Add Flag to enable cpu profiling

README.md

@@ -3,12 +3,14 @@ Because I couldn't find a good program for my usecase, I wrote this simple Go pr
 
 ## Installation
 
-If you have go installed, the easiest way to install is `go get`:
+If you have go installed, the easiest way to install and update is `go get`:
 
 ```
-go get "github.com/JaCoB1123/dupe-finder"
+go get -u "github.com/JaCoB1123/dupe-finder"
 ```
 
+Otherwise you can download the latest binary from the [releases](https://github.com/JaCoB1123/dupe-finder/releases) page.
+
 ## Usage
 
 dupe-finder supports the following options:
@@ -34,15 +36,15 @@ dupe-finder supports the following options:
 Find all duplicate files in `~/` and save the results to `dupes.json`
 ```
 > dupe-finder --to-file dupes.json ~/
-``̀`
+```
 
 Load previous results from `dupes.json` and delete all duplicates located in ~/.cache
 ```
 > dupe-finder --from-file dupes.json --delete-dupes-in ~/.cache
-``̀`
+```
 
 Find all duplicate files in `~/' and `/mnt/EXT`. Prompt which file to keep for each set of duplicates and move the others to /dupes/.
 ```
 > dupe-finder --delete-prompt --move-files /dupes/ ~/ /mnt/EXT
-``̀`
+```
 

file.go

@@ -0,0 +1,57 @@
+package main
+
+import (
+	"crypto/sha1"
+	"encoding/base64"
+	"io"
+	"os"
+	"path/filepath"
+	"strconv"
+)
+
+func remove(path string) {
+	if !*force {
+		return
+	}
+
+	if *moveToFolder == "" {
+		os.Remove(path)
+		return
+	}
+
+	moveButDontOvewrite(path, *moveToFolder)
+}
+
+func moveButDontOvewrite(path string, targetPath string) {
+	num := 0
+
+	filename := filepath.Base(path)
+
+	target := filepath.Join(targetPath, filename)
+
+	for {
+		_, err := os.Stat(target)
+		if os.IsNotExist(err) {
+			os.Rename(path, target)
+			return
+		}
+
+		target = filepath.Join(targetPath, filename+"."+strconv.Itoa(num))
+		num++
+	}
+}
+
+func calculateHash(path string) (string, error) {
+	f, err := os.Open(path)
+	if err != nil {
+		return "", err
+	}
+	defer f.Close()
+
+	h := sha1.New()
+	if _, err := io.Copy(h, f); err != nil {
+		return "", err
+	}
+
+	return base64.RawStdEncoding.EncodeToString(h.Sum(nil)), nil
+}

filesmap.go

@@ -0,0 +1,110 @@
+package main
+
+import (
+	"flag"
+	"fmt"
+	"log"
+	"os"
+	"path/filepath"
+	"sync"
+)
+
+// 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
+
+	FilesHashing chan fileEntry
+
+	FilesIncoming chan fileEntry
+
+	FilesHashed chan fileEntry
+
+	lock sync.Mutex
+}
+
+func newFilesMap() *FilesMap {
+	return &FilesMap{
+		FilesBySize:   map[int64]string{},
+		FilesByHash:   map[string][]string{},
+		FilesHashed:   make(chan fileEntry),
+		FilesIncoming: make(chan fileEntry),
+		FilesHashing:  make(chan fileEntry),
+	}
+}
+
+func (fm *FilesMap) IncomingWorker() {
+	for file := range fm.FilesIncoming {
+		if *verbose {
+			fmt.Println("Incoming", file.path)
+		}
+
+		prevFile, ok := fm.FilesBySize[file.size]
+		if !ok {
+			fm.FilesBySize[file.size] = file.path
+			continue
+		}
+
+		if prevFile != "" {
+			fm.FilesHashing <- fileEntry{prevFile, file.size, ""}
+		}
+
+		fm.FilesBySize[file.size] = ""
+
+		fm.FilesHashing <- file
+	}
+	close(fm.FilesHashing)
+}
+
+func (fm *FilesMap) HashingWorker(wg *sync.WaitGroup) {
+	for file := range fm.FilesHashing {
+		if *verbose {
+			fmt.Println("Hashing", file.path)
+		}
+
+		hash, err := calculateHash(file.path)
+
+		if err != nil {
+			log.Printf("Error calculating Hash for %s: %v\n", file, err)
+			continue
+		}
+
+		file.hash = hash
+		fm.FilesHashed <- file
+	}
+	wg.Done()
+}
+
+func (fm *FilesMap) HashedWorker(done chan bool) {
+	for file := range fm.FilesHashed {
+		if *verbose {
+			fmt.Println("Finishing", file.path)
+		}
+
+		fm.lock.Lock()
+		if _, ok := fm.FilesByHash[file.hash]; ok {
+			fm.FilesByHash[file.hash] = append(fm.FilesByHash[file.hash], file.path)
+		} else {
+			fm.FilesByHash[file.hash] = []string{file.path}
+		}
+		fm.lock.Unlock()
+	}
+
+	done <- true
+}
+
+func (fm *FilesMap) WalkDirectories() {
+	for _, path := range flag.Args() {
+		filepath.Walk(path, func(path string, info os.FileInfo, err error) error {
+			if info.IsDir() {
+				return nil
+			}
+
+			fm.FilesIncoming <- fileEntry{path, info.Size(), ""}
+			return nil
+		})
+	}
+
+	close(fm.FilesIncoming)
+}

main.go

@@ -2,17 +2,18 @@ package main
 
 import (
 	"bufio"
-	"crypto/sha256"
-	"encoding/base64"
 	"encoding/json"
 	"flag"
 	"fmt"
-	"io"
 	"io/ioutil"
+	"log"
 	"os"
 	"path/filepath"
+	"runtime"
+	"runtime/pprof"
 	"strconv"
 	"strings"
+	"sync"
 )
 
 var fromFile = flag.String("from-file", "", "Load results file from <path>")
@@ -22,74 +23,59 @@ var promptForDelete = flag.Bool("delete-prompt", false, "Ask which file to keep
 var moveToFolder = flag.String("move-files", "", "Move files to <path> instead of deleting them")
 var force = flag.Bool("force", false, "Actually delete files. Without this options, the files to be deleted are only printed")
 var verbose = flag.Bool("verbose", false, "Output additional information")
-
-func Delete(path string) {
-	if !*force {
-		return
-	}
-
-	if *moveToFolder == "" {
-		os.Remove(path)
-		return
-	}
-
-	MoveButDontOvewrite(path, *moveToFolder)
-}
-
-func MoveButDontOvewrite(path string, targetPath string) {
-	num := 0
-
-	filename := filepath.Base(path)
-
-	target := filepath.Join(targetPath, filename)
-
-	for {
-		_, err := os.Stat(target)
-		if os.IsNotExist(err) {
-			os.Rename(path, target)
-			return
-		}
-
-		target = filepath.Join(targetPath, filename+"."+strconv.Itoa(num))
-		num++
-	}
-}
+var cpuprofile = flag.String("cpuprofile", "", "write cpu profile to file")
 
 func main() {
 	flag.Parse()
 
+	if *cpuprofile != "" {
+		f, err := os.Create(*cpuprofile)
+		if err != nil {
+			log.Fatal(err)
+		}
+		pprof.StartCPUProfile(f)
+		defer pprof.StopCPUProfile()
+	}
+
 	if *verbose {
 		printConfiguration()
 	}
 
 	filesMap := newFilesMap()
 	if *fromFile != "" {
-		fmt.Println("Loading file", *fromFile)
-
 		byteValue, _ := ioutil.ReadFile(*fromFile)
-		err := json.Unmarshal(byteValue, &filesMap.FilesBySize)
+		err := json.Unmarshal(byteValue, &filesMap.FilesByHash)
 		if err != nil {
 			panic(err)
 		}
 	} else {
-		for _, path := range flag.Args() {
-			filepath.Walk(path, func(path string, info os.FileInfo, err error) error {
-				filesMap.Add(path, info)
-				return nil
-			})
+		done := make(chan bool)
+		wg := sync.WaitGroup{}
+		for i := 0; i < runtime.GOMAXPROCS(0); i++ {
+			wg.Add(1)
+			go filesMap.HashingWorker(&wg)
 		}
+
+		go filesMap.IncomingWorker()
+
+		go filesMap.HashedWorker(done)
+
+		filesMap.WalkDirectories()
+
+		wg.Wait()
+		close(filesMap.FilesHashed)
+		<-done
 	}
 
 	if *toFile != "" && *fromFile == "" {
-		json, _ := json.MarshalIndent(filesMap.FilesBySize, "", "  ")
+		json, _ := json.MarshalIndent(filesMap.FilesByHash, "", "  ")
 		ioutil.WriteFile(*toFile, json, 644)
 	}
 
 	if *deleteDupesIn != "" {
 		deleteIn := filepath.Clean(*deleteDupesIn)
-		for size := range filesMap.FilesBySize {
-			for hash := range filesMap.FilesBySize[size] {
-				duplicateFiles := filesMap.FilesBySize[size][hash]
+		for hash := range filesMap.FilesByHash {
+			duplicateFiles := filesMap.FilesByHash[hash]
 			if len(duplicateFiles) <= 1 {
 				continue
 			}
@@ -98,17 +84,15 @@ func main() {
 				if strings.HasPrefix(filepath.Clean(file), deleteIn) {
 					fmt.Println("Would delete ", file)
 					if *force {
-							Delete(file)
-						}
+						remove(file)
 					}
 				}
 			}
 		}
 	} else if *promptForDelete {
 		reader := bufio.NewReader(os.Stdin)
-		for size := range filesMap.FilesBySize {
-			for hash := range filesMap.FilesBySize[size] {
-				duplicateFiles := filesMap.FilesBySize[size][hash]
+		for hash := range filesMap.FilesByHash {
+			duplicateFiles := filesMap.FilesByHash[hash]
 			if len(duplicateFiles) <= 1 {
 				continue
 			}
@@ -138,16 +122,14 @@ func main() {
 				}
 
 				if *force {
-						Delete(file)
-					}
+					remove(file)
 				}
 
 			}
 		}
 	} else {
-		for size := range filesMap.FilesBySize {
-			for hash := range filesMap.FilesBySize[size] {
-				duplicateFiles := filesMap.FilesBySize[size][hash]
+		for hash := range filesMap.FilesByHash {
+			duplicateFiles := filesMap.FilesByHash[hash]
 			if len(duplicateFiles) <= 1 {
 				continue
 			}
@@ -158,7 +140,6 @@ func main() {
 			fmt.Println()
 		}
 	}
-	}
 }
 
 func printConfiguration() {
@@ -175,81 +156,8 @@ func printConfiguration() {
 	fmt.Println()
 }
 
-// FilesMap is a struct for listing files by Size and Hash to search for duplicates
-type FilesMap struct {
-	FilesBySize map[int64]map[string][]string
-}
-
-// Add a file to the Map and calculate hash on demand
-func (fm *FilesMap) Add(path string, info os.FileInfo) error {
-	if info.IsDir() {
-		return nil
-	}
-
-	fileInfo := path
-
-	filesByHash := fm.FilesBySize[info.Size()]
-
-	// first file with same size
-	// => create new map for size
-	if filesByHash == nil {
-		filesByHash = map[string][]string{}
-		fm.FilesBySize[info.Size()] = filesByHash
-		filesByHash[""] = []string{fileInfo}
-		return nil
-	}
-
-	// second file with same size
-	// => calculate hashes for all entries
-	if _, hasEmptyHash := filesByHash[""]; hasEmptyHash {
-		err := appendByFileHash(filesByHash, fileInfo)
-		err2 := appendByFileHash(filesByHash, filesByHash[""][0])
-
-		delete(filesByHash, "")
-
-		if err != nil {
-			return err
-		}
-
-		return err2
-	}
-
-	// for later files always append by hash
-	return appendByFileHash(filesByHash, fileInfo)
-}
-
-func appendByFileHash(filesByHash map[string][]string, fileInfo string) error {
-	hash, err := calculateHash(fileInfo)
-
-	if err != nil {
-		return err
-	}
-
-	if _, ok := filesByHash[hash]; ok {
-		filesByHash[hash] = append(filesByHash[hash], fileInfo)
-	} else {
-		filesByHash[hash] = []string{fileInfo}
-	}
-	return nil
-}
-
-func newFilesMap() *FilesMap {
-	return &FilesMap{
-		FilesBySize: map[int64]map[string][]string{},
-	}
-}
-
-func calculateHash(path string) (string, error) {
-	f, err := os.Open(path)
-	if err != nil {
-		return "", err
-	}
-	defer f.Close()
-
-	h := sha256.New()
-	if _, err := io.Copy(h, f); err != nil {
-		return "", err
-	}
-
-	return base64.RawStdEncoding.EncodeToString(h.Sum(nil)), nil
+type fileEntry struct {
+	path string
+	size int64
+	hash string
 }