🎨 Support qwen-vl-plus

common/model-ratio.go

@@ -93,6 +93,7 @@ var ModelRatio = map[string]float64{
 	"qwen-plus":                 1.4286, // ￥0.02 / 1k tokens
 	"qwen-max":                  1.4286, // ￥0.02 / 1k tokens
 	"qwen-max-longcontext":      1.4286, // ￥0.02 / 1k tokens
+	"qwen-vl-plus":              0.5715, // ￥0.008 / 1k tokens
 	"text-embedding-v1":         0.05,   // ￥0.0007 / 1k tokens
 	"SparkDesk":                 1.2858, // ￥0.018 / 1k tokens
 	"360GPT_S2_V9":              0.8572, // ¥0.012 / 1k tokens

providers/ali/base.go

@@ -2,6 +2,7 @@ package ali
 
 import (
 	"fmt"
+	"strings"
 
 	"one-api/providers/base"
 
@@ -28,6 +29,16 @@ type AliProvider struct {
 	base.BaseProvider
 }
 
+func (p *AliProvider) GetFullRequestURL(requestURL string, modelName string) string {
+	baseURL := strings.TrimSuffix(p.GetBaseURL(), "/")
+
+	if modelName == "qwen-vl-plus" {
+		requestURL = "/api/v1/services/aigc/multimodal-generation/generation"
+	}
+
+	return fmt.Sprintf("%s%s", baseURL, requestURL)
+}
+
 // 获取请求头
 func (p *AliProvider) GetRequestHeaders() (headers map[string]string) {
 	headers = make(map[string]string)

providers/ali/chat.go

@@ -26,21 +26,12 @@ func (aliResponse *AliChatResponse) ResponseHandler(resp *http.Response) (OpenAI
 		return
 	}
 
-	choice := types.ChatCompletionChoice{
-		Index: 0,
-		Message: types.ChatCompletionMessage{
-			Role:    "assistant",
-			Content: aliResponse.Output.Text,
-		},
-		FinishReason: aliResponse.Output.FinishReason,
-	}
-
 	OpenAIResponse = types.ChatCompletionResponse{
 		ID:      aliResponse.RequestId,
 		Object:  "chat.completion",
 		Created: common.GetTimestamp(),
 		Model:   aliResponse.Model,
-		Choices: []types.ChatCompletionChoice{choice},
+		Choices: aliResponse.Output.ToChatCompletionChoices(),
 		Usage: &types.Usage{
 			PromptTokens:     aliResponse.Usage.InputTokens,
 			CompletionTokens: aliResponse.Usage.OutputTokens,
@@ -58,10 +49,31 @@ func (p *AliProvider) getChatRequestBody(request *types.ChatCompletionRequest) *
 	messages := make([]AliMessage, 0, len(request.Messages))
 	for i := 0; i < len(request.Messages); i++ {
 		message := request.Messages[i]
-		messages = append(messages, AliMessage{
+		if request.Model != "qwen-vl-plus" {
-			Content: message.StringContent(),
+			messages = append(messages, AliMessage{
-			Role:    strings.ToLower(message.Role),
+				Content: message.StringContent(),
-		})
+				Role:    strings.ToLower(message.Role),
+			})
+		} else {
+			openaiContent := message.ParseContent()
+			var parts []AliMessagePart
+			for _, part := range openaiContent {
+				if part.Type == types.ContentTypeText {
+					parts = append(parts, AliMessagePart{
+						Text: part.Text,
+					})
+				} else if part.Type == types.ContentTypeImageURL {
+					parts = append(parts, AliMessagePart{
+						Image: part.ImageURL.URL,
+					})
+				}
+			}
+			messages = append(messages, AliMessage{
+				Content: parts,
+				Role:    strings.ToLower(message.Role),
+			})
+		}
+
 	}
 
 	enableSearch := false
@@ -77,6 +89,7 @@ func (p *AliProvider) getChatRequestBody(request *types.ChatCompletionRequest) *
 			Messages: messages,
 		},
 		Parameters: AliParameters{
+			ResultFormat:      "message",
 			EnableSearch:      enableSearch,
 			IncrementalOutput: request.Stream,
 		},
@@ -87,6 +100,7 @@ func (p *AliProvider) getChatRequestBody(request *types.ChatCompletionRequest) *
 func (p *AliProvider) ChatAction(request *types.ChatCompletionRequest, isModelMapped bool, promptTokens int) (usage *types.Usage, errWithCode *types.OpenAIErrorWithStatusCode) {
 
 	requestBody := p.getChatRequestBody(request)
+
 	fullRequestURL := p.GetFullRequestURL(p.ChatCompletions, request.Model)
 	headers := p.GetRequestHeaders()
 	if request.Stream {
@@ -134,10 +148,15 @@ func (p *AliProvider) ChatAction(request *types.ChatCompletionRequest, isModelMa
 
 // 阿里云响应转OpenAI响应
 func (p *AliProvider) streamResponseAli2OpenAI(aliResponse *AliChatResponse) *types.ChatCompletionStreamResponse {
+	// chatChoice := aliResponse.Output.ToChatCompletionChoices()
+	// jsonBody, _ := json.MarshalIndent(chatChoice, "", "  ")
+	// fmt.Println("requestBody:", string(jsonBody))
 	var choice types.ChatCompletionStreamChoice
-	choice.Delta.Content = aliResponse.Output.Text
+	choice.Index = aliResponse.Output.Choices[0].Index
-	if aliResponse.Output.FinishReason != "null" {
+	choice.Delta.Content = aliResponse.Output.Choices[0].Message.StringContent()
-		finishReason := aliResponse.Output.FinishReason
+	// fmt.Println("choice.Delta.Content:", chatChoice[0].Message)
+	if aliResponse.Output.Choices[0].FinishReason != "null" {
+		finishReason := aliResponse.Output.Choices[0].FinishReason
 		choice.FinishReason = &finishReason
 	}
 
@@ -200,7 +219,8 @@ func (p *AliProvider) sendStreamRequest(req *http.Request, model string) (usage
 		stopChan <- true
 	}()
 	common.SetEventStreamHeaders(p.Context)
-	// lastResponseText := ""
+	lastResponseText := ""
+	index := 0
 	p.Context.Stream(func(w io.Writer) bool {
 		select {
 		case data := <-dataChan:
@@ -216,9 +236,11 @@ func (p *AliProvider) sendStreamRequest(req *http.Request, model string) (usage
 				usage.TotalTokens = aliResponse.Usage.InputTokens + aliResponse.Usage.OutputTokens
 			}
 			aliResponse.Model = model
+			aliResponse.Output.Choices[0].Index = index
+			index++
 			response := p.streamResponseAli2OpenAI(&aliResponse)
-			// response.Choices[0].Delta.Content = strings.TrimPrefix(response.Choices[0].Delta.Content, lastResponseText)
+			response.Choices[0].Delta.Content = strings.TrimPrefix(response.Choices[0].Delta.Content, lastResponseText)
-			// lastResponseText = aliResponse.Output.Text
+			lastResponseText = aliResponse.Output.Choices[0].Message.StringContent()
 			jsonResponse, err := json.Marshal(response)
 			if err != nil {
 				common.SysError("error marshalling stream response: " + err.Error())

providers/ali/type.go

@@ -1,5 +1,9 @@
 package ali
 
+import (
+	"one-api/types"
+)
+
 type AliError struct {
 	Code      string `json:"code"`
 	Message   string `json:"message"`
@@ -13,10 +17,15 @@ type AliUsage struct {
 }
 
 type AliMessage struct {
-	Content string `json:"content"`
+	Content any    `json:"content"`
 	Role    string `json:"role"`
 }
 
+type AliMessagePart struct {
+	Text  string `json:"text,omitempty"`
+	Image string `json:"image,omitempty"`
+}
+
 type AliInput struct {
 	// Prompt  string       `json:"prompt"`
 	Messages []AliMessage `json:"messages"`
@@ -28,6 +37,7 @@ type AliParameters struct {
 	Seed              uint64  `json:"seed,omitempty"`
 	EnableSearch      bool    `json:"enable_search,omitempty"`
 	IncrementalOutput bool    `json:"incremental_output,omitempty"`
+	ResultFormat      string  `json:"result_format,omitempty"`
 }
 
 type AliChatRequest struct {
@@ -36,9 +46,25 @@ type AliChatRequest struct {
 	Parameters AliParameters `json:"parameters,omitempty"`
 }
 
+type AliChoice struct {
+	FinishReason string                      `json:"finish_reason"`
+	Message      types.ChatCompletionMessage `json:"message"`
+}
+
 type AliOutput struct {
-	Text         string `json:"text"`
+	Choices []types.ChatCompletionChoice `json:"choices"`
-	FinishReason string `json:"finish_reason"`
+}
+
+func (o *AliOutput) ToChatCompletionChoices() []types.ChatCompletionChoice {
+	for i := range o.Choices {
+		_, ok := o.Choices[i].Message.Content.(string)
+		if ok {
+			continue
+		}
+
+		o.Choices[i].Message.Content = o.Choices[i].Message.ParseContent()
+	}
+	return o.Choices
 }
 
 type AliChatResponse struct {

types/chat.go

@@ -27,11 +27,11 @@ func (m ChatCompletionMessage) StringContent() string {
 			if !ok {
 				continue
 			}
-			if contentMap["type"] == "text" {
+
-				if subStr, ok := contentMap["text"].(string); ok {
+			if subStr, ok := contentMap["text"].(string); ok && subStr != "" {
-					contentStr += subStr
+				contentStr += subStr
-				}
 			}
+
 		}
 		return contentStr
 	}
@@ -55,23 +55,26 @@ func (m ChatCompletionMessage) ParseContent() []ChatMessagePart {
 			if !ok {
 				continue
 			}
-			switch contentMap["type"] {
+
-			case ContentTypeText:
+			if subStr, ok := contentMap["text"].(string); ok && subStr != "" {
-				if subStr, ok := contentMap["text"].(string); ok {
+				contentList = append(contentList, ChatMessagePart{
-					contentList = append(contentList, ChatMessagePart{
+					Type: ContentTypeText,
-						Type: ContentTypeText,
+					Text: subStr,
-						Text: subStr,
+				})
-					})
+			} else if subObj, ok := contentMap["image_url"].(map[string]any); ok {
-				}
+				contentList = append(contentList, ChatMessagePart{
-			case ContentTypeImageURL:
+					Type: ContentTypeImageURL,
-				if subObj, ok := contentMap["image_url"].(map[string]any); ok {
+					ImageURL: &ChatMessageImageURL{
-					contentList = append(contentList, ChatMessagePart{
+						URL: subObj["url"].(string),
-						Type: ContentTypeImageURL,
+					},
-						ImageURL: &ChatMessageImageURL{
+				})
-							URL: subObj["url"].(string),
+			} else if subObj, ok := contentMap["image"].(string); ok {
-						},
+				contentList = append(contentList, ChatMessagePart{
-					})
+					Type: ContentTypeImageURL,
-				}
+					ImageURL: &ChatMessageImageURL{
+						URL: subObj,
+					},
+				})
 			}
 		}
 		return contentList