backdev/no-copy
1
0
Fork 0
Code Issues Pull Requests 5 Actions Packages Projects Releases Wiki Activity

Merge branch 'NCBACK-25' into dev

Browse Source 
# Conflicts:
#	src/main/java/ru/soune/nocopy/repository/SimilarImageProjection.java
#	src/main/java/ru/soune/nocopy/service/FileSimilarityService.java
This commit is contained in:
vladp
2026-01-25 18:41:05 +07:00
parent 85449d73fd ab3eef1668
commit c97f205a17
30 changed files with 835 additions and 248 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/main/java/ru/soune/nocopy/service/FileSimilarityService.java
+73 -28
View File
@@ -5,12 +5,14 @@ import org.springframework.data.domain.Page;
import org.springframework.data.domain.PageImpl;
import org.springframework.data.domain.Pageable;
import org.springframework.stereotype.Service;
import ru.soune.nocopy.dto.file.SimilarFileResponse;
import ru.soune.nocopy.dto.file.SimilarFileDTO;
import ru.soune.nocopy.dto.file.SimilarityFilter;
import ru.soune.nocopy.repository.ImageHashRepository;
import ru.soune.nocopy.repository.ImageSimilarityRepository;
import ru.soune.nocopy.repository.SimilarImageProjection;
import ru.soune.nocopy.util.FileUtil;
import java.util.AbstractMap;
import java.util.ArrayList;
import java.util.Comparator;
import java.util.List;
@@ -19,26 +21,28 @@ import java.util.stream.Collectors;
@Service
@RequiredArgsConstructor
public class FileSimilarityService {
private final ImageSimilarityRepository repository;
private final ImageHashRepository hashRepository;
public List<SimilarFileResponse> findSimilarFiles(String fileId) {
private final FileUtil fileUtil;
public List<SimilarFileDTO> findSimilarFiles(String fileId) {
var imageHashEntity = hashRepository.findById(fileId)
.orElseThrow(() -> new RuntimeException("Hash not found"));
Integer hash64Hi = imageHashEntity.getHash64Hi();
Integer hash64Lo = imageHashEntity.getHash64Lo();
Long hash64Hi = imageHashEntity.getHash64Hi();
Long hash64Lo = imageHashEntity.getHash64Lo();
List<SimilarImageProjection> candidates =
repository.findCandidates(fileId);
List<SimilarImageProjection> candidates = repository.findCandidates(fileId);
return candidates.stream()
.map(c -> {
Integer cHi = c.getHash64Hi();
Integer cLo = c.getHash64Lo();
Long cHi = c.getHash64Hi();
Long cLo = c.getHash64Lo();
int hamming = hamming64(hash64Hi, hash64Lo, cHi, cLo);
int hamming = fileUtil.hamming64(hash64Hi, hash64Lo, cHi, cLo);
String level;
if (hamming <= 5) {
@@ -49,7 +53,7 @@ public class FileSimilarityService {
level = "DIFFERENT";
}
return SimilarFileResponse.builder()
return SimilarFileDTO.builder()
.fileId(c.getId())
.originalFileName(c.getOriginalFileName())
.fileSize(c.getFileSize())
@@ -62,7 +66,53 @@ public class FileSimilarityService {
.toList();
}
public List<SimilarImageProjection> findDuplicatedByHash(Integer hash64Hi, Integer hash64Lo) {
public List<SimilarFileDTO> findDuplicateByHammingDistance(String fileId, int hammingDistance,
int duplicate, int similar) {
List<SimilarImageProjection> candidates = repository.findCandidates(fileId);
var imageHashEntity = hashRepository.findById(fileId)
.orElseThrow(() -> new RuntimeException("Hash not found"));
Long hash64Hi = imageHashEntity.getHash64Hi();
Long hash64Lo = imageHashEntity.getHash64Lo();
return candidates.stream()
.map(c -> {
Long cHi = c.getHash64Hi();
Long cLo = c.getHash64Lo();
int hamming = fileUtil.hamming64(hash64Hi, hash64Lo, cHi, cLo);
return new AbstractMap.SimpleEntry<>(c, hamming);
})
.filter(entry -> entry.getValue() <= hammingDistance)
.sorted((a, b) -> Integer.compare(a.getValue(), b.getValue()))
.map(entry -> {
SimilarImageProjection similarImageProjection = entry.getKey();
int hamming = entry.getValue();
String level;
if (hamming <= duplicate) {
level = "DUPLICATE";
} else if (hamming <= similar) {
level = "SIMILAR";
} else {
level = "DIFFERENT";
}
return SimilarFileDTO.builder()
.fileId(similarImageProjection.getFileId())
.originalFileName(similarImageProjection.getOriginalFileName())
.fileSize(similarImageProjection.getFileSize())
.hammingDistance(hamming)
.similarityLevel(level)
.ownerId(similarImageProjection.getUserId())
.build();
})
.toList();
}
public List<SimilarImageProjection> findDuplicatedByHash(Long hash64Hi, Long hash64Lo) {
List<SimilarImageProjection> duplicates = repository.findExactDuplicates(hash64Hi,hash64Lo);
if (duplicates.isEmpty()) {
@@ -72,12 +122,12 @@ public class FileSimilarityService {
return duplicates;
}
public Page<SimilarFileResponse> findSimilarFiles(String fileId, SimilarityFilter filter, Pageable pageable) {
public Page<SimilarFileDTO> findSimilarFiles(String fileId, SimilarityFilter filter, Pageable pageable) {
var imageHashEntity = hashRepository.findById(fileId)
.orElseThrow(() -> new RuntimeException("Hash not found"));
Integer hash64Hi = imageHashEntity.getHash64Hi();
Integer hash64Lo = imageHashEntity.getHash64Lo();
Long hash64Hi = imageHashEntity.getHash64Hi();
Long hash64Lo = imageHashEntity.getHash64Lo();
List<SimilarImageProjection> candidates = repository.findCandidates(fileId);
@@ -85,10 +135,10 @@ public class FileSimilarityService {
? filter.getSimilarityLevels()
: List.of("DUPLICATE", "SIMILAR", "DIFFERENT");
List<SimilarFileResponse> allResults = candidates.stream()
List<SimilarFileDTO> allResults = candidates.stream()
.map(c -> createSimilarFileResponse(c, hash64Hi, hash64Lo))
.filter(response -> similarityLevels.contains(response.getSimilarityLevel()))
.sorted(Comparator.comparingInt(SimilarFileResponse::getHammingDistance))
.sorted(Comparator.comparingInt(SimilarFileDTO::getHammingDistance))
.collect(Collectors.toList());
int total = allResults.size();
@@ -98,17 +148,17 @@ public class FileSimilarityService {
int fromIndex = Math.min(page * size, total);
int toIndex = Math.min(fromIndex + size, total);
List<SimilarFileResponse> pageContent = allResults.subList(fromIndex, toIndex);
List<SimilarFileDTO> pageContent = allResults.subList(fromIndex, toIndex);
return new PageImpl<>(pageContent, pageable, total);
}
private SimilarFileResponse createSimilarFileResponse(SimilarImageProjection similarImageProjection,
Integer hash64Hi, Integer hash64Lo) {
Integer imageProjectionHash64Hi = similarImageProjection.getHash64Hi();
Integer similarImageProjectionHash64Lo = similarImageProjection.getHash64Lo();
private SimilarFileDTO createSimilarFileResponse(SimilarImageProjection similarImageProjection,
Long hash64Hi, Long hash64Lo) {
Long imageProjectionHash64Hi = similarImageProjection.getHash64Hi();
Long similarImageProjectionHash64Lo = similarImageProjection.getHash64Lo();
int hamming = hamming64(hash64Hi, hash64Lo, imageProjectionHash64Hi, similarImageProjectionHash64Lo);
int hamming = fileUtil.hamming64(hash64Hi, hash64Lo, imageProjectionHash64Hi, similarImageProjectionHash64Lo);
String level;
if (hamming <= 5) {
@@ -119,7 +169,7 @@ public class FileSimilarityService {
level = "DIFFERENT";
}
return SimilarFileResponse.builder()
return SimilarFileDTO.builder()
.fileId(similarImageProjection.getId())
.originalFileName(similarImageProjection.getOriginalFileName())
.fileSize(similarImageProjection.getFileSize())
@@ -127,9 +177,4 @@ public class FileSimilarityService {
.similarityLevel(level)
.build();
}
private int hamming64(int aHi, int aLo, int bHi, int bLo) {
return Integer.bitCount(aHi ^ bHi)
+ Integer.bitCount(aLo ^ bLo);
}
}