feat(adaptor): 新适配百炼多种图片生成模型

- wan2.6系列生图与编辑,适配多图生成计费
- wan2.5系列生图与编辑
- z-image-turbo生图,适配prompt_extend计费
This commit is contained in:
CaIon
2025-12-29 22:58:32 +08:00
parent 8063897998
commit 48d358faec
16 changed files with 336 additions and 155 deletions

View File

@@ -19,6 +19,22 @@ import (
)
type Adaptor struct {
IsSyncImageModel bool
}
var syncModels = []string{
"z-image",
"qwen-image",
"wan2.6",
}
func isSyncImageModel(modelName string) bool {
for _, m := range syncModels {
if strings.Contains(modelName, m) {
return true
}
}
return false
}
func (a *Adaptor) ConvertGeminiRequest(*gin.Context, *relaycommon.RelayInfo, *dto.GeminiChatRequest) (any, error) {
@@ -45,10 +61,16 @@ func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
case constant.RelayModeRerank:
fullRequestURL = fmt.Sprintf("%s/api/v1/services/rerank/text-rerank/text-rerank", info.ChannelBaseUrl)
case constant.RelayModeImagesGenerations:
fullRequestURL = fmt.Sprintf("%s/api/v1/services/aigc/text2image/image-synthesis", info.ChannelBaseUrl)
if isSyncImageModel(info.OriginModelName) {
fullRequestURL = fmt.Sprintf("%s/api/v1/services/aigc/multimodal-generation/generation", info.ChannelBaseUrl)
} else {
fullRequestURL = fmt.Sprintf("%s/api/v1/services/aigc/text2image/image-synthesis", info.ChannelBaseUrl)
}
case constant.RelayModeImagesEdits:
if isWanModel(info.OriginModelName) {
if isOldWanModel(info.OriginModelName) {
fullRequestURL = fmt.Sprintf("%s/api/v1/services/aigc/image2image/image-synthesis", info.ChannelBaseUrl)
} else if isWanModel(info.OriginModelName) {
fullRequestURL = fmt.Sprintf("%s/api/v1/services/aigc/image-generation/generation", info.ChannelBaseUrl)
} else {
fullRequestURL = fmt.Sprintf("%s/api/v1/services/aigc/multimodal-generation/generation", info.ChannelBaseUrl)
}
@@ -72,7 +94,11 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
req.Set("X-DashScope-Plugin", c.GetString("plugin"))
}
if info.RelayMode == constant.RelayModeImagesGenerations {
req.Set("X-DashScope-Async", "enable")
if isSyncImageModel(info.OriginModelName) {
} else {
req.Set("X-DashScope-Async", "enable")
}
}
if info.RelayMode == constant.RelayModeImagesEdits {
if isWanModel(info.OriginModelName) {
@@ -108,15 +134,25 @@ func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayIn
func (a *Adaptor) ConvertImageRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.ImageRequest) (any, error) {
if info.RelayMode == constant.RelayModeImagesGenerations {
aliRequest, err := oaiImage2Ali(request)
if isSyncImageModel(info.OriginModelName) {
a.IsSyncImageModel = true
}
aliRequest, err := oaiImage2AliImageRequest(info, request, a.IsSyncImageModel)
if err != nil {
return nil, fmt.Errorf("convert image request failed: %w", err)
return nil, fmt.Errorf("convert image request to async ali image request failed: %w", err)
}
return aliRequest, nil
} else if info.RelayMode == constant.RelayModeImagesEdits {
if isWanModel(info.OriginModelName) {
if isOldWanModel(info.OriginModelName) {
return oaiFormEdit2WanxImageEdit(c, info, request)
}
if isSyncImageModel(info.OriginModelName) {
if isWanModel(info.OriginModelName) {
a.IsSyncImageModel = false
} else {
a.IsSyncImageModel = true
}
}
// ali image edit https://bailian.console.aliyun.com/?tab=api#/api/?type=model&url=2976416
// 如果用户使用表单,则需要解析表单数据
if strings.Contains(c.Request.Header.Get("Content-Type"), "multipart/form-data") {
@@ -126,9 +162,9 @@ func (a *Adaptor) ConvertImageRequest(c *gin.Context, info *relaycommon.RelayInf
}
return aliRequest, nil
} else {
aliRequest, err := oaiImage2Ali(request)
aliRequest, err := oaiImage2AliImageRequest(info, request, a.IsSyncImageModel)
if err != nil {
return nil, fmt.Errorf("convert image request failed: %w", err)
return nil, fmt.Errorf("convert image request to async ali image request failed: %w", err)
}
return aliRequest, nil
}
@@ -169,13 +205,9 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, info *relaycom
default:
switch info.RelayMode {
case constant.RelayModeImagesGenerations:
err, usage = aliImageHandler(c, resp, info)
err, usage = aliImageHandler(a, c, resp, info)
case constant.RelayModeImagesEdits:
if isWanModel(info.OriginModelName) {
err, usage = aliImageHandler(c, resp, info)
} else {
err, usage = aliImageEditHandler(c, resp, info)
}
err, usage = aliImageHandler(a, c, resp, info)
case constant.RelayModeRerank:
err, usage = RerankHandler(c, resp, info)
default:

View File

@@ -1,6 +1,13 @@
package ali
import "github.com/QuantumNous/new-api/dto"
import (
"strings"
"github.com/QuantumNous/new-api/dto"
"github.com/QuantumNous/new-api/logger"
"github.com/QuantumNous/new-api/service"
"github.com/gin-gonic/gin"
)
type AliMessage struct {
Content any `json:"content"`
@@ -65,6 +72,7 @@ type AliUsage struct {
InputTokens int `json:"input_tokens"`
OutputTokens int `json:"output_tokens"`
TotalTokens int `json:"total_tokens"`
ImageCount int `json:"image_count,omitempty"`
}
type TaskResult struct {
@@ -75,14 +83,78 @@ type TaskResult struct {
}
type AliOutput struct {
TaskId string `json:"task_id,omitempty"`
TaskStatus string `json:"task_status,omitempty"`
Text string `json:"text"`
FinishReason string `json:"finish_reason"`
Message string `json:"message,omitempty"`
Code string `json:"code,omitempty"`
Results []TaskResult `json:"results,omitempty"`
Choices []map[string]any `json:"choices,omitempty"`
TaskId string `json:"task_id,omitempty"`
TaskStatus string `json:"task_status,omitempty"`
Text string `json:"text"`
FinishReason string `json:"finish_reason"`
Message string `json:"message,omitempty"`
Code string `json:"code,omitempty"`
Results []TaskResult `json:"results,omitempty"`
Choices []struct {
FinishReason string `json:"finish_reason,omitempty"`
Message struct {
Role string `json:"role,omitempty"`
Content []AliMediaContent `json:"content,omitempty"`
ReasoningContent string `json:"reasoning_content,omitempty"`
} `json:"message,omitempty"`
} `json:"choices,omitempty"`
}
func (o *AliOutput) ChoicesToOpenAIImageDate(c *gin.Context, responseFormat string) []dto.ImageData {
var imageData []dto.ImageData
if len(o.Choices) > 0 {
for _, choice := range o.Choices {
var data dto.ImageData
for _, content := range choice.Message.Content {
if content.Image != "" {
if strings.HasPrefix(content.Image, "http") {
var b64Json string
if responseFormat == "b64_json" {
_, b64, err := service.GetImageFromUrl(content.Image)
if err != nil {
logger.LogError(c, "get_image_data_failed: "+err.Error())
continue
}
b64Json = b64
}
data.Url = content.Image
data.B64Json = b64Json
} else {
data.B64Json = content.Image
}
} else if content.Text != "" {
data.RevisedPrompt = content.Text
}
}
imageData = append(imageData, data)
}
}
return imageData
}
func (o *AliOutput) ResultToOpenAIImageDate(c *gin.Context, responseFormat string) []dto.ImageData {
var imageData []dto.ImageData
for _, data := range o.Results {
var b64Json string
if responseFormat == "b64_json" {
_, b64, err := service.GetImageFromUrl(data.Url)
if err != nil {
logger.LogError(c, "get_image_data_failed: "+err.Error())
continue
}
b64Json = b64
} else {
b64Json = data.B64Image
}
imageData = append(imageData, dto.ImageData{
Url: data.Url,
B64Json: b64Json,
RevisedPrompt: "",
})
}
return imageData
}
type AliResponse struct {
@@ -92,18 +164,26 @@ type AliResponse struct {
}
type AliImageRequest struct {
Model string `json:"model"`
Input any `json:"input"`
Parameters any `json:"parameters,omitempty"`
ResponseFormat string `json:"response_format,omitempty"`
Model string `json:"model"`
Input any `json:"input"`
Parameters AliImageParameters `json:"parameters,omitempty"`
ResponseFormat string `json:"response_format,omitempty"`
}
type AliImageParameters struct {
Size string `json:"size,omitempty"`
N int `json:"n,omitempty"`
Steps string `json:"steps,omitempty"`
Scale string `json:"scale,omitempty"`
Watermark *bool `json:"watermark,omitempty"`
Size string `json:"size,omitempty"`
N int `json:"n,omitempty"`
Steps string `json:"steps,omitempty"`
Scale string `json:"scale,omitempty"`
Watermark *bool `json:"watermark,omitempty"`
PromptExtend *bool `json:"prompt_extend,omitempty"`
}
func (p *AliImageParameters) PromptExtendValue() bool {
if p != nil && p.PromptExtend != nil {
return *p.PromptExtend
}
return false
}
type AliImageInput struct {

View File

@@ -21,17 +21,25 @@ import (
"github.com/gin-gonic/gin"
)
func oaiImage2Ali(request dto.ImageRequest) (*AliImageRequest, error) {
func oaiImage2AliImageRequest(info *relaycommon.RelayInfo, request dto.ImageRequest, isSync bool) (*AliImageRequest, error) {
var imageRequest AliImageRequest
imageRequest.Model = request.Model
imageRequest.ResponseFormat = request.ResponseFormat
logger.LogJson(context.Background(), "oaiImage2Ali request extra", request.Extra)
logger.LogDebug(context.Background(), "oaiImage2Ali request isSync: "+fmt.Sprintf("%v", isSync))
if request.Extra != nil {
if val, ok := request.Extra["parameters"]; ok {
err := common.Unmarshal(val, &imageRequest.Parameters)
if err != nil {
return nil, fmt.Errorf("invalid parameters field: %w", err)
}
} else {
// 兼容没有parameters字段的情况从openai标准字段中提取参数
imageRequest.Parameters = AliImageParameters{
Size: strings.Replace(request.Size, "x", "*", -1),
N: int(request.N),
Watermark: request.Watermark,
}
}
if val, ok := request.Extra["input"]; ok {
err := common.Unmarshal(val, &imageRequest.Input)
@@ -41,23 +49,44 @@ func oaiImage2Ali(request dto.ImageRequest) (*AliImageRequest, error) {
}
}
if imageRequest.Parameters == nil {
imageRequest.Parameters = AliImageParameters{
Size: strings.Replace(request.Size, "x", "*", -1),
N: int(request.N),
Watermark: request.Watermark,
if strings.Contains(request.Model, "z-image") {
// z-image 开启prompt_extend后按2倍计费
if imageRequest.Parameters.PromptExtendValue() {
info.PriceData.AddOtherRatio("prompt_extend", 2)
}
}
if imageRequest.Input == nil {
imageRequest.Input = AliImageInput{
Prompt: request.Prompt,
// 检查n参数
if imageRequest.Parameters.N != 0 {
info.PriceData.AddOtherRatio("n", float64(imageRequest.Parameters.N))
}
// 同步图片模型和异步图片模型请求格式不一样
if isSync {
if imageRequest.Input == nil {
imageRequest.Input = AliImageInput{
Messages: []AliMessage{
{
Role: "user",
Content: []AliMediaContent{
{
Text: request.Prompt,
},
},
},
},
}
}
} else {
if imageRequest.Input == nil {
imageRequest.Input = AliImageInput{
Prompt: request.Prompt,
}
}
}
return &imageRequest, nil
}
func getImageBase64sFromForm(c *gin.Context, fieldName string) ([]string, error) {
mf := c.Request.MultipartForm
if mf == nil {
@@ -199,6 +228,8 @@ func asyncTaskWait(c *gin.Context, info *relaycommon.RelayInfo, taskID string) (
var taskResponse AliResponse
var responseBody []byte
time.Sleep(time.Duration(5) * time.Second)
for {
logger.LogDebug(c, fmt.Sprintf("asyncTaskWait step %d/%d, wait %d seconds", step, maxStep, waitSeconds))
step++
@@ -238,32 +269,17 @@ func responseAli2OpenAIImage(c *gin.Context, response *AliResponse, originBody [
Created: info.StartTime.Unix(),
}
for _, data := range response.Output.Results {
var b64Json string
if responseFormat == "b64_json" {
_, b64, err := service.GetImageFromUrl(data.Url)
if err != nil {
logger.LogError(c, "get_image_data_failed: "+err.Error())
continue
}
b64Json = b64
} else {
b64Json = data.B64Image
}
imageResponse.Data = append(imageResponse.Data, dto.ImageData{
Url: data.Url,
B64Json: b64Json,
RevisedPrompt: "",
})
if len(response.Output.Results) > 0 {
imageResponse.Data = response.Output.ResultToOpenAIImageDate(c, responseFormat)
} else if len(response.Output.Choices) > 0 {
imageResponse.Data = response.Output.ChoicesToOpenAIImageDate(c, responseFormat)
}
var mapResponse map[string]any
_ = common.Unmarshal(originBody, &mapResponse)
imageResponse.Extra = mapResponse
imageResponse.Metadata = originBody
return &imageResponse
}
func aliImageHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*types.NewAPIError, *dto.Usage) {
func aliImageHandler(a *Adaptor, c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*types.NewAPIError, *dto.Usage) {
responseFormat := c.GetString("response_format")
var aliTaskResponse AliResponse
@@ -282,66 +298,49 @@ func aliImageHandler(c *gin.Context, resp *http.Response, info *relaycommon.Rela
return types.NewError(errors.New(aliTaskResponse.Message), types.ErrorCodeBadResponse), nil
}
aliResponse, originRespBody, err := asyncTaskWait(c, info, aliTaskResponse.Output.TaskId)
if err != nil {
return types.NewError(err, types.ErrorCodeBadResponse), nil
}
var (
aliResponse *AliResponse
originRespBody []byte
)
if aliResponse.Output.TaskStatus != "SUCCEEDED" {
return types.WithOpenAIError(types.OpenAIError{
Message: aliResponse.Output.Message,
Type: "ali_error",
Param: "",
Code: aliResponse.Output.Code,
}, resp.StatusCode), nil
}
fullTextResponse := responseAli2OpenAIImage(c, aliResponse, originRespBody, info, responseFormat)
jsonResponse, err := common.Marshal(fullTextResponse)
if err != nil {
return types.NewError(err, types.ErrorCodeBadResponseBody), nil
}
service.IOCopyBytesGracefully(c, resp, jsonResponse)
return nil, &dto.Usage{}
}
func aliImageEditHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*types.NewAPIError, *dto.Usage) {
var aliResponse AliResponse
responseBody, err := io.ReadAll(resp.Body)
if err != nil {
return types.NewOpenAIError(err, types.ErrorCodeReadResponseBodyFailed, http.StatusInternalServerError), nil
}
service.CloseResponseBodyGracefully(resp)
err = common.Unmarshal(responseBody, &aliResponse)
if err != nil {
return types.NewOpenAIError(err, types.ErrorCodeBadResponseBody, http.StatusInternalServerError), nil
}
if aliResponse.Message != "" {
logger.LogError(c, "ali_task_failed: "+aliResponse.Message)
return types.NewError(errors.New(aliResponse.Message), types.ErrorCodeBadResponse), nil
}
var fullTextResponse dto.ImageResponse
if len(aliResponse.Output.Choices) > 0 {
fullTextResponse = dto.ImageResponse{
Created: info.StartTime.Unix(),
Data: []dto.ImageData{
{
Url: aliResponse.Output.Choices[0]["message"].(map[string]any)["content"].([]any)[0].(map[string]any)["image"].(string),
B64Json: "",
},
},
if a.IsSyncImageModel {
aliResponse = &aliTaskResponse
originRespBody = responseBody
} else {
// 异步图片模型需要轮询任务结果
aliResponse, originRespBody, err = asyncTaskWait(c, info, aliTaskResponse.Output.TaskId)
if err != nil {
return types.NewError(err, types.ErrorCodeBadResponse), nil
}
if aliResponse.Output.TaskStatus != "SUCCEEDED" {
return types.WithOpenAIError(types.OpenAIError{
Message: aliResponse.Output.Message,
Type: "ali_error",
Param: "",
Code: aliResponse.Output.Code,
}, resp.StatusCode), nil
}
}
var mapResponse map[string]any
_ = common.Unmarshal(responseBody, &mapResponse)
fullTextResponse.Extra = mapResponse
jsonResponse, err := common.Marshal(fullTextResponse)
//logger.LogDebug(c, "ali_async_task_result: "+string(originRespBody))
if a.IsSyncImageModel {
logger.LogDebug(c, "ali_sync_image_result: "+string(originRespBody))
} else {
logger.LogDebug(c, "ali_async_image_result: "+string(originRespBody))
}
imageResponses := responseAli2OpenAIImage(c, aliResponse, originRespBody, info, responseFormat)
// 可能生成多张图片修正计费数量n
if aliResponse.Usage.ImageCount != 0 {
info.PriceData.AddOtherRatio("n", float64(aliResponse.Usage.ImageCount))
} else if len(imageResponses.Data) != 0 {
info.PriceData.AddOtherRatio("n", float64(len(imageResponses.Data)))
}
jsonResponse, err := common.Marshal(imageResponses)
if err != nil {
return types.NewError(err, types.ErrorCodeBadResponseBody), nil
}
service.IOCopyBytesGracefully(c, resp, jsonResponse)
return nil, &dto.Usage{}
}

View File

@@ -26,14 +26,22 @@ func oaiFormEdit2WanxImageEdit(c *gin.Context, info *relaycommon.RelayInfo, requ
if wanInput.Images, err = getImageBase64sFromForm(c, "image"); err != nil {
return nil, fmt.Errorf("get image base64s from form failed: %w", err)
}
wanParams := WanImageParameters{
//wanParams := WanImageParameters{
// N: int(request.N),
//}
imageRequest.Input = wanInput
imageRequest.Parameters = AliImageParameters{
N: int(request.N),
}
imageRequest.Input = wanInput
imageRequest.Parameters = wanParams
info.PriceData.AddOtherRatio("n", float64(imageRequest.Parameters.N))
return &imageRequest, nil
}
func isOldWanModel(modelName string) bool {
return strings.Contains(modelName, "wan") && !strings.Contains(modelName, "wan2.6")
}
func isWanModel(modelName string) bool {
return strings.Contains(modelName, "wan")
}