Files
Assistant/adapters/multimodal/client_test.go
T
m 5594525893
CI / test (push) Successful in 1m8s
CI / test (pull_request) Failing after 52s
Handle reasoning-only vision responses
2026-08-25 12:36:57 +03:00

79 lines
3.1 KiB
Go

package multimodal
import (
"context"
"encoding/json"
"io"
"net/http"
"net/http/httptest"
"strings"
"testing"
)
func TestSpeechClientSendsMultipartAndReturnsTranscript(t *testing.T) {
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if r.URL.Path != "/inference" || r.Method != http.MethodPost {
t.Fatalf("unexpected speech request: %s %s", r.Method, r.URL.Path)
}
if err := r.ParseMultipartForm(1 << 20); err != nil {
t.Fatal(err)
}
file, _, err := r.FormFile("file")
if err != nil {
t.Fatal(err)
}
defer file.Close()
data, _ := io.ReadAll(file)
if string(data) != "audio" {
t.Fatalf("audio payload = %q", data)
}
_ = json.NewEncoder(w).Encode(map[string]string{"text": "распознанная фраза"})
}))
defer server.Close()
text, err := (SpeechClient{BaseURL: server.URL, Client: server.Client()}).Transcribe(context.Background(), "voice.wav", "audio/wav", []byte("audio"))
if err != nil || text != "распознанная фраза" {
t.Fatalf("transcript=%q err=%v", text, err)
}
}
func TestVisionClientSendsImageDataAndInstruction(t *testing.T) {
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
var payload struct {
Messages []struct {
Content []struct {
Type string `json:"type"`
Text string `json:"text"`
ImageURL *struct {
URL string `json:"url"`
} `json:"image_url"`
} `json:"content"`
} `json:"messages"`
}
if err := json.NewDecoder(r.Body).Decode(&payload); err != nil {
t.Fatal(err)
}
if len(payload.Messages) != 1 || len(payload.Messages[0].Content) != 2 || !strings.Contains(payload.Messages[0].Content[0].Text, "Опиши") || payload.Messages[0].Content[1].ImageURL == nil || !strings.HasPrefix(payload.Messages[0].Content[1].ImageURL.URL, "data:image/png;base64,") {
t.Fatalf("unexpected vision payload: %+v", payload)
}
_ = json.NewEncoder(w).Encode(map[string]any{"choices": []any{map[string]any{"message": map[string]string{"content": "на изображении круг"}}}})
}))
defer server.Close()
text, err := (VisionClient{BaseURL: server.URL + "/v1", Client: server.Client()}).Describe(context.Background(), "vision-model", "image/png", []byte{0, 1, 2}, "Опиши изображение")
if err != nil || text != "на изображении круг" {
t.Fatalf("description=%q err=%v", text, err)
}
}
func TestVisionClientUsesReasoningContentWhenProviderOmitsContent(t *testing.T) {
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
_ = json.NewEncoder(w).Encode(map[string]any{"choices": []any{map[string]any{
"message": map[string]string{"content": "", "reasoning_content": "чёрный квадрат"},
}}})
}))
defer server.Close()
text, err := (VisionClient{BaseURL: server.URL, Client: server.Client()}).Describe(context.Background(), "vision-model", "image/png", []byte("image"), "Опиши изображение")
if err != nil || text != "чёрный квадрат" {
t.Fatalf("description=%q err=%v", text, err)
}
}