Init

.gitignore

@@ -0,0 +1,2 @@
+/model
+voskcli

README.md

@@ -0,0 +1 @@
+#VoskCLI

go.mod

@@ -0,0 +1,14 @@
+module git.helcel.net/helcel/voskcli
+
+go 1.20
+
+require (
+	github.com/alphacep/vosk-api/go v0.3.46-0.20230429023501-12f29a3415e4
+	github.com/m7shapan/njson v1.0.8
+)
+
+require (
+	github.com/tidwall/gjson v1.12.1 // indirect
+	github.com/tidwall/match v1.1.1 // indirect
+	github.com/tidwall/pretty v1.2.0 // indirect
+)

go.sum

@@ -0,0 +1,13 @@
+github.com/alphacep/vosk-api/go v0.3.46-0.20230429023501-12f29a3415e4 h1:MGwc3N+btl1ayMYPbe5VyK0vifwwmgsUbDxZd9auO2Y=
+github.com/alphacep/vosk-api/go v0.3.46-0.20230429023501-12f29a3415e4/go.mod h1:9X8IJsHnFk/b1xyvjlZifo+ZL5VTAx3LW+JQce/eRcA=
+github.com/google/go-cmp v0.5.6 h1:BKbKCqvP6I+rmFHt06ZmyQtvB8xAkWdhFyr0ZUNZcxQ=
+github.com/google/go-cmp v0.5.6/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/m7shapan/njson v1.0.8 h1:5FNWjy8WhZ53Ev4ricYPtZfAv7o3/7wAKL+zhXxknvg=
+github.com/m7shapan/njson v1.0.8/go.mod h1:h3P0qmwsv8+s+cE07cXuTgtyKCrBIAEUger0hA71Fy0=
+github.com/tidwall/gjson v1.12.1 h1:ikuZsLdhr8Ws0IdROXUS1Gi4v9Z4pGqpX/CvJkxvfpo=
+github.com/tidwall/gjson v1.12.1/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
+github.com/tidwall/match v1.1.1 h1:+Ho715JplO36QYgwN9PGYNhgZvoUSc9X2c80KVTi+GA=
+github.com/tidwall/match v1.1.1/go.mod h1:eRSPERbgtNPcGhD8UCthc6PmLEQXEWd3PRB5JTxsfmM=
+github.com/tidwall/pretty v1.2.0 h1:RWIZEg2iJ8/g6fDDYzMpobmaoGh5OLl4AXtGUGPcqCs=
+github.com/tidwall/pretty v1.2.0/go.mod h1:ITEVvHYasfjBbM0u2Pg8T2nJnzm8xPwvNhhsoaGGjNU=
+golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=

voskcli.go

@@ -0,0 +1,177 @@
+package main
+
+import (
+	"fmt"
+	"io"
+	"os"
+	"os/exec"
+	"os/signal"
+	"syscall"
+	njson "github.com/m7shapan/njson"
+	vosk "github.com/alphacep/vosk-api/go"
+)
+
+func usage() {
+	fmt.Println(`Usage: voskcli [FILE...]
+	
+--mic=NAME       Specify the audio device (use 'arecord -L' to find mic name).
+--model=PATH     Path to the model to use`)
+}
+
+func fatal(a ...any) {
+	fmt.Fprintln(os.Stderr, "voskcli:", fmt.Sprint(a...))
+	os.Exit(1)
+}
+func warn(a ...any) {
+	fmt.Fprintln(os.Stderr, "voskcli: WARNING:", fmt.Sprint(a...))
+}
+
+func print(a ...any) {
+	fmt.Fprintln(os.Stdout, fmt.Sprint(a...))
+}
+
+
+func parseResults(json string) string {
+	type ResultJson struct {
+		Text string `njson:"text"`
+		Words []string `njson:"result.#.word"`
+		Confs []float64 `njson:"result.#.conf"`
+		Starts []float64 `njson:"result.#.start"`
+		Ends []float64 `njson:"result.#.end"`
+		Confidence float64 `njson:"confidence"`  // only with alternatives
+	}
+	var s struct {
+		Alternatives []ResultJson `njson:"alternatives"`
+
+		// copy paste of ResultJson
+		Text string `njson:"text"`
+		Words []string `njson:"result.#.word"`
+		Confs []float64 `njson:"result.#.conf"`
+		Starts []float64 `njson:"result.#.start"`
+		Ends []float64 `njson:"result.#.end"`
+		Confidence float64 `njson:"confidence"`  // only with alternatives
+
+		ParText string `njson:"partial"`
+		ParWords []string `njson:"partial_result.#.word"`
+		ParConfs []float64 `njson:"partial_result.#.conf"`
+		ParStarts []float64 `njson:"partial_result.#.start"`
+		ParEnds []float64 `njson:"partial_result.#.end"`
+	}
+	err := njson.Unmarshal([]byte(json), &s)
+	if err != nil {
+		panic(err)
+	}
+	print(json);
+	return s.Text
+}
+
+
+func getMic(mic string) string {
+	if mic == "" {
+		mic = "default"
+	}
+	return mic
+}
+
+func record(mic string) (io.Reader, error) {
+	cmd := exec.Command("arecord", "-q", "-fS16_LE", "-c1", "-r16000", "-D", mic)
+	cmd.Stderr = os.Stderr
+	stdout, err := cmd.StdoutPipe()
+	if err != nil {
+		return nil, err
+	}
+	err = cmd.Start()
+	if err != nil {
+		return nil, err
+	}
+	return stdout, nil
+}
+
+type Action struct {
+	Tags []string
+	Text string
+}
+
+func main() {
+	var opts struct {
+		Mic string
+		Model string
+		Verbose bool
+	}
+	var model *vosk.VoskModel
+	{
+		m := os.Getenv("VOSK_MODEL")
+		if m == "" {
+			fatal("you need to install the a vosk model and set $VOSK_MODEL")
+		}
+		if opts.Verbose {
+			fmt.Fprintln(os.Stderr, "Model: " + m)
+		}
+
+		var err error
+		model, err = vosk.NewModel(m)
+		if err != nil {
+			fatal(err)
+		}
+	}
+	defer model.Free()
+
+	var r *vosk.VoskRecognizer
+	{
+		var err error
+		r, err = vosk.NewRecognizer(model, float64(16000))
+		if err != nil {
+			panic(err)
+		}
+		r.SetWords(1)
+		r.SetMaxAlternatives(3)//3 ? 10 ?
+	}
+	defer r.Free()
+
+	var mic string
+	var audio io.Reader
+
+	mic = getMic(opts.Mic)
+	if opts.Verbose {
+		fmt.Fprintln(os.Stderr, "Microphone: " + mic)
+	}
+	var err error
+	audio, err = record(mic)
+	if err != nil {
+		fatal(err)
+	}
+
+	terminate := make(chan os.Signal, 1)
+	signal.Notify(terminate, os.Interrupt, syscall.SIGHUP, syscall.SIGINT, syscall.SIGTERM)
+
+	for {
+		select {
+			case <-terminate: return
+			default:
+		}
+		chunk := make([]byte, 4096)
+		_, err := io.ReadFull(audio, chunk)
+		if err != nil {
+			if err != io.EOF && err != io.ErrUnexpectedEOF {
+				panic(err)
+			}
+			r, err := record(mic)
+			if err != nil {
+				warn(err)
+			}
+			audio = r
+			continue
+		}
+
+		var res string
+		if  (r.AcceptWaveform(chunk) == 1) {
+			res = r.FinalResult()
+			parseResults(res);
+		} else {
+			res = r.PartialResult() // r.Result()
+			parseResults(res);
+		}
+	}
+
+	
+}