Mirror/MarginaliaSearch
1
0
Fork 0
mirror of https://github.com/MarginaliaSearch/MarginaliaSearch.git synced 2025-02-24 05:18:58 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Fixing LSH deduplication bug.

Browse Source
This commit is contained in:
Viktor Lofgren 2023-03-06 19:32:37 +01:00
parent 43f3380cb9
commit 384de2e54b
1 changed files with 7 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
crawl/converting-process/src/main/java/nu/marginalia/converting/processor/logic/LshDocumentDeduplicator.java
View File

@ -11,10 +11,14 @@ import java.util.List;
import java.util.Set; import java.util.Set;
import java.util.stream.Collectors; import java.util.stream.Collectors;
/** Deduplicates documents based on their LSH
*
* @see EasyLSH
*/
@Singleton @Singleton
public class LshDocumentDeduplicator { public class LshDocumentDeduplicator {
private final int DISTANCE_THRESHOLD = 4; private final int DISTANCE_THRESHOLD = 2;
private final Logger logger = LoggerFactory.getLogger(getClass()); private final Logger logger = LoggerFactory.getLogger(getClass());
public void deduplicate(List<ProcessedDocument> documents) { public void deduplicate(List<ProcessedDocument> documents) {
@ -40,13 +44,13 @@ public class LshDocumentDeduplicator {
return false; return false;
} }
if (EasyLSH.hammingDistance(thisDoc.lshHash, otherDoc.lshHash) < DISTANCE_THRESHOLD) if (EasyLSH.hammingDistance(thisDoc.details.hashCode, otherDoc.details.hashCode) > DISTANCE_THRESHOLD)
return false; return false;
if (thisDoc.url.path.length() if (thisDoc.url.path.length()
< otherDoc.url.path.length()) < otherDoc.url.path.length())
{ {
logger.info("{} duplicates {}", otherDoc.url, thisDoc.url); logger.debug("{} duplicates {}", otherDoc.url, thisDoc.url);
otherDoc.state = EdgeUrlState.DISQUALIFIED; otherDoc.state = EdgeUrlState.DISQUALIFIED;
otherDoc.stateReason = "Duplicate"; otherDoc.stateReason = "Duplicate";