From 3f99abb502d373cb4b520b212afd4be9ed4da07f Mon Sep 17 00:00:00 2001
From: Grail Finder <wohilas@gmail.com>
Date: Sun, 26 Feb 2023 14:48:15 +0600
Subject: Feat: use md5 hash to save on space with shorter names

---
 ffmpeg.go    | 14 ++++++++++++--
 main.go      | 37 +++++++++++++++++++++++--------------
 main_test.go | 26 ++++++++++++++++++++++++++
 workers.go   |  3 ++-
 4 files changed, 63 insertions(+), 17 deletions(-)

diff --git a/ffmpeg.go b/ffmpeg.go
index 7c5b6fd..4dead58 100644
--- a/ffmpeg.go
+++ b/ffmpeg.go
@@ -6,6 +6,16 @@ import (
 	ffmpeg "github.com/u2takey/ffmpeg-go"
 )
 
+func buildFFmpegCall(ut *Utterance) string {
+	return fmt.Sprintf(
+		`yes no | ffmpeg -i "%s" -ss %s -to %s \
+		-metadata text_source="%s" \
+		-ar 22050 "%s"`,
+		ut.FD.AudioPath, ut.LeftTime, ut.RightTime,
+		ut.FD.VttPath, ut.OutPath,
+	)
+}
+
 func cutoutClipAndTranscode(ut *Utterance) error {
 	err := ffmpeg.Input(ut.FD.AudioPath,
 		ffmpeg.KwArgs{
@@ -19,9 +29,9 @@ func cutoutClipAndTranscode(ut *Utterance) error {
 	return err
 }
 
-func cutOnEqualParts(filepath, segment string) error {
+func cutOnEqualParts(filepath, outname, segment string) error {
 	err := ffmpeg.Input(filepath).
-		Output(filepath+"%03d.opus",
+		Output(outname+"%02d.opus",
 			ffmpeg.KwArgs{
 				"c":                "copy",
 				"map":              0,
diff --git a/main.go b/main.go
index fe7ac06..4cfa197 100644
--- a/main.go
+++ b/main.go
@@ -1,6 +1,7 @@
 package main
 
 import (
+	"crypto/md5"
 	"flag"
 	"fmt"
 	"os"
@@ -12,7 +13,6 @@ import (
 
 const (
 	subExt          = ".vtt"
-	outdir          = "/mnt/nvme/data"
 	ffCmdOut        = "data/ff_commands"
 	timeSep         = "-->"
 	metadataPath    = "data/metadata.json"
@@ -20,6 +20,10 @@ const (
 	segmentSize     = "00:08:00"
 )
 
+var (
+	outdir = "/mnt/nvme/data"
+)
+
 type Utterance struct {
 	LeftTime  string
 	RightTime string
@@ -59,6 +63,16 @@ func mapToCSV(req map[string]string) [][]string {
 	return resp
 }
 
+func hashStr(req string) string {
+	return fmt.Sprintf("%x", md5.Sum([]byte(req)))
+}
+
+func fpathToOutname(req string) string {
+	basename := filepath.Base(req)
+	h := hashStr(basename)
+	return path.Join(outdir, h+".opus")
+}
+
 func linesToUtterances(lines []string, fd *FileData) []*Utterance {
 	resp := []*Utterance{}
 	for i, line := range lines {
@@ -107,16 +121,6 @@ func adequateTimes(left, right string) bool {
 	return false
 }
 
-func buildFFmpegCall(ut *Utterance) string {
-	return fmt.Sprintf(
-		`yes no | ffmpeg -i "%s" -ss %s -to %s \
-		-metadata text_source="%s" \
-		-ar 22050 "%s"`,
-		ut.FD.AudioPath, ut.LeftTime, ut.RightTime,
-		ut.FD.VttPath, ut.OutPath,
-	)
-}
-
 func utterancesToFileTextMap(utterances []*Utterance) map[string]string {
 	resp := make(map[string]string)
 	for _, ut := range utterances {
@@ -163,7 +167,7 @@ func getFileList(dirpath string, filter string) []string {
 	return resp
 }
 
-func equalSlice(dirpath string) {
+func equalSliceRun(dirpath string) {
 	auFiles := getFileList(dirpath, "opus")
 
 	fQueue := make(chan string, len(auFiles))
@@ -171,7 +175,7 @@ func equalSlice(dirpath string) {
 		fQueue <- fpath
 	}
 
-	workers := 3
+	workers := 20
 	for i := 0; i < workers; i++ {
 		go cutterQueue(fQueue, i)
 	}
@@ -190,17 +194,22 @@ func equalSlice(dirpath string) {
 func main() {
 	vttFilepath := flag.String("f", "", "path to a vtt file")
 	vttDir := flag.String("d", "", "path to a vtt dir")
+	outDirArg := flag.String("o", "", "output dir")
 	sliceAudioDir := flag.String("slice-audio-dir", "",
 		"for equal segmentation only without subs")
 	flag.Parse()
 
+	if outDirArg != nil && *outDirArg != "" {
+		outdir = *outDirArg
+	}
+
 	utterances := []*Utterance{}
 	if vttDir != nil && *vttDir != "" {
 		utterances = dirRun(*vttDir)
 	} else if vttFilepath != nil && *vttFilepath != "" {
 		utterances = oneFileRun(*vttFilepath)
 	} else if sliceAudioDir != nil && *sliceAudioDir != "" {
-		equalSlice(*sliceAudioDir)
+		equalSliceRun(*sliceAudioDir)
 		return
 	} else {
 		fmt.Println("no flags provided;")
diff --git a/main_test.go b/main_test.go
index e494dc6..83a1649 100644
--- a/main_test.go
+++ b/main_test.go
@@ -63,5 +63,31 @@ func TestFullyIncludes(t *testing.T) {
 			}
 		})
 	}
+}
 
+func TestHashStr(t *testing.T) {
+	cases := []struct {
+		Input       string
+		Want        string
+		Description string
+	}{
+		{
+			Input:       "Let's Play Re：Kinder [Rpg Maker Horror] #5 - Im Meer der traurigen Erinnerungen [iU9Un035p3A].opus",
+			Want:        "b047af2cb102a2a6b15007108f99cfef",
+			Description: "sum of filename example",
+		},
+		{
+			Input:       "",
+			Want:        "d41d8cd98f00b204e9800998ecf8427e",
+			Description: "sum of empty string",
+		},
+	}
+	for i, tc := range cases {
+		t.Run(fmt.Sprintf("run: #%d; %q", i, tc.Description), func(t *testing.T) {
+			got := hashStr(tc.Input)
+			if got != tc.Want {
+				t.Errorf("want: %v; got: %v", tc.Want, got)
+			}
+		})
+	}
 }
diff --git a/workers.go b/workers.go
index 91f9eaf..4255e33 100644
--- a/workers.go
+++ b/workers.go
@@ -28,6 +28,7 @@ func cutterQueue(fQueue chan string, workerID int) {
 			return
 		}
 		fpath := <-fQueue
-		cutOnEqualParts(fpath, segmentSize)
+		outname := fpathToOutname(fpath)
+		cutOnEqualParts(fpath, outname, segmentSize)
 	}
 }
-- 
cgit v1.2.3