func main() {
var src, ext, reg string
flag.StringVar(&src, "src", "./", "Source directory of the corpus")
flag.StringVar(&ext, "ext", "wav", "extension of the corpus with default (wav)")
flag.StringVar(®, "reg", "", "regular expression for searching files)")
flag.Parse()
fmt.Printf("src:%s\next:%s\nreg:%s\n", src, ext, reg)
if !kaldi.DirExist(src) {
fmt.Printf("Error: No Exist src directory: %s\n", src)
return
}
pattern := "*" + reg + "*." + ext
file_cmd := kaldi.JoinArgs(
"find",
src+"/",
"-type f ",
"-iname "+pattern)
fmt.Println("Find: ", file_cmd)
output, _ := kaldi.BashOutput(file_cmd)
files := strings.Split(string(output[:]), "\n")
total_len := 0.0
count := 0
for _, file := range files {
cmd := kaldi.JoinArgs(
"sox",
strings.Trim(file, "\n "),
"-n stat",
"2>&1",
"|awk '/Length/ {print $3}'")
output, err := kaldi.BashOutput(cmd)
if err != nil {
fmt.Println("Missed file:", file)
continue
}
length, err := strconv.ParseFloat(strings.Trim(string(output[:]), "\n \t"), 64)
if err != nil {
fmt.Println("Missed file:", file)
continue
}
count++
total_len += length
}
fmt.Println("Total Count: ", len(files))
fmt.Println("Total Len(hours): ", total_len/3600)
fmt.Println("Wave Count: ", count)
fmt.Println("Missed Count: ", len(files)-count)
}
func MakeMfcc(files []string, dst string) error {
dir, _ := filepath.Abs(dst)
kaldi.InsureDir(dir)
log := filepath.Join(dir, "log")
kaldi.InsureDir(log)
kaldi.BashRun("rm " + dir + "/*.*")
jobs := make(chan string, 100)
var wg sync.WaitGroup
for i := 0; i < 10; i++ {
wg.Add(1)
go func(id int, job <-chan string) {
defer wg.Done()
for f := range jobs {
wav, _ := filepath.Abs(f)
cmd := kaldi.JoinArgs("cd ~/rev_kaldi; ./steps/make_mfcc.sh",
"--nj 8", wav, log, dir)
fmt.Println(cmd)
kaldi.BashRun(cmd)
}
}(i, jobs)
}
for _, f := range files {
jobs <- f
}
close(jobs)
wg.Wait()
return nil
}
func ConstructData(data, param string) {
cmd := kaldi.JoinArgs(
"find", param, "-type f", "-iname \"feats.scp\"")
output, _ := kaldi.BashOutput(cmd)
files := strings.Split(strings.Trim(string(output), " \n"), "\n")
if len(files) == 0 {
return
}
for _, f := range files {
dst := strings.Replace(f, "param/", "data/", 1)
kaldi.InsureDir(path.Dir(dst))
cmd := kaldi.JoinArgs("cp", f, dst)
fmt.Println(cmd)
kaldi.BashRun(cmd)
}
}
func MergeScp(dir string) error {
file := path.Join(dir, "feats.scp")
os.Rename(file, file+".bakup")
cmd := kaldi.JoinArgs(
"cat", path.Join(dir, "*.scp"),
"|sort",
"> ", file)
return kaldi.BashRun(cmd)
}
func Clear(param string) error {
keys := []string{"*cmvn*", "feats.scp*"}
for _, key := range keys {
cmd := kaldi.JoinArgs(
"find", param, "-type f", "-iname", "\""+key+"\"",
"|xargs rm")
fmt.Println(cmd)
err := kaldi.BashRun(cmd)
if err != nil {
fmt.Println(err)
}
}
return nil
}
func SyncPs(svr string) []ProcInfo {
procs := []ProcInfo{}
cmd_str := kaldi.JoinArgs("ssh",
svr,
"\"ps -u ren -o comm\"")
out, err := kaldi.BashOutput(cmd_str)
if err != nil {
// log.Println("Err:", err)
} else {
str := strings.Trim(string(out), "\n ")
ps := strings.Split(str, "\n")
for _, v := range ps {
procs = append(procs, ProcInfo{svr, v, 1})
}
}
return procs
}
func ApplyCmd(c Config) {
file, err := os.Open(c.List)
kaldi.Trace().Println("list :", c.List)
if err != nil {
kaldi.Err().Println("Open list file :", err)
return
}
defer file.Close()
scanner := bufio.NewScanner(file)
for scanner.Scan() {
item := strings.Trim(scanner.Text(), " \n")
if len(item) != 0 {
cmd := kaldi.JoinArgs(c.Cmd, c.Args, item)
err := kaldi.CpuBashRun(cmd)
if err != nil {
kaldi.Err().Println("Cmd Err:", cmd, "\t", err)
}
}
}
}