Follow naming conventions of the audio API (#7)

Names of the packages and types should follow the names
used in https://platform.openai.com/docs/api-reference/audio
and be consistent with the naming strategy in other packages.

This is a breaking change but given the low current usage,
we are breaking it now before it's impossible to break.

whisper/whisper.go → audio/audio.go

@@ -1,6 +1,6 @@
-// Package whisper implements a client for OpenAI's Whisper
+// Package audio implements a client for OpenAI's Whisper
 // audio transcriber.
-package whisper
+package audio
 
 import (
 	"context"
@@ -11,16 +11,16 @@ import (
 	"github.com/rakyll/openai-go"
 )
 
-const defaultCreateCompletionsEndpoint = "https://api.openai.com/v1/audio/transcriptions"
+const defaultCreateTranscriptionEndpoint = "https://api.openai.com/v1/audio/transcriptions"
 
 // Client is a client to communicate with Open AI's ChatGPT APIs.
 type Client struct {
 	s     *openai.Session
 	model string
 
-	// CreateCompletionsEndpoint allows overriding the default API endpoint.
+	// CreateTranscriptionEndpoint allows overriding the default API endpoint.
 	// Set this field before using the client.
-	CreateCompletionEndpoint string
+	CreateTranscriptionEndpoint string
 }
 
 // NewClient creates a new default client that uses the given session
@@ -30,24 +30,24 @@ func NewClient(session *openai.Session, model string) *Client {
 		model = "whisper-1"
 	}
 	return &Client{
-		s:                        session,
-		model:                    model,
-		CreateCompletionEndpoint: defaultCreateCompletionsEndpoint,
+		s:                           session,
+		model:                       model,
+		CreateTranscriptionEndpoint: defaultCreateTranscriptionEndpoint,
 	}
 }
 
-type CreateCompletionParams struct {
+type CreateTranscriptionParams struct {
 	Model       string
 	Language    string
 	Audio       io.Reader
 	AudioFormat string // such as "mp3" or "wav", etc.
 }
 
-type CreateCompletionResponse struct {
+type CreateTranscriptionResponse struct {
 	Text string `json:"text,omitempty"`
 }
 
-func (c *Client) Transcribe(ctx context.Context, p *CreateCompletionParams) (*CreateCompletionResponse, error) {
+func (c *Client) CreateTranscription(ctx context.Context, p *CreateTranscriptionParams) (*CreateTranscriptionResponse, error) {
 	if p.AudioFormat == "" {
 		return nil, fmt.Errorf("audio format is required")
 	}
@@ -59,6 +59,6 @@ func (c *Client) Transcribe(ctx context.Context, p *CreateCompletionParams) (*Cr
 	if p.Language != "" {
 		params.Set("language", p.Language)
 	}
-	var r CreateCompletionResponse
-	return &r, c.s.Upload(ctx, c.CreateCompletionEndpoint, p.Audio, p.AudioFormat, params, &r)
+	var r CreateTranscriptionResponse
+	return &r, c.s.Upload(ctx, c.CreateTranscriptionEndpoint, p.Audio, p.AudioFormat, params, &r)
 }

examples/whisper/main.go → examples/audio/main.go

@@ -6,12 +6,14 @@ import (
 	"os"
 
 	"github.com/rakyll/openai-go"
-	"github.com/rakyll/openai-go/whisper"
+	"github.com/rakyll/openai-go/audio"
 )
 
 func main() {
-	sesh := openai.NewSession(os.Getenv("OPENAI_API_KEY"))
-	wc := whisper.NewClient(sesh, "")
+	ctx := context.Background()
+
+	s := openai.NewSession(os.Getenv("OPENAI_API_KEY"))
+	client := audio.NewClient(s, "")
 	filePath := os.Getenv("AUDIO_FILE_PATH")
 	if filePath == "" {
 		log.Fatal("must provide an AUDIO_FILE_PATH env var")
@@ -21,7 +23,7 @@ func main() {
 		log.Fatalf("error opening audio file: %v", err)
 	}
 	defer f.Close()
-	resp, err := wc.Transcribe(context.TODO(), &whisper.CreateCompletionParams{
+	resp, err := client.CreateTranscription(ctx, &audio.CreateTranscriptionParams{
 		Language:    "en",
 		Audio:       f,
 		AudioFormat: "mp3",

examples/image/main.go

@@ -34,6 +34,6 @@ func main() {
 		if err != nil {
 			log.Fatalf("ReadAll error: %v", err)
 		}
-		log.Println(string(data[1:4]))
+		_ = data // use data
 	}
 }