go : exposed various parts to the Go Interface (#697)

This commit is contained in:
Brian Murray
2023-04-14 18:52:10 +03:00
committed by GitHub
parent 463e46338c
commit 6704a81255
4 changed files with 30 additions and 7 deletions
+12 -3
View File
@@ -111,6 +111,11 @@ func (context *context) SetMaxSegmentLength(n uint) {
context.params.SetMaxSegmentLength(int(n))
}
// Set token timestamps flag
func (context *context) SetTokenTimestamps(b bool) {
context.params.SetTokenTimestamps(b)
}
// Set max tokens per segment (0 = no limit)
func (context *context) SetMaxTokensPerSegment(n uint) {
context.params.SetMaxTokensPerSegment(int(n))
@@ -280,10 +285,14 @@ func toSegment(ctx *whisper.Context, n int) Segment {
func toTokens(ctx *whisper.Context, n int) []Token {
result := make([]Token, ctx.Whisper_full_n_tokens(n))
for i := 0; i < len(result); i++ {
data := ctx.Whisper_full_get_token_data(n, i)
result[i] = Token{
Id: int(ctx.Whisper_full_get_token_id(n, i)),
Text: strings.TrimSpace(ctx.Whisper_full_get_token_text(n, i)),
P: ctx.Whisper_full_get_token_p(n, i),
Id: int(ctx.Whisper_full_get_token_id(n, i)),
Text: ctx.Whisper_full_get_token_text(n, i),
P: ctx.Whisper_full_get_token_p(n, i),
Start: time.Duration(data.T0()) * time.Millisecond * 10,
End: time.Duration(data.T1()) * time.Millisecond * 10,
}
}
return result
+5 -3
View File
@@ -41,6 +41,7 @@ type Context interface {
SetTokenThreshold(float32) // Set timestamp token probability threshold
SetTokenSumThreshold(float32) // Set timestamp token sum probability threshold
SetMaxSegmentLength(uint) // Set max segment length in characters
SetTokenTimestamps(bool) // Set token timestamps flag
SetMaxTokensPerSegment(uint) // Set max tokens per segment (0 = no limit)
// Process mono audio data and return any errors.
@@ -85,7 +86,8 @@ type Segment struct {
// Token is a text or special token
type Token struct {
Id int
Text string
P float32
Id int
Text string
P float32
Start, End time.Duration
}