mirror of
https://github.com/infiniflow/ragflow.git
synced 2026-07-10 13:45:44 +08:00
### What problem does this PR solve? This PR implements ASR and TTS support for the ZhipuAI Go driver. The ZhipuAI model config already advertises `glm-asr-2512` as an ASR model, but the Go driver returned `zhipu, no such method` from `TranscribeAudio`. This adds the documented audio transcription endpoint suffix and sends multipart transcription requests with `model`, `stream=false`, and `file` fields. Per maintainer review, this also adds the ZhipuAI TTS endpoint suffix and implements `AudioSpeech` / `AudioSpeechWithSender` for `glm-tts`. Closes #15133 ### Type of change - [x] Bug Fix (non-breaking change which fixes an issue) - [x] New Feature (non-breaking change which adds functionality)
340 lines
9.4 KiB
Go
340 lines
9.4 KiB
Go
package models
|
|
|
|
import (
|
|
"encoding/json"
|
|
"io"
|
|
"net/http"
|
|
"net/http/httptest"
|
|
"os"
|
|
"strings"
|
|
"testing"
|
|
)
|
|
|
|
func newZhipuAIForTest(baseURL string) *ZhipuAIModel {
|
|
return NewZhipuAIModel(
|
|
map[string]string{"default": baseURL},
|
|
URLSuffix{ASR: "audio/transcriptions", TTS: "audio/speech"},
|
|
)
|
|
}
|
|
|
|
func writeZhipuAITestAudio(t *testing.T) string {
|
|
t.Helper()
|
|
|
|
file, err := os.CreateTemp(t.TempDir(), "speech-*.mp3")
|
|
if err != nil {
|
|
t.Fatalf("create temp audio: %v", err)
|
|
}
|
|
defer file.Close()
|
|
|
|
if _, err = file.WriteString("fake audio"); err != nil {
|
|
t.Fatalf("write temp audio: %v", err)
|
|
}
|
|
return file.Name()
|
|
}
|
|
|
|
func TestZhipuAITranscribeAudio(t *testing.T) {
|
|
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
|
if r.Method != http.MethodPost {
|
|
t.Errorf("method=%s", r.Method)
|
|
return
|
|
}
|
|
if r.URL.Path != "/audio/transcriptions" {
|
|
t.Errorf("path=%s", r.URL.Path)
|
|
return
|
|
}
|
|
if got := r.Header.Get("Authorization"); got != "Bearer test-key" {
|
|
t.Errorf("Authorization=%q", got)
|
|
return
|
|
}
|
|
if got := r.Header.Get("Content-Type"); !strings.HasPrefix(got, "multipart/form-data") {
|
|
t.Errorf("Content-Type=%q", got)
|
|
return
|
|
}
|
|
|
|
if err := r.ParseMultipartForm(1024 * 1024); err != nil {
|
|
t.Errorf("ParseMultipartForm: %v", err)
|
|
return
|
|
}
|
|
if got := r.FormValue("model"); got != "glm-asr-2512" {
|
|
t.Errorf("model=%q", got)
|
|
}
|
|
if got := r.MultipartForm.Value["model"]; len(got) != 1 {
|
|
t.Errorf("model values=%v", got)
|
|
}
|
|
if got := r.FormValue("stream"); got != "false" {
|
|
t.Errorf("stream=%q", got)
|
|
}
|
|
if got := r.MultipartForm.Value["stream"]; len(got) != 1 {
|
|
t.Errorf("stream values=%v", got)
|
|
}
|
|
if got := r.FormValue("prompt"); got != "previous transcript" {
|
|
t.Errorf("prompt=%q", got)
|
|
}
|
|
if got := r.FormValue("user_id"); got != "12345" {
|
|
t.Errorf("user_id=%q", got)
|
|
}
|
|
if got := r.MultipartForm.Value["hotwords"]; len(got) != 2 || got[0] != "RAGFlow" || got[1] != "ZhipuAI" {
|
|
t.Errorf("hotwords=%v", got)
|
|
}
|
|
if got := r.MultipartForm.Value["file"]; len(got) != 0 {
|
|
t.Errorf("file values=%v", got)
|
|
}
|
|
if _, _, err := r.FormFile("file"); err != nil {
|
|
t.Errorf("file field: %v", err)
|
|
return
|
|
}
|
|
|
|
_ = json.NewEncoder(w).Encode(map[string]string{"text": "hello world"})
|
|
}))
|
|
defer srv.Close()
|
|
|
|
apiKey := "test-key"
|
|
modelName := "glm-asr-2512"
|
|
file := writeZhipuAITestAudio(t)
|
|
resp, err := newZhipuAIForTest(srv.URL).TranscribeAudio(
|
|
&modelName,
|
|
&file,
|
|
&APIConfig{ApiKey: &apiKey},
|
|
&ASRConfig{Params: map[string]interface{}{
|
|
"prompt": "previous transcript",
|
|
"hotwords": []string{"RAGFlow", "ZhipuAI"},
|
|
"model": "ignored-model",
|
|
"stream": true,
|
|
"file": "ignored-file",
|
|
"user_id": 12345,
|
|
"nil_value": nil,
|
|
}},
|
|
)
|
|
if err != nil {
|
|
t.Fatalf("TranscribeAudio: %v", err)
|
|
}
|
|
if resp.Text != "hello world" {
|
|
t.Errorf("Text=%q", resp.Text)
|
|
}
|
|
}
|
|
|
|
func TestZhipuAITranscribeAudioValidation(t *testing.T) {
|
|
apiKey := "test-key"
|
|
modelName := "glm-asr-2512"
|
|
file := "speech.mp3"
|
|
|
|
tests := []struct {
|
|
name string
|
|
modelName *string
|
|
file *string
|
|
apiConfig *APIConfig
|
|
want string
|
|
}{
|
|
{name: "missing api key", modelName: &modelName, file: &file, apiConfig: &APIConfig{}, want: "api key is required"},
|
|
{name: "missing model", modelName: nil, file: &file, apiConfig: &APIConfig{ApiKey: &apiKey}, want: "model name is required"},
|
|
{name: "missing file", modelName: &modelName, file: nil, apiConfig: &APIConfig{ApiKey: &apiKey}, want: "file is required"},
|
|
}
|
|
|
|
for _, tt := range tests {
|
|
t.Run(tt.name, func(t *testing.T) {
|
|
_, err := newZhipuAIForTest("http://unused").TranscribeAudio(tt.modelName, tt.file, tt.apiConfig, nil)
|
|
if err == nil || !strings.Contains(err.Error(), tt.want) {
|
|
t.Fatalf("error=%v, want %q", err, tt.want)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestZhipuAITranscribeAudioRequiresASRSuffix(t *testing.T) {
|
|
apiKey := "test-key"
|
|
modelName := "glm-asr-2512"
|
|
file := writeZhipuAITestAudio(t)
|
|
_, err := NewZhipuAIModel(map[string]string{"default": "http://unused"}, URLSuffix{}).TranscribeAudio(
|
|
&modelName,
|
|
&file,
|
|
&APIConfig{ApiKey: &apiKey},
|
|
nil,
|
|
)
|
|
if err == nil || !strings.Contains(err.Error(), "ASR URL suffix is not configured") {
|
|
t.Fatalf("error=%v", err)
|
|
}
|
|
}
|
|
|
|
func TestZhipuAITranscribeAudioHTTPError(t *testing.T) {
|
|
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
|
w.WriteHeader(http.StatusBadRequest)
|
|
_, _ = io.WriteString(w, `{"error":"bad request"}`)
|
|
}))
|
|
defer srv.Close()
|
|
|
|
apiKey := "test-key"
|
|
modelName := "glm-asr-2512"
|
|
file := writeZhipuAITestAudio(t)
|
|
_, err := newZhipuAIForTest(srv.URL).TranscribeAudio(&modelName, &file, &APIConfig{ApiKey: &apiKey}, nil)
|
|
if err == nil || !strings.Contains(err.Error(), "ZhipuAI ASR API error: 400 Bad Request") {
|
|
t.Fatalf("error=%v", err)
|
|
}
|
|
}
|
|
|
|
func TestZhipuAIAudioSpeech(t *testing.T) {
|
|
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
|
if r.Method != http.MethodPost {
|
|
t.Errorf("method=%s", r.Method)
|
|
return
|
|
}
|
|
if r.URL.Path != "/audio/speech" {
|
|
t.Errorf("path=%s", r.URL.Path)
|
|
return
|
|
}
|
|
if got := r.Header.Get("Authorization"); got != "Bearer test-key" {
|
|
t.Errorf("Authorization=%q", got)
|
|
return
|
|
}
|
|
if got := r.Header.Get("Content-Type"); !strings.HasPrefix(got, "application/json") {
|
|
t.Errorf("Content-Type=%q", got)
|
|
return
|
|
}
|
|
|
|
var body map[string]interface{}
|
|
if err := json.NewDecoder(r.Body).Decode(&body); err != nil {
|
|
t.Errorf("Decode: %v", err)
|
|
return
|
|
}
|
|
if body["model"] != "glm-tts" {
|
|
t.Errorf("model=%v", body["model"])
|
|
}
|
|
if body["input"] != "hello" {
|
|
t.Errorf("input=%v", body["input"])
|
|
}
|
|
if body["stream"] != false {
|
|
t.Errorf("stream=%v", body["stream"])
|
|
}
|
|
if body["voice"] != "zhipu" {
|
|
t.Errorf("voice=%v", body["voice"])
|
|
}
|
|
if body["response_format"] != "mp3" {
|
|
t.Errorf("response_format=%v", body["response_format"])
|
|
}
|
|
|
|
_, _ = w.Write([]byte("audio-bytes"))
|
|
}))
|
|
defer srv.Close()
|
|
|
|
apiKey := "test-key"
|
|
modelName := "glm-tts"
|
|
content := "hello"
|
|
resp, err := newZhipuAIForTest(srv.URL).AudioSpeech(
|
|
&modelName,
|
|
&content,
|
|
&APIConfig{ApiKey: &apiKey},
|
|
&TTSConfig{Format: "mp3", Params: map[string]interface{}{
|
|
"voice": "zhipu",
|
|
"model": "ignored-model",
|
|
"input": "ignored-input",
|
|
"stream": true,
|
|
"response_format": "wav",
|
|
}},
|
|
)
|
|
if err != nil {
|
|
t.Fatalf("AudioSpeech: %v", err)
|
|
}
|
|
if string(resp.Audio) != "audio-bytes" {
|
|
t.Errorf("Audio=%q", string(resp.Audio))
|
|
}
|
|
}
|
|
|
|
func TestZhipuAIAudioSpeechWithSender(t *testing.T) {
|
|
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
|
var body map[string]interface{}
|
|
if err := json.NewDecoder(r.Body).Decode(&body); err != nil {
|
|
t.Errorf("Decode: %v", err)
|
|
return
|
|
}
|
|
if body["stream"] != true {
|
|
t.Errorf("stream=%v", body["stream"])
|
|
}
|
|
_, _ = w.Write([]byte("chunk-one"))
|
|
_, _ = w.Write([]byte("chunk-two"))
|
|
}))
|
|
defer srv.Close()
|
|
|
|
apiKey := "test-key"
|
|
modelName := "glm-tts"
|
|
content := "hello"
|
|
var chunks []string
|
|
err := newZhipuAIForTest(srv.URL).AudioSpeechWithSender(
|
|
&modelName,
|
|
&content,
|
|
&APIConfig{ApiKey: &apiKey},
|
|
nil,
|
|
func(content *string, reasoning *string) error {
|
|
if content != nil {
|
|
chunks = append(chunks, *content)
|
|
}
|
|
if reasoning != nil {
|
|
t.Errorf("reasoning=%q", *reasoning)
|
|
}
|
|
return nil
|
|
},
|
|
)
|
|
if err != nil {
|
|
t.Fatalf("AudioSpeechWithSender: %v", err)
|
|
}
|
|
if strings.Join(chunks, "") != "chunk-onechunk-two" {
|
|
t.Errorf("chunks=%q", strings.Join(chunks, ""))
|
|
}
|
|
}
|
|
|
|
func TestZhipuAIAudioSpeechValidation(t *testing.T) {
|
|
apiKey := "test-key"
|
|
modelName := "glm-tts"
|
|
content := "hello"
|
|
|
|
tests := []struct {
|
|
name string
|
|
modelName *string
|
|
audioContent *string
|
|
apiConfig *APIConfig
|
|
want string
|
|
}{
|
|
{name: "missing api key", modelName: &modelName, audioContent: &content, apiConfig: &APIConfig{}, want: "api key is required"},
|
|
{name: "missing model", modelName: nil, audioContent: &content, apiConfig: &APIConfig{ApiKey: &apiKey}, want: "model name is required"},
|
|
{name: "missing content", modelName: &modelName, audioContent: nil, apiConfig: &APIConfig{ApiKey: &apiKey}, want: "audio content is empty"},
|
|
}
|
|
|
|
for _, tt := range tests {
|
|
t.Run(tt.name, func(t *testing.T) {
|
|
_, err := newZhipuAIForTest("http://unused").AudioSpeech(tt.modelName, tt.audioContent, tt.apiConfig, nil)
|
|
if err == nil || !strings.Contains(err.Error(), tt.want) {
|
|
t.Fatalf("error=%v, want %q", err, tt.want)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestZhipuAIAudioSpeechRequiresTTSSuffix(t *testing.T) {
|
|
apiKey := "test-key"
|
|
modelName := "glm-tts"
|
|
content := "hello"
|
|
_, err := NewZhipuAIModel(map[string]string{"default": "http://unused"}, URLSuffix{}).AudioSpeech(
|
|
&modelName,
|
|
&content,
|
|
&APIConfig{ApiKey: &apiKey},
|
|
nil,
|
|
)
|
|
if err == nil || !strings.Contains(err.Error(), "TTS URL suffix is not configured") {
|
|
t.Fatalf("error=%v", err)
|
|
}
|
|
}
|
|
|
|
func TestZhipuAIAudioSpeechHTTPError(t *testing.T) {
|
|
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
|
w.WriteHeader(http.StatusBadRequest)
|
|
_, _ = io.WriteString(w, `{"error":"bad request"}`)
|
|
}))
|
|
defer srv.Close()
|
|
|
|
apiKey := "test-key"
|
|
modelName := "glm-tts"
|
|
content := "hello"
|
|
_, err := newZhipuAIForTest(srv.URL).AudioSpeech(&modelName, &content, &APIConfig{ApiKey: &apiKey}, nil)
|
|
if err == nil || !strings.Contains(err.Error(), "ZhipuAI TTS API error: 400 Bad Request") {
|
|
t.Fatalf("error=%v", err)
|
|
}
|
|
}
|