go : adding features to the go-whisper example, go ci, etc (#384)

* Updated bindings so they can be used in third pary packages.

* Updated makefiles to set FMA flag on optionally, for xeon E5 on Darwin

* Added test script

* Changes for examples

* Reverted

* Made the NewContext method private
This commit is contained in:
David Thorpe
2023-01-07 19:21:43 +00:00
committed by GitHub
parent f30b5d322c
commit f078a6f20e
10 changed files with 370 additions and 31 deletions

View File

@ -11,7 +11,7 @@ import (
wav "github.com/go-audio/wav"
)
func Process(model whisper.Model, path string, lang string, speedup, tokens bool) error {
func Process(model whisper.Model, path string, flags *Flags) error {
var data []float32
// Create processing context
@ -20,14 +20,20 @@ func Process(model whisper.Model, path string, lang string, speedup, tokens bool
return err
}
// Set the parameters
if err := flags.SetParams(context); err != nil {
return err
}
// Open the file
fmt.Fprintf(flags.Output(), "Loading %q\n", path)
fh, err := os.Open(path)
if err != nil {
return err
}
defer fh.Close()
// Decode the WAV file
// Decode the WAV file - load the full buffer
dec := wav.NewDecoder(fh)
if buf, err := dec.FullPCMBuffer(); err != nil {
return err
@ -39,42 +45,83 @@ func Process(model whisper.Model, path string, lang string, speedup, tokens bool
data = buf.AsFloat32Buffer().Data
}
// Set the parameters
// Segment callback when -tokens is specified
var cb whisper.SegmentCallback
if lang != "" {
if err := context.SetLanguage(lang); err != nil {
return err
}
}
if speedup {
context.SetSpeedup(true)
}
if tokens {
if flags.IsTokens() {
cb = func(segment whisper.Segment) {
fmt.Printf("%02d [%6s->%6s] ", segment.Num, segment.Start.Truncate(time.Millisecond), segment.End.Truncate(time.Millisecond))
fmt.Fprintf(flags.Output(), "%02d [%6s->%6s] ", segment.Num, segment.Start.Truncate(time.Millisecond), segment.End.Truncate(time.Millisecond))
for _, token := range segment.Tokens {
fmt.Printf("%q ", token.Text)
if flags.IsColorize() && context.IsText(token) {
fmt.Fprint(flags.Output(), Colorize(token.Text, int(token.P*24.0)), " ")
} else {
fmt.Fprint(flags.Output(), token.Text, " ")
}
}
fmt.Println("")
fmt.Fprintln(flags.Output(), "")
fmt.Fprintln(flags.Output(), "")
}
}
// Process the data
fmt.Fprintf(flags.Output(), " ...processing %q\n", path)
if err := context.Process(data, cb); err != nil {
return err
}
// Print out the results
switch {
case flags.GetOut() == "srt":
return OutputSRT(os.Stdout, context)
case flags.GetOut() == "none":
return nil
default:
return Output(os.Stdout, context, flags.IsColorize())
}
}
// Output text as SRT file
func OutputSRT(w io.Writer, context whisper.Context) error {
n := 1
for {
segment, err := context.NextSegment()
if err == io.EOF {
break
return nil
} else if err != nil {
return err
}
fmt.Printf("[%6s->%6s] %s\n", segment.Start.Truncate(time.Millisecond), segment.End.Truncate(time.Millisecond), segment.Text)
fmt.Fprintln(w, n)
fmt.Fprintln(w, srtTimestamp(segment.Start), " --> ", srtTimestamp(segment.End))
fmt.Fprintln(w, segment.Text)
fmt.Fprintln(w, "")
n++
}
// Return success
return nil
}
// Output text to terminal
func Output(w io.Writer, context whisper.Context, colorize bool) error {
for {
segment, err := context.NextSegment()
if err == io.EOF {
return nil
} else if err != nil {
return err
}
fmt.Fprintf(w, "[%6s->%6s]", segment.Start.Truncate(time.Millisecond), segment.End.Truncate(time.Millisecond))
if colorize {
for _, token := range segment.Tokens {
if !context.IsText(token) {
continue
}
fmt.Fprint(w, " ", Colorize(token.Text, int(token.P*24.0)))
}
fmt.Fprint(w, "\n")
} else {
fmt.Fprintln(w, " ", segment.Text)
}
}
}
// Return srtTimestamp
func srtTimestamp(t time.Duration) string {
return fmt.Sprintf("%02d:%02d:%02d,%03d", t/time.Hour, (t%time.Hour)/time.Minute, (t%time.Minute)/time.Second, (t%time.Second)/time.Millisecond)
}