重构前
This commit is contained in:
@@ -0,0 +1,277 @@
|
||||
package ali
|
||||
|
||||
import (
|
||||
"errors"
|
||||
"fmt"
|
||||
"io"
|
||||
"net/http"
|
||||
"strings"
|
||||
|
||||
"github.com/QuantumNous/new-api/common"
|
||||
"github.com/QuantumNous/new-api/relay/channel"
|
||||
"github.com/QuantumNous/new-api/relay/channel/claude"
|
||||
"github.com/QuantumNous/new-api/relay/channel/openai"
|
||||
relaycommon "github.com/QuantumNous/new-api/relay/common"
|
||||
"github.com/QuantumNous/new-api/relay/constant"
|
||||
"github.com/QuantumNous/new-api/relaykit/dto"
|
||||
"github.com/QuantumNous/new-api/relaykit/types"
|
||||
"github.com/QuantumNous/new-api/service"
|
||||
"github.com/QuantumNous/new-api/setting/model_setting"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
"github.com/samber/lo"
|
||||
)
|
||||
|
||||
type Adaptor struct {
|
||||
IsSyncImageModel bool
|
||||
}
|
||||
|
||||
const aliAnthropicMessagesModelsEnv = "ALI_ANTHROPIC_MESSAGES_MODELS"
|
||||
const defaultAliAnthropicMessagesModels = "qwen,deepseek-v4,kimi,glm,minimax-m"
|
||||
|
||||
/*
|
||||
var syncModels = []string{
|
||||
"z-image",
|
||||
"qwen-image",
|
||||
"wan2.6",
|
||||
}
|
||||
*/
|
||||
func supportsAliAnthropicMessages(modelName string) bool {
|
||||
normalizedModelName := strings.ToLower(strings.TrimSpace(modelName))
|
||||
if normalizedModelName == "" {
|
||||
return false
|
||||
}
|
||||
|
||||
return lo.SomeBy(aliAnthropicMessagesModelPatterns(), func(pattern string) bool {
|
||||
return strings.Contains(normalizedModelName, pattern)
|
||||
})
|
||||
}
|
||||
|
||||
func aliAnthropicMessagesModelPatterns() []string {
|
||||
configuredModels := common.GetEnvOrDefaultString(aliAnthropicMessagesModelsEnv, defaultAliAnthropicMessagesModels)
|
||||
return lo.FilterMap(strings.Split(configuredModels, ","), func(item string, _ int) (string, bool) {
|
||||
pattern := strings.ToLower(strings.TrimSpace(item))
|
||||
return pattern, pattern != ""
|
||||
})
|
||||
}
|
||||
|
||||
var syncModels = []string{
|
||||
"z-image",
|
||||
"qwen-image",
|
||||
"wan2.6",
|
||||
}
|
||||
|
||||
func isSyncImageModel(modelName string) bool {
|
||||
return model_setting.IsSyncImageModel(modelName)
|
||||
}
|
||||
|
||||
func (a *Adaptor) ConvertGeminiRequest(*gin.Context, *relaycommon.RelayInfo, *dto.GeminiChatRequest) (any, error) {
|
||||
//TODO implement me
|
||||
return nil, errors.New("not implemented")
|
||||
}
|
||||
|
||||
func (a *Adaptor) ConvertClaudeRequest(c *gin.Context, info *relaycommon.RelayInfo, req *dto.ClaudeRequest) (any, error) {
|
||||
if supportsAliAnthropicMessages(info.UpstreamModelName) {
|
||||
return req, nil
|
||||
}
|
||||
|
||||
result, err := service.ConvertRequest(c, info, types.RelayFormatOpenAI, req)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
oaiReq, ok := result.Value.(*dto.GeneralOpenAIRequest)
|
||||
if !ok {
|
||||
return nil, fmt.Errorf("expected OpenAI chat completions request, got %T", result.Value)
|
||||
}
|
||||
if info.SupportStreamOptions && info.IsStream {
|
||||
oaiReq.StreamOptions = &dto.StreamOptions{IncludeUsage: true}
|
||||
}
|
||||
return a.ConvertOpenAIRequest(c, info, oaiReq)
|
||||
}
|
||||
|
||||
func (a *Adaptor) Init(info *relaycommon.RelayInfo) {
|
||||
}
|
||||
|
||||
func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
|
||||
var fullRequestURL string
|
||||
switch info.RelayFormat {
|
||||
case types.RelayFormatClaude:
|
||||
if supportsAliAnthropicMessages(info.UpstreamModelName) {
|
||||
fullRequestURL = fmt.Sprintf("%s/apps/anthropic/v1/messages", info.ChannelBaseUrl)
|
||||
} else {
|
||||
fullRequestURL = fmt.Sprintf("%s/compatible-mode/v1/chat/completions", info.ChannelBaseUrl)
|
||||
}
|
||||
default:
|
||||
switch info.RelayMode {
|
||||
case constant.RelayModeEmbeddings:
|
||||
fullRequestURL = fmt.Sprintf("%s/compatible-mode/v1/embeddings", info.ChannelBaseUrl)
|
||||
case constant.RelayModeRerank:
|
||||
fullRequestURL = fmt.Sprintf("%s/api/v1/services/rerank/text-rerank/text-rerank", info.ChannelBaseUrl)
|
||||
case constant.RelayModeResponses:
|
||||
fullRequestURL = fmt.Sprintf("%s/api/v2/apps/protocols/compatible-mode/v1/responses", info.ChannelBaseUrl)
|
||||
case constant.RelayModeImagesGenerations:
|
||||
if isSyncImageModel(info.UpstreamModelName) {
|
||||
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 isOldWanModel(info.UpstreamModelName) {
|
||||
fullRequestURL = fmt.Sprintf("%s/api/v1/services/aigc/image2image/image-synthesis", info.ChannelBaseUrl)
|
||||
} else if isWanModel(info.UpstreamModelName) {
|
||||
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)
|
||||
}
|
||||
case constant.RelayModeCompletions:
|
||||
fullRequestURL = fmt.Sprintf("%s/compatible-mode/v1/completions", info.ChannelBaseUrl)
|
||||
default:
|
||||
fullRequestURL = fmt.Sprintf("%s/compatible-mode/v1/chat/completions", info.ChannelBaseUrl)
|
||||
}
|
||||
}
|
||||
|
||||
return fullRequestURL, nil
|
||||
}
|
||||
|
||||
func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *relaycommon.RelayInfo) error {
|
||||
channel.SetupApiRequestHeader(info, c, req)
|
||||
req.Set("Authorization", "Bearer "+info.ApiKey)
|
||||
if info.IsStream {
|
||||
req.Set("X-DashScope-SSE", "enable")
|
||||
}
|
||||
if c.GetString("plugin") != "" {
|
||||
req.Set("X-DashScope-Plugin", c.GetString("plugin"))
|
||||
}
|
||||
if info.RelayMode == constant.RelayModeImagesGenerations {
|
||||
if isSyncImageModel(info.UpstreamModelName) {
|
||||
|
||||
} else {
|
||||
req.Set("X-DashScope-Async", "enable")
|
||||
}
|
||||
}
|
||||
if info.RelayMode == constant.RelayModeImagesEdits {
|
||||
if isWanModel(info.UpstreamModelName) {
|
||||
req.Set("X-DashScope-Async", "enable")
|
||||
}
|
||||
req.Set("Content-Type", "application/json")
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
|
||||
if request == nil {
|
||||
return nil, errors.New("request is nil")
|
||||
}
|
||||
// docs: https://bailian.console.aliyun.com/?tab=api#/api/?type=model&url=2712216
|
||||
// fix: InternalError.Algo.InvalidParameter: The value of the enable_thinking parameter is restricted to True.
|
||||
//if strings.Contains(request.Model, "thinking") {
|
||||
// request.EnableThinking = true
|
||||
// request.Stream = true
|
||||
// info.IsStream = true
|
||||
//}
|
||||
//// fix: ali parameter.enable_thinking must be set to false for non-streaming calls
|
||||
//if !info.IsStream {
|
||||
// request.EnableThinking = false
|
||||
//}
|
||||
|
||||
switch info.RelayMode {
|
||||
default:
|
||||
aliReq := requestOpenAI2Ali(*request, info.UpstreamModelName)
|
||||
return aliReq, nil
|
||||
}
|
||||
}
|
||||
|
||||
func (a *Adaptor) ConvertImageRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.ImageRequest) (any, error) {
|
||||
if info.RelayMode == constant.RelayModeImagesGenerations {
|
||||
if isSyncImageModel(info.UpstreamModelName) {
|
||||
a.IsSyncImageModel = true
|
||||
}
|
||||
aliRequest, err := oaiImage2AliImageRequest(info, request, a.IsSyncImageModel)
|
||||
if err != nil {
|
||||
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 isOldWanModel(info.UpstreamModelName) {
|
||||
return oaiFormEdit2WanxImageEdit(c, info, request)
|
||||
}
|
||||
if isSyncImageModel(info.UpstreamModelName) {
|
||||
if isWanModel(info.UpstreamModelName) {
|
||||
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") {
|
||||
aliRequest, err := oaiFormEdit2AliImageEdit(c, info, request)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("convert image edit form request failed: %w", err)
|
||||
}
|
||||
return aliRequest, nil
|
||||
} else {
|
||||
aliRequest, err := oaiImage2AliImageRequest(info, request, a.IsSyncImageModel)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("convert image request to async ali image request failed: %w", err)
|
||||
}
|
||||
return aliRequest, nil
|
||||
}
|
||||
}
|
||||
return nil, fmt.Errorf("unsupported image relay mode: %d", info.RelayMode)
|
||||
}
|
||||
|
||||
func (a *Adaptor) ConvertRerankRequest(c *gin.Context, relayMode int, request dto.RerankRequest) (any, error) {
|
||||
return ConvertRerankRequest(request), nil
|
||||
}
|
||||
|
||||
func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.EmbeddingRequest) (any, error) {
|
||||
return request, nil
|
||||
}
|
||||
|
||||
func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.AudioRequest) (io.Reader, error) {
|
||||
//TODO implement me
|
||||
return nil, errors.New("not implemented")
|
||||
}
|
||||
|
||||
func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
|
||||
return request, nil
|
||||
}
|
||||
|
||||
func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error) {
|
||||
return channel.DoApiRequest(a, c, info, requestBody)
|
||||
}
|
||||
|
||||
func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (usage any, err *types.NewAPIError) {
|
||||
switch info.RelayFormat {
|
||||
case types.RelayFormatClaude:
|
||||
if supportsAliAnthropicMessages(info.UpstreamModelName) {
|
||||
adaptor := claude.Adaptor{}
|
||||
return adaptor.DoResponse(c, resp, info)
|
||||
}
|
||||
|
||||
adaptor := openai.Adaptor{}
|
||||
return adaptor.DoResponse(c, resp, info)
|
||||
default:
|
||||
switch info.RelayMode {
|
||||
case constant.RelayModeImagesGenerations:
|
||||
err, usage = aliImageHandler(a, c, resp, info)
|
||||
case constant.RelayModeImagesEdits:
|
||||
err, usage = aliImageHandler(a, c, resp, info)
|
||||
case constant.RelayModeRerank:
|
||||
err, usage = RerankHandler(c, resp, info)
|
||||
default:
|
||||
adaptor := openai.Adaptor{}
|
||||
usage, err = adaptor.DoResponse(c, resp, info)
|
||||
}
|
||||
return usage, err
|
||||
}
|
||||
}
|
||||
|
||||
func (a *Adaptor) GetModelList() []string {
|
||||
return ModelList
|
||||
}
|
||||
|
||||
func (a *Adaptor) GetChannelName() string {
|
||||
return ChannelName
|
||||
}
|
||||
@@ -0,0 +1,161 @@
|
||||
package ali
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"testing"
|
||||
|
||||
"github.com/QuantumNous/new-api/common"
|
||||
relaycommon "github.com/QuantumNous/new-api/relay/common"
|
||||
"github.com/QuantumNous/new-api/relay/constant"
|
||||
relayhelper "github.com/QuantumNous/new-api/relay/helper"
|
||||
"github.com/QuantumNous/new-api/relaykit/dto"
|
||||
"github.com/gin-gonic/gin"
|
||||
"github.com/stretchr/testify/assert"
|
||||
"github.com/stretchr/testify/require"
|
||||
"github.com/tidwall/gjson"
|
||||
)
|
||||
|
||||
func TestConvertOpenAIRequestFiltersThinkingBudgetByUpstreamModel(t *testing.T) {
|
||||
tests := []struct {
|
||||
name string
|
||||
requestModel string
|
||||
upstreamModel string
|
||||
budget string
|
||||
wantBudget bool
|
||||
wantValue int64
|
||||
}{
|
||||
{
|
||||
name: "qwen",
|
||||
requestModel: "qwen-plus",
|
||||
upstreamModel: "qwen-plus",
|
||||
budget: "128",
|
||||
wantBudget: true,
|
||||
wantValue: 128,
|
||||
},
|
||||
{
|
||||
name: "qwq explicit zero",
|
||||
requestModel: "qwq-32b",
|
||||
upstreamModel: "qwq-32b",
|
||||
budget: "0",
|
||||
wantBudget: true,
|
||||
wantValue: 0,
|
||||
},
|
||||
{
|
||||
name: "unsupported upstream overrides qwen request",
|
||||
requestModel: "qwen-plus",
|
||||
upstreamModel: "deepseek-r1",
|
||||
budget: "128",
|
||||
wantBudget: false,
|
||||
},
|
||||
}
|
||||
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
request := &dto.GeneralOpenAIRequest{
|
||||
Model: tt.requestModel,
|
||||
EnableThinking: json.RawMessage(`true`),
|
||||
ThinkingBudget: json.RawMessage(tt.budget),
|
||||
}
|
||||
info := &relaycommon.RelayInfo{
|
||||
ChannelMeta: &relaycommon.ChannelMeta{
|
||||
UpstreamModelName: tt.upstreamModel,
|
||||
},
|
||||
}
|
||||
|
||||
convertedValue, err := (&Adaptor{}).ConvertOpenAIRequest(nil, info, request)
|
||||
require.NoError(t, err)
|
||||
converted, ok := convertedValue.(*dto.GeneralOpenAIRequest)
|
||||
require.True(t, ok)
|
||||
|
||||
if tt.wantBudget {
|
||||
assert.Equal(t, tt.budget, string(converted.ThinkingBudget))
|
||||
} else {
|
||||
assert.Nil(t, converted.ThinkingBudget)
|
||||
}
|
||||
|
||||
encoded, err := common.Marshal(converted)
|
||||
require.NoError(t, err)
|
||||
|
||||
assert.True(t, gjson.GetBytes(encoded, "enable_thinking").Bool())
|
||||
value := gjson.GetBytes(encoded, "thinking_budget")
|
||||
assert.Equal(t, tt.wantBudget, value.Exists())
|
||||
if tt.wantBudget {
|
||||
assert.Equal(t, tt.wantValue, value.Int())
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
func TestConvertOpenAIRequestPreservesExplicitZeroForMappedQwenModel(t *testing.T) {
|
||||
const (
|
||||
clientModel = "customer-model"
|
||||
upstreamModel = "Qwen/Qwen3-235B-A22B-Thinking-2507"
|
||||
)
|
||||
|
||||
c, _ := gin.CreateTestContext(httptest.NewRecorder())
|
||||
c.Set("model_mapping", `{"customer-model":"Qwen/Qwen3-235B-A22B-Thinking-2507"}`)
|
||||
|
||||
request := &dto.GeneralOpenAIRequest{
|
||||
Model: clientModel,
|
||||
EnableThinking: json.RawMessage(`true`),
|
||||
ThinkingBudget: json.RawMessage(`0`),
|
||||
}
|
||||
info := &relaycommon.RelayInfo{
|
||||
OriginModelName: clientModel,
|
||||
ChannelMeta: &relaycommon.ChannelMeta{
|
||||
UpstreamModelName: clientModel,
|
||||
},
|
||||
}
|
||||
|
||||
err := relayhelper.ModelMappedHelper(c, info, request)
|
||||
require.NoError(t, err)
|
||||
assert.True(t, info.IsModelMapped)
|
||||
assert.Equal(t, upstreamModel, info.UpstreamModelName)
|
||||
assert.Equal(t, upstreamModel, request.Model)
|
||||
|
||||
convertedValue, err := (&Adaptor{}).ConvertOpenAIRequest(c, info, request)
|
||||
require.NoError(t, err)
|
||||
converted, ok := convertedValue.(*dto.GeneralOpenAIRequest)
|
||||
require.True(t, ok)
|
||||
assert.Equal(t, json.RawMessage(`0`), converted.ThinkingBudget)
|
||||
|
||||
encoded, err := common.Marshal(converted)
|
||||
require.NoError(t, err)
|
||||
|
||||
value := gjson.GetBytes(encoded, "thinking_budget")
|
||||
assert.True(t, value.Exists())
|
||||
assert.Equal(t, int64(0), value.Int())
|
||||
}
|
||||
|
||||
func TestMappedAliImageModelUsesUpstreamProtocol(t *testing.T) {
|
||||
c, _ := gin.CreateTestContext(httptest.NewRecorder())
|
||||
c.Request = httptest.NewRequest(http.MethodPost, "/v1/images/generations", nil)
|
||||
|
||||
info := &relaycommon.RelayInfo{
|
||||
RelayMode: constant.RelayModeImagesGenerations,
|
||||
OriginModelName: "customer-image-model",
|
||||
ChannelMeta: &relaycommon.ChannelMeta{
|
||||
ChannelBaseUrl: "https://dashscope.aliyuncs.com",
|
||||
UpstreamModelName: "qwen-image-3.0-pro",
|
||||
},
|
||||
}
|
||||
|
||||
adaptor := &Adaptor{}
|
||||
url, err := adaptor.GetRequestURL(info)
|
||||
require.NoError(t, err)
|
||||
assert.Equal(t, "https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation", url)
|
||||
|
||||
header := http.Header{}
|
||||
require.NoError(t, adaptor.SetupRequestHeader(c, &header, info))
|
||||
assert.Empty(t, header.Get("X-DashScope-Async"))
|
||||
|
||||
converted, err := adaptor.ConvertImageRequest(c, info, dto.ImageRequest{
|
||||
Model: info.UpstreamModelName,
|
||||
Prompt: "poster",
|
||||
})
|
||||
require.NoError(t, err)
|
||||
assert.True(t, adaptor.IsSyncImageModel)
|
||||
assert.IsType(t, &AliImageRequest{}, converted)
|
||||
}
|
||||
@@ -0,0 +1,14 @@
|
||||
package ali
|
||||
|
||||
var ModelList = []string{
|
||||
"qwen-turbo",
|
||||
"qwen-plus",
|
||||
"qwen-max",
|
||||
"qwen-max-longcontext",
|
||||
"qwq-32b",
|
||||
"qwen3-235b-a22b",
|
||||
"text-embedding-v1",
|
||||
"gte-rerank-v2",
|
||||
}
|
||||
|
||||
var ChannelName = "ali"
|
||||
@@ -0,0 +1,236 @@
|
||||
package ali
|
||||
|
||||
import (
|
||||
"strings"
|
||||
|
||||
"github.com/QuantumNous/new-api/logger"
|
||||
"github.com/QuantumNous/new-api/relaykit/dto"
|
||||
"github.com/QuantumNous/new-api/service"
|
||||
"github.com/gin-gonic/gin"
|
||||
)
|
||||
|
||||
type AliMessage struct {
|
||||
Content any `json:"content"`
|
||||
Role string `json:"role"`
|
||||
}
|
||||
|
||||
type AliMediaContent struct {
|
||||
Image string `json:"image,omitempty"`
|
||||
Text string `json:"text,omitempty"`
|
||||
}
|
||||
|
||||
type AliInput struct {
|
||||
Prompt string `json:"prompt,omitempty"`
|
||||
//History []AliMessage `json:"history,omitempty"`
|
||||
Messages []AliMessage `json:"messages"`
|
||||
}
|
||||
|
||||
type AliParameters struct {
|
||||
TopP float64 `json:"top_p,omitempty"`
|
||||
TopK int `json:"top_k,omitempty"`
|
||||
Seed uint64 `json:"seed,omitempty"`
|
||||
EnableSearch bool `json:"enable_search,omitempty"`
|
||||
IncrementalOutput bool `json:"incremental_output,omitempty"`
|
||||
}
|
||||
|
||||
type AliChatRequest struct {
|
||||
Model string `json:"model"`
|
||||
Input AliInput `json:"input,omitempty"`
|
||||
Parameters AliParameters `json:"parameters,omitempty"`
|
||||
}
|
||||
|
||||
type AliEmbeddingRequest struct {
|
||||
Model string `json:"model"`
|
||||
Input struct {
|
||||
Texts []string `json:"texts"`
|
||||
} `json:"input"`
|
||||
Parameters *struct {
|
||||
TextType string `json:"text_type,omitempty"`
|
||||
} `json:"parameters,omitempty"`
|
||||
}
|
||||
|
||||
type AliEmbedding struct {
|
||||
Embedding []float64 `json:"embedding"`
|
||||
TextIndex int `json:"text_index"`
|
||||
}
|
||||
|
||||
type AliEmbeddingResponse struct {
|
||||
Output struct {
|
||||
Embeddings []AliEmbedding `json:"embeddings"`
|
||||
} `json:"output"`
|
||||
Usage AliUsage `json:"usage"`
|
||||
AliError
|
||||
}
|
||||
|
||||
type AliError struct {
|
||||
Code string `json:"code"`
|
||||
Message string `json:"message"`
|
||||
RequestId string `json:"request_id"`
|
||||
}
|
||||
|
||||
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 {
|
||||
B64Image string `json:"b64_image,omitempty"`
|
||||
Url string `json:"url,omitempty"`
|
||||
Code string `json:"code,omitempty"`
|
||||
Message string `json:"message,omitempty"`
|
||||
}
|
||||
|
||||
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 []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 {
|
||||
Output AliOutput `json:"output"`
|
||||
Usage AliUsage `json:"usage"`
|
||||
AliError
|
||||
}
|
||||
|
||||
type AliImageRequest struct {
|
||||
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"`
|
||||
PromptExtend *bool `json:"prompt_extend,omitempty"`
|
||||
ThinkingMode *bool `json:"thinking_mode,omitempty"`
|
||||
EnableSequential *bool `json:"enable_sequential,omitempty"`
|
||||
BboxList any `json:"bbox_list,omitempty"`
|
||||
ColorPalette any `json:"color_palette,omitempty"`
|
||||
Seed *int `json:"seed,omitempty"`
|
||||
}
|
||||
|
||||
func (p *AliImageParameters) PromptExtendValue() bool {
|
||||
if p != nil && p.PromptExtend != nil {
|
||||
return *p.PromptExtend
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
type AliImageInput struct {
|
||||
Prompt string `json:"prompt,omitempty"`
|
||||
NegativePrompt string `json:"negative_prompt,omitempty"`
|
||||
Messages []AliMessage `json:"messages,omitempty"`
|
||||
}
|
||||
|
||||
type WanImageInput struct {
|
||||
Prompt string `json:"prompt"` // 必需：文本提示词，描述生成图像中期望包含的元素和视觉特点
|
||||
Images []string `json:"images"` // 必需：图像URL数组，长度不超过2，支持HTTP/HTTPS URL或Base64编码
|
||||
NegativePrompt string `json:"negative_prompt,omitempty"` // 可选：反向提示词，描述不希望在画面中看到的内容
|
||||
}
|
||||
|
||||
type WanImageParameters struct {
|
||||
N int `json:"n,omitempty"` // 生成图片数量，取值范围1-4，默认4
|
||||
Watermark *bool `json:"watermark,omitempty"` // 是否添加水印标识，默认false
|
||||
Seed int `json:"seed,omitempty"` // 随机数种子，取值范围[0, 2147483647]
|
||||
Strength float64 `json:"strength,omitempty"` // 修改幅度 0.0-1.0，默认0.5（部分模型支持）
|
||||
}
|
||||
|
||||
type AliRerankParameters struct {
|
||||
TopN *int `json:"top_n,omitempty"`
|
||||
ReturnDocuments *bool `json:"return_documents,omitempty"`
|
||||
}
|
||||
|
||||
type AliRerankInput struct {
|
||||
Query string `json:"query"`
|
||||
Documents []any `json:"documents"`
|
||||
}
|
||||
|
||||
type AliRerankRequest struct {
|
||||
Model string `json:"model"`
|
||||
Input AliRerankInput `json:"input"`
|
||||
Parameters AliRerankParameters `json:"parameters,omitempty"`
|
||||
}
|
||||
|
||||
type AliRerankResponse struct {
|
||||
Output struct {
|
||||
Results []dto.RerankResponseResult `json:"results"`
|
||||
} `json:"output"`
|
||||
Usage AliUsage `json:"usage"`
|
||||
RequestId string `json:"request_id"`
|
||||
AliError
|
||||
}
|
||||
@@ -0,0 +1,348 @@
|
||||
package ali
|
||||
|
||||
import (
|
||||
"encoding/base64"
|
||||
"errors"
|
||||
"fmt"
|
||||
"io"
|
||||
"mime/multipart"
|
||||
"net/http"
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
"github.com/QuantumNous/new-api/common"
|
||||
"github.com/QuantumNous/new-api/logger"
|
||||
relaycommon "github.com/QuantumNous/new-api/relay/common"
|
||||
"github.com/QuantumNous/new-api/relaykit/dto"
|
||||
"github.com/QuantumNous/new-api/relaykit/types"
|
||||
"github.com/QuantumNous/new-api/service"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
"github.com/samber/lo"
|
||||
)
|
||||
|
||||
func oaiImage2AliImageRequest(info *relaycommon.RelayInfo, request dto.ImageRequest, isSync bool) (*AliImageRequest, error) {
|
||||
var imageRequest AliImageRequest
|
||||
imageRequest.Model = request.Model
|
||||
imageRequest.ResponseFormat = request.ResponseFormat
|
||||
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(lo.FromPtrOr(request.N, uint(1))),
|
||||
Watermark: request.Watermark,
|
||||
}
|
||||
}
|
||||
if val, ok := request.Extra["input"]; ok {
|
||||
err := common.Unmarshal(val, &imageRequest.Input)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("invalid input field: %w", err)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
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)
|
||||
}
|
||||
}
|
||||
|
||||
// Parameters may come from Extra["parameters"], bypassing the standard
|
||||
// top-level n validation; enforce the same bound before it becomes a
|
||||
// billing multiplier.
|
||||
if imageRequest.Parameters.N < 0 || imageRequest.Parameters.N > dto.MaxImageN {
|
||||
return nil, fmt.Errorf("parameters.n must be an integer between 1 and %d", dto.MaxImageN)
|
||||
}
|
||||
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 {
|
||||
if _, err := c.MultipartForm(); err != nil {
|
||||
return nil, fmt.Errorf("failed to parse image edit form request: %w", err)
|
||||
}
|
||||
mf = c.Request.MultipartForm
|
||||
}
|
||||
|
||||
var imageFiles []*multipart.FileHeader
|
||||
var exists bool
|
||||
|
||||
// First check for standard "image" field
|
||||
if imageFiles, exists = mf.File["image"]; !exists || len(imageFiles) == 0 {
|
||||
// If not found, check for "image[]" field
|
||||
if imageFiles, exists = mf.File["image[]"]; !exists || len(imageFiles) == 0 {
|
||||
// If still not found, iterate through all fields to find any that start with "image["
|
||||
foundArrayImages := false
|
||||
for fieldName, files := range mf.File {
|
||||
if strings.HasPrefix(fieldName, "image[") && len(files) > 0 {
|
||||
foundArrayImages = true
|
||||
imageFiles = append(imageFiles, files...)
|
||||
}
|
||||
}
|
||||
|
||||
// If no image fields found at all
|
||||
if !foundArrayImages && (len(imageFiles) == 0) {
|
||||
return nil, errors.New("image is required")
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if len(imageFiles) == 0 {
|
||||
return nil, errors.New("image is required")
|
||||
}
|
||||
|
||||
//if len(imageFiles) > 1 {
|
||||
// return nil, errors.New("only one image is supported for qwen edit")
|
||||
//}
|
||||
|
||||
// 获取base64编码的图片
|
||||
var imageBase64s []string
|
||||
for _, file := range imageFiles {
|
||||
image, err := file.Open()
|
||||
if err != nil {
|
||||
return nil, errors.New("failed to open image file")
|
||||
}
|
||||
|
||||
// 读取文件内容
|
||||
imageData, err := io.ReadAll(image)
|
||||
if err != nil {
|
||||
return nil, errors.New("failed to read image file")
|
||||
}
|
||||
|
||||
// 获取MIME类型
|
||||
mimeType := http.DetectContentType(imageData)
|
||||
|
||||
// 编码为base64
|
||||
base64Data := base64.StdEncoding.EncodeToString(imageData)
|
||||
|
||||
// 构造data URL格式
|
||||
dataURL := fmt.Sprintf("data:%s;base64,%s", mimeType, base64Data)
|
||||
imageBase64s = append(imageBase64s, dataURL)
|
||||
image.Close()
|
||||
}
|
||||
return imageBase64s, nil
|
||||
}
|
||||
|
||||
func oaiFormEdit2AliImageEdit(c *gin.Context, info *relaycommon.RelayInfo, request dto.ImageRequest) (*AliImageRequest, error) {
|
||||
var imageRequest AliImageRequest
|
||||
imageRequest.Model = request.Model
|
||||
imageRequest.ResponseFormat = request.ResponseFormat
|
||||
|
||||
imageBase64s, err := getImageBase64sFromForm(c, "image")
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("get image base64s from form failed: %w", err)
|
||||
}
|
||||
//dto.MediaContent{}
|
||||
mediaContents := make([]AliMediaContent, len(imageBase64s))
|
||||
for i, b64 := range imageBase64s {
|
||||
mediaContents[i] = AliMediaContent{
|
||||
Image: b64,
|
||||
}
|
||||
}
|
||||
mediaContents = append(mediaContents, AliMediaContent{
|
||||
Text: request.Prompt,
|
||||
})
|
||||
imageRequest.Input = AliImageInput{
|
||||
Messages: []AliMessage{
|
||||
{
|
||||
Role: "user",
|
||||
Content: mediaContents,
|
||||
},
|
||||
},
|
||||
}
|
||||
imageRequest.Parameters = AliImageParameters{
|
||||
N: int(lo.FromPtrOr(request.N, uint(1))),
|
||||
Watermark: request.Watermark,
|
||||
}
|
||||
return &imageRequest, nil
|
||||
}
|
||||
|
||||
func updateTask(info *relaycommon.RelayInfo, taskID string) (*AliResponse, error, []byte) {
|
||||
url := fmt.Sprintf("%s/api/v1/tasks/%s", info.ChannelBaseUrl, taskID)
|
||||
|
||||
var aliResponse AliResponse
|
||||
|
||||
req, err := http.NewRequest("GET", url, nil)
|
||||
if err != nil {
|
||||
return &aliResponse, err, nil
|
||||
}
|
||||
|
||||
req.Header.Set("Authorization", "Bearer "+info.ApiKey)
|
||||
|
||||
client := &http.Client{}
|
||||
resp, err := client.Do(req)
|
||||
if err != nil {
|
||||
common.SysLog("updateTask client.Do err: " + err.Error())
|
||||
return &aliResponse, err, nil
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
|
||||
responseBody, err := io.ReadAll(resp.Body)
|
||||
|
||||
var response AliResponse
|
||||
err = common.Unmarshal(responseBody, &response)
|
||||
if err != nil {
|
||||
common.SysLog("updateTask NewDecoder err: " + err.Error())
|
||||
return &aliResponse, err, nil
|
||||
}
|
||||
|
||||
return &response, nil, responseBody
|
||||
}
|
||||
|
||||
func asyncTaskWait(c *gin.Context, info *relaycommon.RelayInfo, taskID string) (*AliResponse, []byte, error) {
|
||||
waitSeconds := 10
|
||||
step := 0
|
||||
maxStep := 20
|
||||
|
||||
var taskResponse AliResponse
|
||||
var responseBody []byte
|
||||
|
||||
time.Sleep(time.Duration(5) * time.Second)
|
||||
|
||||
for {
|
||||
logger.LogDebug(c, "asyncTaskWait step %d/%d, wait %d seconds", step, maxStep, waitSeconds)
|
||||
step++
|
||||
rsp, err, body := updateTask(info, taskID)
|
||||
responseBody = body
|
||||
if err != nil {
|
||||
logger.LogWarn(c, "asyncTaskWait UpdateTask err: "+err.Error())
|
||||
time.Sleep(time.Duration(waitSeconds) * time.Second)
|
||||
continue
|
||||
}
|
||||
|
||||
if rsp.Output.TaskStatus == "" {
|
||||
return &taskResponse, responseBody, nil
|
||||
}
|
||||
|
||||
switch rsp.Output.TaskStatus {
|
||||
case "FAILED":
|
||||
fallthrough
|
||||
case "CANCELED":
|
||||
fallthrough
|
||||
case "SUCCEEDED":
|
||||
fallthrough
|
||||
case "UNKNOWN":
|
||||
return rsp, responseBody, nil
|
||||
}
|
||||
if step >= maxStep {
|
||||
break
|
||||
}
|
||||
time.Sleep(time.Duration(waitSeconds) * time.Second)
|
||||
}
|
||||
|
||||
return nil, nil, fmt.Errorf("aliAsyncTaskWait timeout")
|
||||
}
|
||||
|
||||
func responseAli2OpenAIImage(c *gin.Context, response *AliResponse, originBody []byte, info *relaycommon.RelayInfo, responseFormat string) *dto.ImageResponse {
|
||||
imageResponse := dto.ImageResponse{
|
||||
Created: info.StartTime.Unix(),
|
||||
}
|
||||
|
||||
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)
|
||||
}
|
||||
|
||||
imageResponse.Metadata = originBody
|
||||
return &imageResponse
|
||||
}
|
||||
|
||||
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
|
||||
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, &aliTaskResponse)
|
||||
if err != nil {
|
||||
return types.NewOpenAIError(err, types.ErrorCodeBadResponseBody, http.StatusInternalServerError), nil
|
||||
}
|
||||
|
||||
if aliTaskResponse.Message != "" {
|
||||
logger.LogError(c, "ali_async_task_failed: "+aliTaskResponse.Message)
|
||||
return types.NewError(errors.New(aliTaskResponse.Message), types.ErrorCodeBadResponse), nil
|
||||
}
|
||||
|
||||
var (
|
||||
aliResponse *AliResponse
|
||||
originRespBody []byte
|
||||
)
|
||||
|
||||
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
|
||||
}
|
||||
}
|
||||
|
||||
if a.IsSyncImageModel {
|
||||
logger.LogDebug(c, "ali_sync_image_result: %s", originRespBody)
|
||||
} else {
|
||||
logger.LogDebug(c, "ali_async_image_result: %s", originRespBody)
|
||||
}
|
||||
|
||||
imageResponses := responseAli2OpenAIImage(c, aliResponse, originRespBody, info, responseFormat)
|
||||
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{}
|
||||
}
|
||||
@@ -0,0 +1,49 @@
|
||||
package ali
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"strings"
|
||||
|
||||
"github.com/QuantumNous/new-api/common"
|
||||
relaycommon "github.com/QuantumNous/new-api/relay/common"
|
||||
"github.com/QuantumNous/new-api/relaykit/dto"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
"github.com/samber/lo"
|
||||
)
|
||||
|
||||
func oaiFormEdit2WanxImageEdit(c *gin.Context, info *relaycommon.RelayInfo, request dto.ImageRequest) (*AliImageRequest, error) {
|
||||
var err error
|
||||
var imageRequest AliImageRequest
|
||||
imageRequest.Model = request.Model
|
||||
imageRequest.ResponseFormat = request.ResponseFormat
|
||||
wanInput := WanImageInput{
|
||||
Prompt: request.Prompt,
|
||||
}
|
||||
|
||||
if err := common.UnmarshalBodyReusable(c, &wanInput); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
if wanInput.Images, err = getImageBase64sFromForm(c, "image"); err != nil {
|
||||
return nil, fmt.Errorf("get image base64s from form failed: %w", err)
|
||||
}
|
||||
//wanParams := WanImageParameters{
|
||||
// N: int(request.N),
|
||||
//}
|
||||
imageRequest.Input = wanInput
|
||||
imageRequest.Parameters = AliImageParameters{
|
||||
N: int(lo.FromPtrOr(request.N, uint(1))),
|
||||
}
|
||||
info.PriceData.AddOtherRatio("n", float64(imageRequest.Parameters.N))
|
||||
|
||||
return &imageRequest, nil
|
||||
}
|
||||
|
||||
func isOldWanModel(modelName string) bool {
|
||||
return strings.Contains(modelName, "wan") &&
|
||||
!lo.SomeBy([]string{"wan2.6", "wan2.7"}, func(v string) bool { return strings.Contains(modelName, v) })
|
||||
}
|
||||
|
||||
func isWanModel(modelName string) bool {
|
||||
return strings.Contains(modelName, "wan")
|
||||
}
|
||||
@@ -0,0 +1,75 @@
|
||||
package ali
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"io"
|
||||
"net/http"
|
||||
|
||||
relaycommon "github.com/QuantumNous/new-api/relay/common"
|
||||
"github.com/QuantumNous/new-api/relaykit/dto"
|
||||
"github.com/QuantumNous/new-api/relaykit/types"
|
||||
"github.com/QuantumNous/new-api/service"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
)
|
||||
|
||||
func ConvertRerankRequest(request dto.RerankRequest) *AliRerankRequest {
|
||||
returnDocuments := request.ReturnDocuments
|
||||
if returnDocuments == nil {
|
||||
t := true
|
||||
returnDocuments = &t
|
||||
}
|
||||
return &AliRerankRequest{
|
||||
Model: request.Model,
|
||||
Input: AliRerankInput{
|
||||
Query: request.Query,
|
||||
Documents: request.Documents,
|
||||
},
|
||||
Parameters: AliRerankParameters{
|
||||
TopN: request.TopN,
|
||||
ReturnDocuments: returnDocuments,
|
||||
},
|
||||
}
|
||||
}
|
||||
|
||||
func RerankHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*types.NewAPIError, *dto.Usage) {
|
||||
responseBody, err := io.ReadAll(resp.Body)
|
||||
if err != nil {
|
||||
return types.NewOpenAIError(err, types.ErrorCodeReadResponseBodyFailed, http.StatusInternalServerError), nil
|
||||
}
|
||||
service.CloseResponseBodyGracefully(resp)
|
||||
|
||||
var aliResponse AliRerankResponse
|
||||
err = json.Unmarshal(responseBody, &aliResponse)
|
||||
if err != nil {
|
||||
return types.NewOpenAIError(err, types.ErrorCodeBadResponseBody, http.StatusInternalServerError), nil
|
||||
}
|
||||
|
||||
if aliResponse.Code != "" {
|
||||
return types.WithOpenAIError(types.OpenAIError{
|
||||
Message: aliResponse.Message,
|
||||
Type: aliResponse.Code,
|
||||
Param: aliResponse.RequestId,
|
||||
Code: aliResponse.Code,
|
||||
}, resp.StatusCode), nil
|
||||
}
|
||||
|
||||
usage := dto.Usage{
|
||||
PromptTokens: aliResponse.Usage.TotalTokens,
|
||||
CompletionTokens: 0,
|
||||
TotalTokens: aliResponse.Usage.TotalTokens,
|
||||
}
|
||||
rerankResponse := dto.RerankResponse{
|
||||
Results: aliResponse.Output.Results,
|
||||
Usage: usage,
|
||||
}
|
||||
|
||||
jsonResponse, err := json.Marshal(rerankResponse)
|
||||
if err != nil {
|
||||
return types.NewError(err, types.ErrorCodeBadResponseBody), nil
|
||||
}
|
||||
c.Writer.Header().Set("Content-Type", "application/json")
|
||||
c.Writer.WriteHeader(resp.StatusCode)
|
||||
c.Writer.Write(jsonResponse)
|
||||
return nil, &usage
|
||||
}
|
||||
@@ -0,0 +1,36 @@
|
||||
package ali
|
||||
|
||||
import (
|
||||
"github.com/QuantumNous/new-api/relaykit/dto"
|
||||
"github.com/samber/lo"
|
||||
)
|
||||
|
||||
// https://help.aliyun.com/document_detail/613695.html?spm=a2c4g.2399480.0.0.1adb778fAdzP9w#341800c0f8w0r
|
||||
|
||||
const EnableSearchModelSuffix = "-internet"
|
||||
|
||||
func requestOpenAI2Ali(request dto.GeneralOpenAIRequest, upstreamModelName string) *dto.GeneralOpenAIRequest {
|
||||
modelName := upstreamModelName
|
||||
if modelName == "" {
|
||||
modelName = request.Model
|
||||
}
|
||||
if !dto.IsQwenThinkingBudgetModel(modelName) {
|
||||
request.ThinkingBudget = nil
|
||||
}
|
||||
|
||||
// DashScope rejects top_p at the 0 and 1 boundaries, so an explicit value is
|
||||
// clamped into the open interval. The clamp stays at two decimals because
|
||||
// some models on the platform reject a third decimal with
|
||||
// "top_p参数非法：限制小数点[2]位".
|
||||
//
|
||||
// A request that omits top_p is left untouched: injecting a value would
|
||||
// silently replace the model's own default with near-greedy decoding.
|
||||
if request.TopP != nil {
|
||||
if *request.TopP >= 1 {
|
||||
request.TopP = lo.ToPtr(0.99)
|
||||
} else if *request.TopP <= 0 {
|
||||
request.TopP = lo.ToPtr(0.01)
|
||||
}
|
||||
}
|
||||
return &request
|
||||
}
|
||||
@@ -0,0 +1,59 @@
|
||||
package ali
|
||||
|
||||
import (
|
||||
"testing"
|
||||
|
||||
"github.com/QuantumNous/new-api/relaykit/dto"
|
||||
"github.com/samber/lo"
|
||||
"github.com/stretchr/testify/assert"
|
||||
)
|
||||
|
||||
func TestRequestOpenAI2AliTopP(t *testing.T) {
|
||||
tests := []struct {
|
||||
name string
|
||||
topP *float64
|
||||
want *float64
|
||||
}{
|
||||
{
|
||||
name: "omitted top_p is not injected",
|
||||
topP: nil,
|
||||
want: nil,
|
||||
},
|
||||
{
|
||||
name: "in-range top_p is preserved",
|
||||
topP: lo.ToPtr(0.8),
|
||||
want: lo.ToPtr(0.8),
|
||||
},
|
||||
{
|
||||
name: "top_p of 1 is clamped to two decimals",
|
||||
topP: lo.ToPtr(1.0),
|
||||
want: lo.ToPtr(0.99),
|
||||
},
|
||||
{
|
||||
name: "top_p above 1 is clamped to two decimals",
|
||||
topP: lo.ToPtr(1.5),
|
||||
want: lo.ToPtr(0.99),
|
||||
},
|
||||
{
|
||||
name: "top_p of 0 is clamped to two decimals",
|
||||
topP: lo.ToPtr(0.0),
|
||||
want: lo.ToPtr(0.01),
|
||||
},
|
||||
{
|
||||
name: "negative top_p is clamped to two decimals",
|
||||
topP: lo.ToPtr(-0.3),
|
||||
want: lo.ToPtr(0.01),
|
||||
},
|
||||
}
|
||||
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
got := requestOpenAI2Ali(dto.GeneralOpenAIRequest{
|
||||
Model: "qwen-plus",
|
||||
TopP: tt.topP,
|
||||
}, "qwen-plus")
|
||||
|
||||
assert.Equal(t, tt.want, got.TopP)
|
||||
})
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user