Files
no-copy/src/main/java/ru/soune/nocopy/service/FileSimilarityService.java
T

136 lines
5.3 KiB
Java
Raw Normal View History

package ru.soune.nocopy.service;
import lombok.RequiredArgsConstructor;
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.SimilarityFilter;
import ru.soune.nocopy.repository.ImageHashRepository;
import ru.soune.nocopy.repository.ImageSimilarityRepository;
import ru.soune.nocopy.repository.SimilarImageProjection;
2026-01-19 15:49:27 +07:00
import java.util.ArrayList;
import java.util.Comparator;
import java.util.List;
import java.util.stream.Collectors;
@Service
@RequiredArgsConstructor
public class FileSimilarityService {
private final ImageSimilarityRepository repository;
private final ImageHashRepository hashRepository;
public List<SimilarFileResponse> findSimilarFiles(String fileId) {
var imageHashEntity = hashRepository.findById(fileId)
.orElseThrow(() -> new RuntimeException("Hash not found"));
Integer hash64Hi = imageHashEntity.getHash64Hi();
Integer hash64Lo = imageHashEntity.getHash64Lo();
List<SimilarImageProjection> candidates =
repository.findCandidates(fileId);
return candidates.stream()
.map(c -> {
Integer cHi = c.getHash64Hi();
Integer cLo = c.getHash64Lo();
int hamming = hamming64(hash64Hi, hash64Lo, cHi, cLo);
String level;
if (hamming <= 5) {
level = "DUPLICATE";
} else if (hamming <= 12) {
level = "SIMILAR";
} else {
level = "DIFFERENT";
}
return SimilarFileResponse.builder()
2026-01-24 10:59:02 +07:00
.fileId(c.getId())
.originalFileName(c.getOriginalFileName())
.fileSize(c.getFileSize())
.hammingDistance(hamming)
.similarityLevel(level)
.build();
})
.sorted((a, b) ->
Integer.compare(a.getHammingDistance(), b.getHammingDistance()))
.toList();
}
2026-01-19 15:49:27 +07:00
public List<SimilarImageProjection> findDuplicatedByHash(Integer hash64Hi, Integer hash64Lo) {
List<SimilarImageProjection> duplicates = repository.findExactDuplicates(hash64Hi,hash64Lo);
if (duplicates.isEmpty()) {
return new ArrayList<>();
}
return duplicates;
}
public Page<SimilarFileResponse> 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();
List<SimilarImageProjection> candidates = repository.findCandidates(fileId);
List<String> similarityLevels = (filter != null && filter.getSimilarityLevels() != null)
? filter.getSimilarityLevels()
: List.of("DUPLICATE", "SIMILAR", "DIFFERENT");
List<SimilarFileResponse> allResults = candidates.stream()
.map(c -> createSimilarFileResponse(c, hash64Hi, hash64Lo))
.filter(response -> similarityLevels.contains(response.getSimilarityLevel()))
.sorted(Comparator.comparingInt(SimilarFileResponse::getHammingDistance))
.collect(Collectors.toList());
int total = allResults.size();
int page = (pageable != null) ? pageable.getPageNumber() : 0;
int size = (pageable != null) ? pageable.getPageSize() : 20;
int fromIndex = Math.min(page * size, total);
int toIndex = Math.min(fromIndex + size, total);
List<SimilarFileResponse> 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();
int hamming = hamming64(hash64Hi, hash64Lo, imageProjectionHash64Hi, similarImageProjectionHash64Lo);
String level;
if (hamming <= 5) {
level = "DUPLICATE";
} else if (hamming <= 12) {
level = "SIMILAR";
} else {
level = "DIFFERENT";
}
return SimilarFileResponse.builder()
2026-01-24 10:59:02 +07:00
.fileId(similarImageProjection.getId())
.originalFileName(similarImageProjection.getOriginalFileName())
.fileSize(similarImageProjection.getFileSize())
.hammingDistance(hamming)
.similarityLevel(level)
.build();
}
private int hamming64(int aHi, int aLo, int bHi, int bLo) {
return Integer.bitCount(aHi ^ bHi)
+ Integer.bitCount(aLo ^ bLo);
}
}