/yacy/search/index/Segment.java also moved portalsearch-dev to yacy-portalsearch to be able to fix problems with new attachment to solr of the search widget
12 years ago
import org.apache.solr.common.SolrDocument;
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
import org.apache.solr.common.SolrInputDocument;
- the webgraph shall store all links which appear on a web page and not all unique links! This made it necessary, that a large portion of the parser and link processing classes must be adopted to carry a different type of link collection which carry a property attribute which are attached to web anchors. - introduction of a new URL class, AnchorURL - the other url classes, DigestURI and MultiProtocolURI had been renamed and refactored to fit into a new document package schema, document.id - cleanup of net.yacy.cora.document package and refactoring
12 years ago
import net.yacy.cora.document.encoding.ASCII;
import net.yacy.cora.document.id.AnchorURL;
import net.yacy.cora.document.id.DigestURL;
import net.yacy.cora.document.id.MultiProtocolURL;
added clickdepth field writing for webgraph core (unfinished)
12 years ago
import net.yacy.cora.federate.solr.ProcessType;
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
import net.yacy.cora.federate.solr.SchemaConfiguration;
import net.yacy.cora.federate.solr.SchemaDeclaration;
added all clickdepth computations for source and target paths in webstructure core
12 years ago
import net.yacy.cora.federate.solr.connector.AbstractSolrConnector;
import net.yacy.cora.federate.solr.connector.SolrConnector;
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
import net.yacy.cora.protocol.Domains;
import net.yacy.cora.protocol.ResponseHeader;
import net.yacy.cora.util.CommonPattern;
- refactoring of log to ConcurrentLog: jdk-based logger tend to block at java.util.logging.Logger.log(Logger.java:476) in concurrent environments. This makes logging a main performance issue. To overcome this problem, this is a add-on to jdk logging to put log entries on a concurrent message queue and log the messages one by one using a separate process. - FTPClient uses the concurrent logging instead of the log4j logger
12 years ago
import net.yacy.cora.util.ConcurrentLog;
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
import net.yacy.document.parser.html.ImageEntry;
added all clickdepth computations for source and target paths in webstructure core
12 years ago
import net.yacy.search.index.Segment;
enhanced and fixed postprocessing
11 years ago
import net.yacy.search.index.Segment.ClickdepthCache;
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
public class WebgraphConfiguration extends SchemaConfiguration implements Serializable {
private static final long serialVersionUID=-499100932212840385L;
/**
* initialize with an empty ConfigurationSet which will cause that all the index
* attributes are used
*/
- generalized SchemaConfiguration into super-class Configuration and adopted other classes which used the configuration-only access for that class - removed many warnings - adjusted logging
12 years ago
public WebgraphConfiguration(boolean lazy) {
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
super();
- generalized SchemaConfiguration into super-class Configuration and adopted other classes which used the configuration-only access for that class - removed many warnings - adjusted logging
12 years ago
this.lazy = lazy;
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
}
/**
* initialize the schema with a given configuration file
* the configuration file simply contains a list of lines with keywords
* or keyword = value lines (while value is a custom Solr field name
* @param configurationFile
- generalized SchemaConfiguration into super-class Configuration and adopted other classes which used the configuration-only access for that class - removed many warnings - adjusted logging
12 years ago
* @throws IOException
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
*/
- generalized SchemaConfiguration into super-class Configuration and adopted other classes which used the configuration-only access for that class - removed many warnings - adjusted logging
12 years ago
public WebgraphConfiguration(final File configurationFile, boolean lazy) throws IOException {
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
super(configurationFile);
this.lazy = lazy;
// check consistency: compare with YaCyField enum
if (this.isEmpty()) return;
Iterator<Entry> it = this.entryIterator();
for (SchemaConfiguration.Entry etr = it.next(); it.hasNext(); etr = it.next()) {
try {
WebgraphSchema f = WebgraphSchema.valueOf(etr.key());
f.setSolrFieldName(etr.getValue());
Added 'final' for all exception blocks as this helps the Java compiler to optimize memory usage Conflicts: source/net/yacy/search/Switchboard.java
12 years ago
} catch (final IllegalArgumentException e) {
- refactoring of log to ConcurrentLog: jdk-based logger tend to block at java.util.logging.Logger.log(Logger.java:476) in concurrent environments. This makes logging a main performance issue. To overcome this problem, this is a add-on to jdk logging to put log entries on a concurrent message queue and log the messages one by one using a separate process. - FTPClient uses the concurrent logging instead of the log4j logger
12 years ago
ConcurrentLog.fine("SolrWebgraphWriter", "solr schema file " + configurationFile.getAbsolutePath() + " defines unknown attribute '" + etr.toString() + "'");
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
it.remove();
}
}
// check consistency the other way: look if all enum constants in SolrField appear in the configuration file
for (SchemaDeclaration field: WebgraphSchema.values()) {
if (this.get(field.name()) == null) {
- refactoring of log to ConcurrentLog: jdk-based logger tend to block at java.util.logging.Logger.log(Logger.java:476) in concurrent environments. This makes logging a main performance issue. To overcome this problem, this is a add-on to jdk logging to put log entries on a concurrent message queue and log the messages one by one using a separate process. - FTPClient uses the concurrent logging instead of the log4j logger
12 years ago
ConcurrentLog.warn("SolrWebgraphWriter", " solr schema file " + configurationFile.getAbsolutePath() + " is missing declaration for '" + field.name() + "'");
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
}
}
}
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
public static class Subgraph {
added the anchor text within web pages to the searcheable entities of a web page. This can be of benefit for the ranking if these fields are used for boosts.
11 years ago
public final ArrayList<String>[] urlProtocols, urlStubs, urlAnchorTexts;
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
public final ArrayList<SolrInputDocument> edges;
@SuppressWarnings("unchecked")
public Subgraph(int inboundSize, int outboundSize) {
this.urlProtocols = new ArrayList[]{new ArrayList<String>(inboundSize), new ArrayList<String>(outboundSize)};
this.urlStubs = new ArrayList[]{new ArrayList<String>(inboundSize), new ArrayList<String>(outboundSize)};
added the anchor text within web pages to the searcheable entities of a web page. This can be of benefit for the ranking if these fields are used for boosts.
11 years ago
this.urlAnchorTexts = new ArrayList[]{new ArrayList<String>(inboundSize), new ArrayList<String>(outboundSize)};
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
this.edges = new ArrayList<SolrInputDocument>(inboundSize + outboundSize);
}
}
added url_file_name_s in default collection schema for the file name without the file extension. This part of the file path is removed from the multi-field url_paths_sxt, which has now not the file name as last part of the path list. The same applies to the new fields source_file_name_s and target_file_name_s in the webgraph schema.
12 years ago
public void addEdges(
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
final Subgraph subgraph,
- the webgraph shall store all links which appear on a web page and not all unique links! This made it necessary, that a large portion of the parser and link processing classes must be adopted to carry a different type of link collection which carry a property attribute which are attached to web anchors. - introduction of a new URL class, AnchorURL - the other url classes, DigestURI and MultiProtocolURI had been renamed and refactored to fit into a new document package schema, document.id - cleanup of net.yacy.cora.document package and refactoring
12 years ago
final DigestURL source, final ResponseHeader responseHeader, Map<String, Pattern> collections, int clickdepth_source,
final List<ImageEntry> images, final boolean inbound, final Collection<AnchorURL> links,
Patch the citation index for links with canonical tags. This shall fulfill the following requirement: If a document A links to B and B contains a 'canonical C', then the citation rank computation shall consider that A links to C and B does not link to C. To do so, we first must collect all canonical links, find all references to them, get the anchor list of the documents and patch the citation reference of these links.
11 years ago
final String sourceName) {
added url_file_name_s in default collection schema for the file name without the file extension. This part of the file path is removed from the multi-field url_paths_sxt, which has now not the file name as last part of the path list. The same applies to the new fields source_file_name_s and target_file_name_s in the webgraph schema.
12 years ago
boolean allAttr = this.isEmpty();
fixed a NPE in surrogat processing
11 years ago
boolean generalNofollow = responseHeader == null ? false : responseHeader.get("X-Robots-Tag", "").indexOf("nofollow") >= 0;
enhanced webgraph processing
11 years ago
int target_order = 0;
- the webgraph shall store all links which appear on a web page and not all unique links! This made it necessary, that a large portion of the parser and link processing classes must be adopted to carry a different type of link collection which carry a property attribute which are attached to web anchors. - introduction of a new URL class, AnchorURL - the other url classes, DigestURI and MultiProtocolURI had been renamed and refactored to fit into a new document package schema, document.id - cleanup of net.yacy.cora.document package and refactoring
12 years ago
for (final AnchorURL target_url: links) {
enhanced webgraph processing
11 years ago
SolrInputDocument edge = getEdge(
subgraph, source, responseHeader, collections, clickdepth_source, images, inbound,
sourceName, allAttr, generalNofollow, target_order, target_url);
target_order++;
// add the edge to the subgraph
subgraph.edges.add(edge);
}
}
public SolrInputDocument getEdge(
final Subgraph subgraph,
final DigestURL source, final ResponseHeader responseHeader, Map<String, Pattern> collections, int clickdepth_source,
final List<ImageEntry> images, final boolean inbound,
final String sourceName, boolean allAttr, boolean generalNofollow, int target_order, AnchorURL target_url) {
added clickdepth field writing for webgraph core (unfinished)
12 years ago
enhanced webgraph processing
11 years ago
Set<ProcessType> processTypes = new LinkedHashSet<ProcessType>();
final String name = target_url.getNameProperty(); // the name attribute
final String text = target_url.getTextProperty(); // the text between the <a></a> tag
String rel = target_url.getRelProperty(); // the rel-attribute
int ioidx = inbound ? 0 : 1;
if (generalNofollow) {
// patch the rel attribute since the header makes nofollow valid for all links
if (rel.length() == 0) rel = "nofollow"; else if (rel.indexOf("nofollow") < 0) rel += ",nofollow";
}
// index organization
StringBuilder idi = new StringBuilder(8);
idi.append(Integer.toHexString((name + text + rel).hashCode()).toLowerCase());
while (idi.length() < 8) idi.insert(0, '0');
String source_id = ASCII.String(source.hash());
String target_id = ASCII.String(target_url.hash());
StringBuilder id = new StringBuilder(source_id).append(target_id).append(idi);
SolrInputDocument edge = new SolrInputDocument();
add(edge, WebgraphSchema.id, id.toString());
add(edge, WebgraphSchema.target_order_i, target_order);
if (allAttr || contains(WebgraphSchema.load_date_dt)) {
Date loadDate = new Date();
Date modDate = responseHeader == null ? new Date() : responseHeader.lastModified();
if (modDate.getTime() > loadDate.getTime()) modDate = loadDate;
add(edge, WebgraphSchema.load_date_dt, loadDate);
}
if (allAttr || contains(WebgraphSchema.last_modified)) add(edge, WebgraphSchema.last_modified, responseHeader == null ? new Date() : responseHeader.lastModified());
final String source_url_string = source.toNormalform(false);
if (allAttr || contains(CollectionSchema.collection_sxt) && collections != null && collections.size() > 0) {
List<String> cs = new ArrayList<String>();
for (Map.Entry<String, Pattern> e: collections.entrySet()) {
if (e.getValue().matcher(source_url_string).matches()) cs.add(e.getKey());
added a feature to set a collection for a crawl result based on a regular expression on th url: the collection attribut for a crawl start may be now either a token or a list of tokens, seperated by ',' where a token is either a string or a pair <string,pattern> where the string is separated to the pattern with a ':' and the string is assigned to the document as collection only if the pattern matches with the url.
12 years ago
}
enhanced webgraph processing
11 years ago
add(edge, WebgraphSchema.collection_sxt, cs);
}
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
enhanced webgraph processing
11 years ago
// add the source attributes
add(edge, WebgraphSchema.source_id_s, source_id);
int pr_source = source_url_string.indexOf("://",0);
if (allAttr || contains(WebgraphSchema.source_protocol_s)) add(edge, WebgraphSchema.source_protocol_s, source_url_string.substring(0, pr_source));
if (allAttr || contains(WebgraphSchema.source_urlstub_s)) add(edge, WebgraphSchema.source_urlstub_s, source_url_string.substring(pr_source + 3));
Map<String, String> source_searchpart = source.getSearchpartMap();
if (source_searchpart == null) {
if (allAttr || contains(WebgraphSchema.source_parameter_count_i)) add(edge, WebgraphSchema.source_parameter_count_i, 0);
} else {
if (allAttr || contains(WebgraphSchema.source_parameter_count_i)) add(edge, WebgraphSchema.source_parameter_count_i, source_searchpart.size());
if (allAttr || contains(WebgraphSchema.source_parameter_key_sxt)) add(edge, WebgraphSchema.source_parameter_key_sxt, source_searchpart.keySet().toArray(new String[source_searchpart.size()]));
if (allAttr || contains(WebgraphSchema.source_parameter_value_sxt)) add(edge, WebgraphSchema.source_parameter_value_sxt, source_searchpart.values().toArray(new String[source_searchpart.size()]));
}
if (allAttr || contains(WebgraphSchema.source_chars_i)) add(edge, WebgraphSchema.source_chars_i, source_url_string.length());
String source_host = null;
if ((source_host = source.getHost()) != null) {
String dnc = Domains.getDNC(source_host);
String subdomOrga = source_host.length() - dnc.length() <= 0 ? "" : source_host.substring(0, source_host.length() - dnc.length() - 1);
int pp = subdomOrga.lastIndexOf('.');
String subdom = (pp < 0) ? "" : subdomOrga.substring(0, pp);
String orga = (pp < 0) ? subdomOrga : subdomOrga.substring(pp + 1);
if (allAttr || contains(WebgraphSchema.source_host_s)) add(edge, WebgraphSchema.source_host_s, source_host);
if (allAttr || contains(WebgraphSchema.source_host_id_s)) add(edge, WebgraphSchema.source_host_id_s, source.hosthash());
if (allAttr || contains(WebgraphSchema.source_host_dnc_s)) add(edge, WebgraphSchema.source_host_dnc_s, dnc);
if (allAttr || contains(WebgraphSchema.source_host_organization_s)) add(edge, WebgraphSchema.source_host_organization_s, orga);
if (allAttr || contains(WebgraphSchema.source_host_organizationdnc_s)) add(edge, WebgraphSchema.source_host_organizationdnc_s, orga + '.' + dnc);
if (allAttr || contains(WebgraphSchema.source_host_subdomain_s)) add(edge, WebgraphSchema.source_host_subdomain_s, subdom);
}
if (allAttr || contains(WebgraphSchema.source_file_ext_s) || contains(WebgraphSchema.source_file_name_s)) {
String source_file_name = source.getFileName();
String source_file_ext = MultiProtocolURL.getFileExtension(source_file_name);
add(edge, WebgraphSchema.source_file_name_s, source_file_name.toLowerCase().endsWith("." + source_file_ext) ? source_file_name.substring(0, source_file_name.length() - source_file_ext.length() - 1) : source_file_name);
add(edge, WebgraphSchema.source_file_ext_s, source_file_ext);
}
if (allAttr || contains(WebgraphSchema.source_path_s)) add(edge, WebgraphSchema.source_path_s, source.getPath());
if (allAttr || contains(WebgraphSchema.source_path_folders_count_i) || contains(WebgraphSchema.source_path_folders_sxt)) {
String[] paths = source.getPaths();
add(edge, WebgraphSchema.source_path_folders_count_i, paths.length);
add(edge, WebgraphSchema.source_path_folders_sxt, paths);
}
- reduce computation in case that specific postprocessing fields are not selected - de-select citation rank computation
11 years ago
if (this.contains(WebgraphSchema.source_clickdepth_i) && this.contains(WebgraphSchema.source_protocol_s) && this.contains(WebgraphSchema.source_urlstub_s) && this.contains(WebgraphSchema.source_id_s)) {
enhanced webgraph processing
11 years ago
add(edge, WebgraphSchema.source_clickdepth_i, clickdepth_source);
if (clickdepth_source < 0 || clickdepth_source > 1) processTypes.add(ProcessType.CLICKDEPTH);
}
// add the source attributes about the target
if (allAttr || contains(WebgraphSchema.target_inbound_b)) add(edge, WebgraphSchema.target_inbound_b, inbound);
if (allAttr || contains(WebgraphSchema.target_name_t)) add(edge, WebgraphSchema.target_name_t, name.length() > 0 ? name : "");
if (allAttr || contains(WebgraphSchema.target_rel_s)) add(edge, WebgraphSchema.target_rel_s, rel.length() > 0 ? rel : "");
if (allAttr || contains(WebgraphSchema.target_relflags_i)) add(edge, WebgraphSchema.target_relflags_i, relEval(rel.length() > 0 ? rel : ""));
if (allAttr || contains(WebgraphSchema.target_linktext_t)) add(edge, WebgraphSchema.target_linktext_t, text.length() > 0 ? text : "");
if (allAttr || contains(WebgraphSchema.target_linktext_charcount_i)) add(edge, WebgraphSchema.target_linktext_charcount_i, text.length());
if (allAttr || contains(WebgraphSchema.target_linktext_wordcount_i)) add(edge, WebgraphSchema.target_linktext_wordcount_i, text.length() > 0 ? CommonPattern.SPACE.split(text).length : 0);
ImageEntry ientry = null;
for (ImageEntry ie: images) {
if (ie.linkurl() != null && ie.linkurl().equals(target_url)) {ientry = ie; break;}
}
String alttext = ientry == null ? "" : ientry.alt();
if (allAttr || contains(WebgraphSchema.target_alt_t)) add(edge, WebgraphSchema.target_alt_t, alttext);
if (allAttr || contains(WebgraphSchema.target_alt_charcount_i)) add(edge, WebgraphSchema.target_alt_charcount_i, alttext.length());
if (allAttr || contains(WebgraphSchema.target_alt_wordcount_i)) add(edge, WebgraphSchema.target_alt_wordcount_i, alttext.length() > 0 ? CommonPattern.SPACE.split(alttext).length : 0);
// add the target attributes
add(edge, WebgraphSchema.target_id_s, target_id);
final String target_url_string = target_url.toNormalform(false);
int pr_target = target_url_string.indexOf("://",0);
subgraph.urlProtocols[ioidx].add(target_url_string.substring(0, pr_target));
subgraph.urlStubs[ioidx].add(target_url_string.substring(pr_target + 3));
subgraph.urlAnchorTexts[ioidx].add(text);
if (allAttr || contains(WebgraphSchema.target_protocol_s)) add(edge, WebgraphSchema.target_protocol_s, target_url_string.substring(0, pr_target));
if (allAttr || contains(WebgraphSchema.target_urlstub_s)) add(edge, WebgraphSchema.target_urlstub_s, target_url_string.substring(pr_target + 3));
Map<String, String> target_searchpart = target_url.getSearchpartMap();
if (target_searchpart == null) {
if (allAttr || contains(WebgraphSchema.target_parameter_count_i)) add(edge, WebgraphSchema.target_parameter_count_i, 0);
} else {
if (allAttr || contains(WebgraphSchema.target_parameter_count_i)) add(edge, WebgraphSchema.target_parameter_count_i, target_searchpart.size());
if (allAttr || contains(WebgraphSchema.target_parameter_key_sxt)) add(edge, WebgraphSchema.target_parameter_key_sxt, target_searchpart.keySet().toArray(new String[target_searchpart.size()]));
if (allAttr || contains(WebgraphSchema.target_parameter_value_sxt)) add(edge, WebgraphSchema.target_parameter_value_sxt, target_searchpart.values().toArray(new String[target_searchpart.size()]));
}
if (allAttr || contains(WebgraphSchema.target_chars_i)) add(edge, WebgraphSchema.target_chars_i, target_url_string.length());
String target_host = null;
if ((target_host = target_url.getHost()) != null) {
String dnc = Domains.getDNC(target_host);
String subdomOrga = target_host.length() - dnc.length() <= 0 ? "" : target_host.substring(0, target_host.length() - dnc.length() - 1);
int pp = subdomOrga.lastIndexOf('.');
String subdom = (pp < 0) ? "" : subdomOrga.substring(0, pp);
String orga = (pp < 0) ? subdomOrga : subdomOrga.substring(pp + 1);
if (allAttr || contains(WebgraphSchema.target_host_s)) add(edge, WebgraphSchema.target_host_s, target_host);
if (allAttr || contains(WebgraphSchema.target_host_id_s)) add(edge, WebgraphSchema.target_host_id_s, target_url.hosthash());
if (allAttr || contains(WebgraphSchema.target_host_dnc_s)) add(edge, WebgraphSchema.target_host_dnc_s, dnc);
if (allAttr || contains(WebgraphSchema.target_host_organization_s)) add(edge, WebgraphSchema.target_host_organization_s, orga);
if (allAttr || contains(WebgraphSchema.target_host_organizationdnc_s)) add(edge, WebgraphSchema.target_host_organizationdnc_s, orga + '.' + dnc);
if (allAttr || contains(WebgraphSchema.target_host_subdomain_s)) add(edge, WebgraphSchema.target_host_subdomain_s, subdom);
}
if (allAttr || contains(WebgraphSchema.target_file_ext_s) || contains(WebgraphSchema.target_file_name_s)) {
String target_file_name = target_url.getFileName();
String target_file_ext = MultiProtocolURL.getFileExtension(target_file_name);
add(edge, WebgraphSchema.target_file_name_s, target_file_name.toLowerCase().endsWith("." + target_file_ext) ? target_file_name.substring(0, target_file_name.length() - target_file_ext.length() - 1) : target_file_name);
add(edge, WebgraphSchema.target_file_ext_s, target_file_ext);
}
if (allAttr || contains(WebgraphSchema.target_path_s)) add(edge, WebgraphSchema.target_path_s, target_url.getPath());
if (allAttr || contains(WebgraphSchema.target_path_folders_count_i) || contains(WebgraphSchema.target_path_folders_sxt)) {
String[] paths = target_url.getPaths();
add(edge, WebgraphSchema.target_path_folders_count_i, paths.length);
add(edge, WebgraphSchema.target_path_folders_sxt, paths);
}
added clickdepth field writing for webgraph core (unfinished)
12 years ago
enhanced webgraph processing
11 years ago
if (this.contains(WebgraphSchema.target_protocol_s) && this.contains(WebgraphSchema.target_urlstub_s) && this.contains(WebgraphSchema.target_id_s)) {
if ((allAttr || contains(WebgraphSchema.target_clickdepth_i))) {
if (target_url.probablyRootURL()) {
boolean lc = this.lazy; this.lazy = false;
add(edge, WebgraphSchema.target_clickdepth_i, 0);
this.lazy = lc;
} else {
add(edge, WebgraphSchema.target_clickdepth_i, 999);
processTypes.add(ProcessType.CLICKDEPTH); // postprocessing needed; this is also needed if the depth is positive; there could be a shortcut
added clickdepth field writing for webgraph core (unfinished)
12 years ago
}
}
enhanced webgraph processing
11 years ago
}
if (allAttr || contains(WebgraphSchema.process_sxt)) {
List<String> pr = new ArrayList<String>();
for (ProcessType t: processTypes) pr.add(t.name());
add(edge, WebgraphSchema.process_sxt, pr);
if (allAttr || contains(CollectionSchema.harvestkey_s)) {
add(edge, CollectionSchema.harvestkey_s, sourceName);
added clickdepth field writing for webgraph core (unfinished)
12 years ago
}
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
}
enhanced webgraph processing
11 years ago
// return the edge
return edge;
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
}
added all clickdepth computations for source and target paths in webstructure core
12 years ago
enhanced webgraph processing
11 years ago
enhanced and fixed postprocessing
11 years ago
public int postprocessing(final Segment segment, ClickdepthCache clickdepthCache, final String harvestkey) {
do a light optimization at the end of a crawl postprocessing
12 years ago
if (!this.contains(WebgraphSchema.process_sxt)) return 0;
enhanced webgraph processing
11 years ago
if (!segment.fulltext().useWebgraph()) return 0;
enhanced and fixed postprocessing
11 years ago
SolrConnector webgraphConnector = segment.fulltext().getWebgraphConnector();
added all clickdepth computations for source and target paths in webstructure core
12 years ago
// that means we must search for those entries.
enhanced and fixed postprocessing
11 years ago
webgraphConnector.commit(true); // make sure that we have latest information that can be found
added all clickdepth computations for source and target paths in webstructure core
12 years ago
//BlockingQueue<SolrDocument> docs = index.fulltext().getSolr().concurrentQuery("*:*", 0, 1000, 60000, 10);
do not use luke requests for remote solr servers if the result is different from normal requests. This happens if the remote solr is actually a solrCloud; in such cases the luke request returns only the result of the single solr peer, not the whole cloud. also done: some refactoring.
11 years ago
String query = (harvestkey == null || !this.contains(WebgraphSchema.harvestkey_s) ? "" : WebgraphSchema.harvestkey_s.getSolrFieldName() + ":\"" + harvestkey + "\" AND ") + WebgraphSchema.process_sxt.getSolrFieldName() + AbstractSolrConnector.CATCHALL_DTERM;
extended also timeout fr webgraph postprocessing
11 years ago
BlockingQueue<SolrDocument> docs = webgraphConnector.concurrentDocumentsByQuery(query, 0, 10000000, 1800000, 100);
added all clickdepth computations for source and target paths in webstructure core
12 years ago
SolrDocument doc;
String protocol, urlstub, id;
- the webgraph shall store all links which appear on a web page and not all unique links! This made it necessary, that a large portion of the parser and link processing classes must be adopted to carry a different type of link collection which carry a property attribute which are attached to web anchors. - introduction of a new URL class, AnchorURL - the other url classes, DigestURI and MultiProtocolURI had been renamed and refactored to fit into a new document package schema, document.id - cleanup of net.yacy.cora.document package and refactoring
12 years ago
DigestURL url;
do a light optimization at the end of a crawl postprocessing
12 years ago
int proccount = 0, proccount_clickdepthchange = 0;
added all clickdepth computations for source and target paths in webstructure core
12 years ago
try {
while ((doc = docs.take()) != AbstractSolrConnector.POISON_DOCUMENT) {
// for each to-be-processed entry work on the process tag
Collection<Object> proctags = doc.getFieldValues(WebgraphSchema.process_sxt.getSolrFieldName());
fixed webgraph postprocessing (did nothing and repeated to do this...)
11 years ago
try {
SolrInputDocument sid = this.toSolrInputDocument(doc);
enhanced and fixed postprocessing
11 years ago
//boolean changed = false;
fixed webgraph postprocessing (did nothing and repeated to do this...)
11 years ago
for (Object tag: proctags) {
added all clickdepth computations for source and target paths in webstructure core
12 years ago
// switch over tag types
ProcessType tagtype = ProcessType.valueOf((String) tag);
if (tagtype == ProcessType.CLICKDEPTH) {
- reduce computation in case that specific postprocessing fields are not selected - de-select citation rank computation
11 years ago
if (this.contains(WebgraphSchema.source_clickdepth_i) && this.contains(WebgraphSchema.source_protocol_s) && this.contains(WebgraphSchema.source_urlstub_s) && this.contains(WebgraphSchema.source_id_s)) {
added all clickdepth computations for source and target paths in webstructure core
12 years ago
protocol = (String) doc.getFieldValue(WebgraphSchema.source_protocol_s.getSolrFieldName());
urlstub = (String) doc.getFieldValue(WebgraphSchema.source_urlstub_s.getSolrFieldName());
id = (String) doc.getFieldValue(WebgraphSchema.source_id_s.getSolrFieldName());
- the webgraph shall store all links which appear on a web page and not all unique links! This made it necessary, that a large portion of the parser and link processing classes must be adopted to carry a different type of link collection which carry a property attribute which are attached to web anchors. - introduction of a new URL class, AnchorURL - the other url classes, DigestURI and MultiProtocolURI had been renamed and refactored to fit into a new document package schema, document.id - cleanup of net.yacy.cora.document package and refactoring
12 years ago
url = new DigestURL(protocol + "://" + urlstub, ASCII.getBytes(id));
enhanced and fixed postprocessing
11 years ago
if (postprocessing_clickdepth(clickdepthCache, sid, url, WebgraphSchema.source_clickdepth_i, 100)) {
proccount_clickdepthchange++;
//changed = true;
}
//ConcurrentLog.info("WebgraphConfiguration", "postprocessing webgraph source id " + id + ", url=" + protocol + "://" + urlstub + ", result: " + (changed ? "changed" : "not changed"));
added all clickdepth computations for source and target paths in webstructure core
12 years ago
}
- reduce computation in case that specific postprocessing fields are not selected - de-select citation rank computation
11 years ago
if (this.contains(WebgraphSchema.target_clickdepth_i) && this.contains(WebgraphSchema.target_protocol_s) && this.contains(WebgraphSchema.target_urlstub_s) && this.contains(WebgraphSchema.target_id_s)) {
added all clickdepth computations for source and target paths in webstructure core
12 years ago
protocol = (String) doc.getFieldValue(WebgraphSchema.target_protocol_s.getSolrFieldName());
urlstub = (String) doc.getFieldValue(WebgraphSchema.target_urlstub_s.getSolrFieldName());
id = (String) doc.getFieldValue(WebgraphSchema.target_id_s.getSolrFieldName());
- the webgraph shall store all links which appear on a web page and not all unique links! This made it necessary, that a large portion of the parser and link processing classes must be adopted to carry a different type of link collection which carry a property attribute which are attached to web anchors. - introduction of a new URL class, AnchorURL - the other url classes, DigestURI and MultiProtocolURI had been renamed and refactored to fit into a new document package schema, document.id - cleanup of net.yacy.cora.document package and refactoring
12 years ago
url = new DigestURL(protocol + "://" + urlstub, ASCII.getBytes(id));
enhanced and fixed postprocessing
11 years ago
if (postprocessing_clickdepth(clickdepthCache, sid, url, WebgraphSchema.target_clickdepth_i, 100)) {
proccount_clickdepthchange++;
//changed = true;
}
//ConcurrentLog.info("WebgraphConfiguration", "postprocessing webgraph target id " + id + ", url=" + protocol + "://" + urlstub + ", result: " + (changed ? "changed" : "not changed"));
added all clickdepth computations for source and target paths in webstructure core
12 years ago
}
}
}
fixed webgraph postprocessing (did nothing and repeated to do this...)
11 years ago
// all processing steps checked, remove the processing tag
sid.removeField(WebgraphSchema.process_sxt.getSolrFieldName());
if (this.contains(WebgraphSchema.harvestkey_s)) sid.removeField(WebgraphSchema.harvestkey_s.getSolrFieldName());
// send back to index
enhanced and fixed postprocessing
11 years ago
webgraphConnector.add(sid);
fixed webgraph postprocessing (did nothing and repeated to do this...)
11 years ago
proccount++;
} catch (Throwable e1) {
wrong logger fix
11 years ago
ConcurrentLog.warn(WebgraphConfiguration.class.getName(), "postprocessing failed", e1);
added all clickdepth computations for source and target paths in webstructure core
12 years ago
}
}
do a light optimization at the end of a crawl postprocessing
12 years ago
ConcurrentLog.info("WebgraphConfiguration", "cleanup_processing: re-calculated " + proccount + " new documents, " + proccount_clickdepthchange + " clickdepth values changed.");
Added 'final' for all exception blocks as this helps the Java compiler to optimize memory usage Conflicts: source/net/yacy/search/Switchboard.java
12 years ago
} catch (final InterruptedException e) {
added all clickdepth computations for source and target paths in webstructure core
12 years ago
}
do a light optimization at the end of a crawl postprocessing
12 years ago
return proccount;
added all clickdepth computations for source and target paths in webstructure core
12 years ago
}
added the generation of 50 (!!) new solr field in the core 'webgraph'. The default schema uses only some of them and the resting search index has now the following properties: - webgraph size will have about 40 times as much entries as default index - the complete index size will increase and may be about the double size of current amount As testing showed, not much indexing performance is lost. The default index will be smaller (moved fields out of it); thus searching can be faster. The new index will cause that some old parts in YaCy can be removed, i.e. specialized webgraph data and the noload crawler. The new index will make it possible to: - search within link texts of linked but not indexed documents (about 20 times of document index in size!!) - get a very detailed link graph - enhance ranking using a complete link graph To get the full access to the new index, the API to solr has now two access points: one with attribute core=collection1 for the default search index and core=webgraph to the new webgraph search index. This is also avaiable for p2p operation but client access is not yet implemented.
12 years ago
/**
* encode a string containing attributes from anchor rel properties binary:
* bit 0: "me" contained in rel
* bit 1: "nofollow" contained in rel
* @param rel
* @return binary encoded information about rel
*/
private static int relEval(final String rels) {
int i = 0;
final String s0 = rels.toLowerCase().trim();
if ("me".equals(s0)) i += 1;
if ("nofollow".equals(s0)) i += 2;
return i;
}
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
/**
* save configuration to file and update enum SolrFields
* @throws IOException
*/
@Override
public void commit() throws IOException {
try {
super.commit();
// make sure the enum SolrField.SolrFieldName is current
Iterator<Entry> it = this.entryIterator();
for (SchemaConfiguration.Entry etr = it.next(); it.hasNext(); etr = it.next()) {
try {
SchemaDeclaration f = WebgraphSchema.valueOf(etr.key());
f.setSolrFieldName(etr.getValue());
Added 'final' for all exception blocks as this helps the Java compiler to optimize memory usage Conflicts: source/net/yacy/search/Switchboard.java
12 years ago
} catch (final IllegalArgumentException e) {
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
continue;
}
}
} catch (final IOException e) {}
}
Merge branch 'master' of ssh://git@gitorious.org/yacy/rc1.git Conflicts: htroot/IndexFederated_p.html source/net/yacy/cora/federate/solr/YaCySchema.java source/net/yacy/peers/Protocol.java source/net/yacy/search/Switchboard.java source/net/yacy/search/index/Segment.java also moved portalsearch-dev to yacy-portalsearch to be able to fix problems with new attachment to solr of the search widget
12 years ago
/**
* Convert a SolrDocument to a SolrInputDocument.
* This is useful if a document from the search index shall be modified and indexed again.
* This shall be used as replacement of ClientUtils.toSolrInputDocument because we remove some fields
* which are created automatically during the indexing process.
* @param doc the solr document
* @return a solr input document
*/
public SolrInputDocument toSolrInputDocument(SolrDocument doc) {
SolrInputDocument sid = new SolrInputDocument();
for (String name: doc.getFieldNames()) {
if (this.contains(name)) { // check each field if enabled in local Solr schema
sid.addField(name, doc.getFieldValue(name), 1.0f);
}
}
return sid;
}
introduced a second core named 'webgraph'. This core will hold the link structure, but is not filled yet. To have the opportunity of a second core, multi-core functionality had to be implemented to the deep-embedded solr: - migrated the solr_40 directory content to a subdirectory 'collection1'; the previously used default core is now called collection1 - added solr_40/webgraph subdirectory as second core - added a servlet configuration for the second core 'webgraph' in /IndexSchema_p.html - added instance handling as addition to solr connections: all solr connectors are now instances of an solr 'instance' object; this required a complete re-design of the solr embedding - migrated also caching and sharding ontop of new instance handling - migrated the search apis to handle now the access to a specific core, the default core named 'collection1' - migrated the remote solr search interface to access shards of cores; for the yacy remote search the default core is now called 'solr'; using the peer address as solr address - migrated the solr backup and restore process: old backups cannot be used after this migration! - redesign of solr instance handling in all methods which access the instances: they cannot hold copies of these instances any more; the must retrieve the actuall connection object every time they want to write to it (this solves also some bugs when switching the index/network) - added another schema 'solr.webgraph.schema', the old solr.keys.list is replaced by solr.collection.schema
12 years ago
}