Group by hamming distance

file.go

@@ -1,10 +1,7 @@
 package main
 
 import (
-	"crypto/sha1"
-	"encoding/base64"
 	"image/jpeg"
-	"io"
 	"os"
 	"path/filepath"
 	"strconv"
@@ -45,30 +42,31 @@ func moveButDontOvewrite(path string, targetPath string) {
 	}
 }
 
-func calculateHash(path string) (string, error) {
+func calculateHash(path string) (uint64, error) {
 	f, err := os.Open(path)
 	if err != nil {
-		return "", err
+		return 0, err
 	}
 	defer f.Close()
 
 	if strings.HasSuffix(path, ".jpg") {
 		img, err := jpeg.Decode(f)
 		if err != nil {
-			return "", err
+			return 0, err
 		}
 		hash, err := goimagehash.DifferenceHash(img)
 		if err != nil {
-			return "", err
+			return 0, err
 		}
 
-		return hash.ToString(), nil
+		return hash.GetHash(), nil
 	}
 
-	h := sha1.New()
+	/*	h := sha1.New()
 		if _, err := io.Copy(h, f); err != nil {
-		return "", err
+			return 0, err
 		}
 
-	return base64.RawStdEncoding.EncodeToString(h.Sum(nil)), nil
+		return base64.RawStdEncoding.EncodeToString(h.Sum(nil)), nil*/
+	return 0, nil
 }

filesmap.go

@@ -14,9 +14,7 @@ import (
 
 // FilesMap is a struct for listing files by Size and Hash to search for duplicates
 type FilesMap struct {
-	FilesBySize map[int64]string
-
-	FilesByHash map[string][]string
+	Files []fileEntry
 
 	FilesHashing chan fileEntry
 
@@ -33,8 +31,6 @@ type FilesMap struct {
 
 func newFilesMap() *FilesMap {
 	return &FilesMap{
-		FilesBySize:  map[int64]string{},
-		FilesByHash:  map[string][]string{},
 		FilesHashed:  make(chan fileEntry, 100000),
 		FilesHashing: make(chan fileEntry),
 		progress:     mpb.New(mpb.WithWidth(64)),
@@ -68,7 +64,7 @@ func (fm *FilesMap) HashedWorker(done chan bool) {
 		}
 
 		fm.lock.Lock()
-		fm.FilesByHash[file.hash] = append(fm.FilesByHash[file.hash], file.path)
+		fm.Files = append(fm.Files, file)
 		fm.lock.Unlock()
 	}
 
@@ -100,22 +96,9 @@ func (fm *FilesMap) WalkDirectories() int {
 				fmt.Println("Incoming", path)
 			}
 
-			prevFile, ok := fm.FilesBySize[size]
-			if !ok {
-				fm.FilesBySize[size] = path
-				return nil
-			}
-
-			if prevFile != "" {
-				sumSize += size
-				fm.FilesHashing <- fileEntry{prevFile, size, ""}
-			}
-
-			fm.FilesBySize[size] = ""
-
 			sumSize += size
 			fm.hashingBar.SetTotal(int64(sumSize), false)
-			fm.FilesHashing <- fileEntry{path, info.Size(), ""}
+			fm.FilesHashing <- fileEntry{path, info.Size(), 0}
 			return nil
 		})
 	}
@@ -128,5 +111,5 @@ func (fm *FilesMap) WalkDirectories() int {
 type fileEntry struct {
 	path string
 	size int64
-	hash string
+	hash uint64
 }

go.mod

@@ -11,5 +11,6 @@ require (
 	github.com/mattn/go-runewidth v0.0.13 // indirect
 	github.com/nfnt/resize v0.0.0-20180221191011-83c6a9932646 // indirect
 	github.com/rivo/uniseg v0.2.0 // indirect
+	github.com/steakknife/hamming v0.0.0-20180906055917-c99c65617cd3 // indirect
 	golang.org/x/sys v0.0.0-20210630005230-0f9fa26af87c // indirect
 )

go.sum

@@ -10,6 +10,8 @@ github.com/nfnt/resize v0.0.0-20180221191011-83c6a9932646 h1:zYyBkD/k9seD2A7fsi6
 github.com/nfnt/resize v0.0.0-20180221191011-83c6a9932646/go.mod h1:jpp1/29i3P1S/RLdc7JQKbRpFeM1dOBd8T9ki5s+AY8=
 github.com/rivo/uniseg v0.2.0 h1:S1pD9weZBuJdFmowNwbpi7BJ8TNftyUImj/0WQi72jY=
 github.com/rivo/uniseg v0.2.0/go.mod h1:J6wj4VEh+S6ZtnVlnTBMWIodfgj8LQOQFoIToxlJtxc=
+github.com/steakknife/hamming v0.0.0-20180906055917-c99c65617cd3 h1:njlZPzLwU639dk2kqnCPPv+wNjq7Xb6EfUxe/oX0/NM=
+github.com/steakknife/hamming v0.0.0-20180906055917-c99c65617cd3/go.mod h1:hpGUWaI9xL8pRQCTXQgocU38Qw1g0Us7n5PxxTwTCYU=
 github.com/vbauerster/mpb/v7 v7.0.5 h1:/CQxyoPjdlON6kqqq3Uq3UUw5tFjuBCjOmLQYaYvBmM=
 github.com/vbauerster/mpb/v7 v7.0.5/go.mod h1:emzg+wTChQAdJgyrDatWRHxji2AnmCrAemByOURuvZs=
 golang.org/x/sys v0.0.0-20210630005230-0f9fa26af87c h1:F1jZWGFhYfh0Ci55sIpILtKKK8p3i2/krTr0H1rg74I=

main.go

@@ -1,20 +1,15 @@
 package main
 
 import (
-	"bufio"
-	"encoding/json"
 	"flag"
 	"fmt"
-	"io/ioutil"
 	"log"
 	"os"
-	"path/filepath"
 	"runtime"
 	"runtime/pprof"
-	"strconv"
-	"strings"
 	"sync"
 
+	"github.com/steakknife/hamming"
 	"github.com/vbauerster/mpb/v7"
 	"github.com/vbauerster/mpb/v7/decor"
 )
@@ -48,11 +43,11 @@ func main() {
 	countFiles := 0
 	filesMap := newFilesMap()
 	if *fromFile != "" {
-		byteValue, _ := ioutil.ReadFile(*fromFile)
-		err := json.Unmarshal(byteValue, &filesMap.FilesByHash)
-		if err != nil {
-			panic(err)
-		}
+		//		byteValue, _ := ioutil.ReadFile(*fromFile)
+		//		err := json.Unmarshal(byteValue, &filesMap.FilesByHash)
+		//		if err != nil {
+		//			panic(err)
+		//		}
 	} else {
 		filesMap.incomingBar = filesMap.progress.AddSpinner(0,
 			mpb.PrependDecorators(
@@ -92,12 +87,12 @@ func main() {
 	}
 
 	if *toFile != "" && *fromFile == "" {
-		json, _ := json.MarshalIndent(filesMap.FilesByHash, "", "  ")
-		ioutil.WriteFile(*toFile, json, 0644)
+		//		json, _ := json.MarshalIndent(filesMap.FilesByHash, "", "  ")
+		//		ioutil.WriteFile(*toFile, json, 0644)
 	}
 
 	if *deleteDupesIn != "" {
-		deleteIn := filepath.Clean(*deleteDupesIn)
+		/*		deleteIn := filepath.Clean(*deleteDupesIn)
 				for hash := range filesMap.FilesByHash {
 					duplicateFiles := filesMap.FilesByHash[hash]
 					if len(duplicateFiles) <= 1 {
@@ -135,9 +130,9 @@ func main() {
 							}
 						}
 					}
-		}
+				}*/
 	} else if *promptForDelete {
-		reader := bufio.NewReader(os.Stdin)
+		/*		reader := bufio.NewReader(os.Stdin)
 				for hash := range filesMap.FilesByHash {
 					duplicateFiles := filesMap.FilesByHash[hash]
 					if len(duplicateFiles) <= 1 {
@@ -172,28 +167,44 @@ func main() {
 							remove(file)
 						}
 					}
-		}
+				}*/
 	} else {
 		countInstances := 0
 		countDupeSets := 0
-		for hash := range filesMap.FilesByHash {
-			duplicateFiles := filesMap.FilesByHash[hash]
-			if len(duplicateFiles) <= 1 {
+
+		for fileIndex := range filesMap.Files {
+			var currentCluster []fileEntry
+			file := filesMap.Files[fileIndex]
+			currentCluster = append(currentCluster, filesMap.Files[fileIndex])
+			for otherIndex := range filesMap.Files {
+				if fileIndex == otherIndex {
+					continue
+				}
+				otherFile := filesMap.Files[otherIndex]
+				var distance = hamming.Uint64(file.hash, otherFile.hash)
+				if distance > 5 {
+					continue
+				}
+
+				currentCluster = append(currentCluster, otherFile)
+			}
+
+			if len(currentCluster) <= 1 {
 				continue
 			}
 
 			countDupeSets++
-			for _, file := range duplicateFiles {
+			for _, file := range currentCluster {
 				countInstances++
-				fmt.Println(file)
+				fmt.Println(file.path)
 			}
 			fmt.Println()
 		}
 
 		fmt.Println("Statistics:")
 		fmt.Println(countFiles, "Files")
-		fmt.Println(len(filesMap.FilesBySize), "Unique Sizes")
-		fmt.Println(len(filesMap.FilesByHash), "Unique Hashes")
+		//		fmt.Println(len(filesMap.FilesBySize), "Unique Sizes")
+		//		fmt.Println(len(filesMap.FilesByHash), "Unique Hashes")
 		fmt.Println(countInstances, "Duplicate Files")
 		fmt.Println(countDupeSets, "Duplicate Sets")
 	}