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

Author: Calcium-Ion

controller/token.go

@@ -1,6 +1,7 @@
 package controller
 
 import (
+	"fmt"
 	"net/http"
 	"strconv"
 	"strings"
@@ -149,6 +150,24 @@ func AddToken(c *gin.Context) {
 		})
 		return
 	}
+	// 非无限额度时，检查额度值是否超出有效范围
+	if !token.UnlimitedQuota {
+		if token.RemainQuota < 0 {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": "额度值不能为负数",
+			})
+			return
+		}
+		maxQuotaValue := int((1000000000 * common.QuotaPerUnit))
+		if token.RemainQuota > maxQuotaValue {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": fmt.Sprintf("额度值超出有效范围，最大值为 %d", maxQuotaValue),
+			})
+			return
+		}
+	}
 	key, err := common.GenerateKey()
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
@@ -216,6 +235,23 @@ func UpdateToken(c *gin.Context) {
 		})
 		return
 	}
+	if !token.UnlimitedQuota {
+		if token.RemainQuota < 0 {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": "额度值不能为负数",
+			})
+			return
+		}
+		maxQuotaValue := int((1000000000 * common.QuotaPerUnit))
+		if token.RemainQuota > maxQuotaValue {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": fmt.Sprintf("额度值超出有效范围，最大值为 %d", maxQuotaValue),
+			})
+			return
+		}
+	}
 	cleanToken, err := model.GetTokenByIds(token.Id, userId)
 	if err != nil {
 		common.ApiError(c, err)
@@ -261,7 +297,6 @@ func UpdateToken(c *gin.Context) {
 		"message": "",
 		"data":    cleanToken,
 	})
-	return
 }
 
 type TokenBatch struct {

dto/openai_image.go

@@ -167,9 +167,9 @@ func (i *ImageRequest) SetModelName(modelName string) {
 }
 
 type ImageResponse struct {
-	Data    []ImageData `json:"data"`
-	Created int64       `json:"created"`
-	Extra   any         `json:"extra,omitempty"`
+	Data     []ImageData     `json:"data"`
+	Created  int64           `json:"created"`
+	Metadata json.RawMessage `json:"metadata,omitempty"`
 }
 type ImageData struct {
 	Url           string `json:"url"`

dto/openai_request.go

@@ -23,6 +23,8 @@ type FormatJsonSchema struct {
 	Strict      json.RawMessage `json:"strict,omitempty"`
 }
 
+// GeneralOpenAIRequest represents a general request structure for OpenAI-compatible APIs.
+// 参数增加规范：无引用的参数必须使用json.RawMessage类型，并添加omitempty标签
 type GeneralOpenAIRequest struct {
 	Model               string            `json:"model,omitempty"`
 	Messages            []Message         `json:"messages,omitempty"`
@@ -82,8 +84,9 @@ type GeneralOpenAIRequest struct {
 	Reasoning json.RawMessage `json:"reasoning,omitempty"`
 	// Ali Qwen Params
 	VlHighResolutionImages json.RawMessage `json:"vl_high_resolution_images,omitempty"`
-	EnableThinking         any             `json:"enable_thinking,omitempty"`
+	EnableThinking         json.RawMessage `json:"enable_thinking,omitempty"`
 	ChatTemplateKwargs     json.RawMessage `json:"chat_template_kwargs,omitempty"`
+	EnableSearch           json.RawMessage `json:"enable_search,omitempty"`
 	// ollama Params
 	Think json.RawMessage `json:"think,omitempty"`
 	// baidu v2

main.go

@@ -188,6 +188,7 @@ func InjectUmamiAnalytics() {
 		analyticsInjectBuilder.WriteString(umamiSiteID)
 		analyticsInjectBuilder.WriteString("\"></script>")
 	}
+	analyticsInjectBuilder.WriteString("<!--Umami QuantumNous-->\n")
 	analyticsInject := analyticsInjectBuilder.String()
 	indexPage = bytes.ReplaceAll(indexPage, []byte("<!--umami-->\n"), []byte(analyticsInject))
 }
@@ -209,6 +210,7 @@ func InjectGoogleAnalytics() {
 		analyticsInjectBuilder.WriteString("');")
 		analyticsInjectBuilder.WriteString("</script>")
 	}
+	analyticsInjectBuilder.WriteString("<!--Google Analytics QuantumNous-->\n")
 	analyticsInject := analyticsInjectBuilder.String()
 	indexPage = bytes.ReplaceAll(indexPage, []byte("<!--Google Analytics-->\n"), []byte(analyticsInject))
 }

relay/audio_handler.go

@@ -70,7 +70,7 @@ func AudioHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *type
 	if usage.(*dto.Usage).CompletionTokenDetails.AudioTokens > 0 || usage.(*dto.Usage).PromptTokensDetails.AudioTokens > 0 {
 		service.PostAudioConsumeQuota(c, info, usage.(*dto.Usage), "")
 	} else {
-		postConsumeQuota(c, info, usage.(*dto.Usage), "")
+		postConsumeQuota(c, info, usage.(*dto.Usage))
 	}
 
 	return nil

relay/channel/ali/adaptor.go

@@ -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:

relay/channel/ali/dto.go

@@ -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 {