Initial open-source release (MIT): image2api AI gateway

Full Go backend + Vue 3 frontend, OpenAI-compatible API, multi-provider account pools, billing/admin, Docker one-command deploy with auto HTTPS. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-25 22:59:04 +08:00
commit 606caaf047
142 changed files with 33648 additions and 0 deletions
@@ -0,0 +1,603 @@
+package handler
+
+import (
+	"errors"
+	"net/http"
+
+	"backend/internal/service"
+	"github.com/gin-gonic/gin"
+)
+
+type UserGenerationHandler struct {
+	userGen *service.UserGenerationService
+	admin   *service.AdminReadService
+}
+
+func NewUserGenerationHandler(userGen *service.UserGenerationService, admin *service.AdminReadService) *UserGenerationHandler {
+	return &UserGenerationHandler{
+		userGen: userGen,
+		admin:   admin,
+	}
+}
+
+// MyImages returns the current user's own recently generated images (scoped to
+// their owner directory) — used by the showcase "选择已生成" picker so an admin
+// only sees their own images, not everyone's.
+func (h *UserGenerationHandler) MyImages(c *gin.Context) {
+	user := currentUser(c)
+	if user == nil {
+		c.JSON(http.StatusUnauthorized, gin.H{"detail": "未登录或会话已过期"})
+		return
+	}
+	items, err := h.admin.RecentImagesOwned(c.Request.Context(), service.OwnerDir(user), 60)
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"detail": "failed to load images"})
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{"data": items})
+}
+
+func (h *UserGenerationHandler) Generate(c *gin.Context) {
+	user := currentUser(c)
+	if user == nil {
+		c.JSON(http.StatusUnauthorized, gin.H{"detail": "未登录或会话已过期"})
+		return
+	}
+
+	var body struct {
+		Model           string   `json:"model"`
+		Prompt          string   `json:"prompt"`
+		Ratio           string   `json:"ratio"`
+		Resolution      string   `json:"resolution"`
+		Duration        string   `json:"duration"`
+		ReferenceImages []string `json:"reference_images"`
+	}
+	if err := c.ShouldBindJSON(&body); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"detail": "invalid request body"})
+		return
+	}
+
+	resp, err := h.userGen.Generate(c.Request.Context(), user, service.UserGenerateRequest{
+		Model:           body.Model,
+		Prompt:          body.Prompt,
+		Ratio:           body.Ratio,
+		Resolution:      body.Resolution,
+		Duration:        body.Duration,
+		ReferenceImages: body.ReferenceImages,
+	})
+	if err != nil {
+		switch {
+		case errors.Is(err, service.ErrUnknownModel):
+			c.JSON(http.StatusNotFound, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrUnsupportedParams):
+			c.JSON(http.StatusBadRequest, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrInsufficientFunds):
+			c.JSON(http.StatusPaymentRequired, gin.H{"detail": "积分不足"})
+		case errors.Is(err, service.ErrNoProviderAccount):
+			c.JSON(http.StatusServiceUnavailable, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrProviderAuth), errors.Is(err, service.ErrProviderTemporary):
+			c.JSON(http.StatusServiceUnavailable, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrProviderQuota):
+			c.JSON(http.StatusTooManyRequests, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrConcurrencyFull):
+			c.JSON(http.StatusTooManyRequests, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrProviderExecution):
+			c.JSON(http.StatusBadGateway, gin.H{"detail": err.Error()})
+		default:
+			if err.Error() == "已有正在生成的任务,请稍候" {
+				c.JSON(http.StatusConflict, gin.H{"detail": err.Error()})
+				return
+			}
+			c.JSON(http.StatusBadRequest, gin.H{"detail": err.Error()})
+		}
+		return
+	}
+	c.JSON(http.StatusOK, resp)
+}
+
+func (h *UserGenerationHandler) Test(c *gin.Context) {
+	user := currentUser(c)
+	if user == nil {
+		c.JSON(http.StatusUnauthorized, gin.H{"detail": "未登录或会话已过期"})
+		return
+	}
+	if user.Role != "admin" {
+		c.JSON(http.StatusForbidden, gin.H{"detail": "需要管理员权限"})
+		return
+	}
+
+	var body struct {
+		Model           string   `json:"model"`
+		Prompt          string   `json:"prompt"`
+		Ratio           string   `json:"ratio"`
+		Resolution      string   `json:"resolution"`
+		Duration        string   `json:"duration"`
+		ReferenceImages []string `json:"reference_images"`
+	}
+	if err := c.ShouldBindJSON(&body); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"detail": "invalid request body"})
+		return
+	}
+
+	resp, err := h.userGen.AdminTest(c.Request.Context(), user, service.UserGenerateRequest{
+		Model:           body.Model,
+		Prompt:          body.Prompt,
+		Ratio:           body.Ratio,
+		Resolution:      body.Resolution,
+		Duration:        body.Duration,
+		ReferenceImages: body.ReferenceImages,
+	})
+	if err != nil {
+		switch {
+		case errors.Is(err, service.ErrUnknownModel):
+			c.JSON(http.StatusNotFound, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrUnsupportedParams):
+			c.JSON(http.StatusBadRequest, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrProviderQuota):
+			c.JSON(http.StatusTooManyRequests, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrConcurrencyFull):
+			c.JSON(http.StatusTooManyRequests, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrNoProviderAccount):
+			c.JSON(http.StatusServiceUnavailable, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrProviderAuth), errors.Is(err, service.ErrProviderTemporary):
+			c.JSON(http.StatusServiceUnavailable, gin.H{"detail": err.Error()})
+		case errors.Is(err, service.ErrProviderExecution):
+			c.JSON(http.StatusBadGateway, gin.H{"detail": err.Error()})
+		default:
+			c.JSON(http.StatusBadRequest, gin.H{"detail": err.Error()})
+		}
+		return
+	}
+	c.JSON(http.StatusOK, resp)
+}
+
+func (h *UserGenerationHandler) MyJobs(c *gin.Context) {
+	user := currentUser(c)
+	if user == nil {
+		c.JSON(http.StatusOK, gin.H{"pending": nil, "latest": nil})
+		return
+	}
+	data, err := h.userGen.MyJobs(c.Request.Context(), user, c.Query("source"))
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"detail": "failed to load jobs"})
+		return
+	}
+	c.JSON(http.StatusOK, data)
+}
+
+func (h *UserGenerationHandler) Logs(c *gin.Context) {
+	user := currentUser(c)
+	if user == nil {
+		c.JSON(http.StatusUnauthorized, gin.H{"detail": "未登录或会话已过期"})
+		return
+	}
+
+	limit := parseInt(c.Query("limit"), 50)
+	offset := parseInt(c.Query("offset"), 0)
+	kind := c.Query("kind")
+	status := c.Query("status")
+	// Secure-by-default: always scope to the caller's OWN records. This endpoint
+	// serves the front-end 日志 / 创作记录 pages, so an admin viewing their personal
+	// records must NOT see other users' work. Only an admin who explicitly opts
+	// into the full view (?scope=all — the admin 日志 page) sees everyone's logs.
+	// API-key ("v1") usage IS included for the caller's own records so the user
+	// can audit their key's calls on /mylogs; the image-only 创作记录 gallery still
+	// hides them client-side (they have no stored file).
+	userID := user.ID
+	excludeSource := ""
+	if user.Role == "admin" && c.Query("scope") == "all" {
+		userID = ""
+	}
+	// 来源筛选: "v1" = API key, "user" = 前台画图, "admin" = 测试模型. 始终生效 ——
+	// 普通用户已被 userID 限定为本人记录,按来源服务端筛选 + 分页(/mylogs 翻全部历史)。
+	source := c.Query("source")
+	// 创作记录 gallery passes has_file=1 so server-side pagination counts only
+	// rows with real media (success + stored file), not failed/pending events.
+	hasFile := c.Query("has_file") == "1" || c.Query("has_file") == "true"
+
+	items, total, stats, err := h.admin.Logs(c.Request.Context(), limit, offset, kind, status, nil, userID, excludeSource, source, hasFile)
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"detail": "failed to load logs"})
+		return
+	}
+	// Resolve user_id -> display name (mirrors admin.py / AdminReadHandler.Logs).
+	// Without this the log table showed every row as "匿名".
+	nameByID, err := h.admin.UserNameMap(c.Request.Context())
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"detail": "failed to load logs"})
+		return
+	}
+
+	out := make([]gin.H, 0, len(items))
+	for _, item := range items {
+		var userName any
+		if item.UserID == "" {
+			userName = "匿名"
+		} else if name, ok := nameByID[item.UserID]; ok {
+			userName = name
+		} else {
+			userName = item.UserID
+		}
+		out = append(out, gin.H{
+			"id":         item.ID,
+			"ts":         item.TS.Unix(),
+			"kind":       item.Kind,
+			"status":     item.Status,
+			"model":      item.Model,
+			"provider":   item.Provider,
+			"prompt":     item.Prompt,
+			"ratio":      item.Ratio,
+			"resolution": item.Resolution,
+			"duration":   item.Duration,
+			"refs":       item.Refs,
+			"source":     emptyStringNil(item.Source),
+			"user_id":    emptyStringNil(item.UserID),
+			"user_name":  userName,
+			"cost":       item.Cost,
+			"elapsed_ms": item.ElapsedMS,
+			"file":       emptyStringNil(item.File),
+			"error":      emptyStringNil(item.Error),
+			"created_at": unixSec(item.CreatedAt),
+			"updated_at": unixSec(item.UpdatedAt),
+		})
+	}
+
+	c.JSON(http.StatusOK, gin.H{
+		"data":   out,
+		"total":  total,
+		"limit":  limit,
+		"offset": offset,
+		"stats":  stats,
+	})
+}
+
+func (h *UserGenerationHandler) VideoPresets(c *gin.Context) {
+	c.JSON(http.StatusOK, gin.H{
+		"data": []gin.H{
+			{
+				"key":                  "gemini-veo31",
+				"label":                "Veo31",
+				"type":                 "video",
+				"provider":             "adobe",
+				"durations":            []string{"4s", "6s", "8s"},
+				"ratios":               []string{"16x9", "9x16"},
+				"resolutions":          []string{"720p", "1080p"},
+				"max_reference_images": 2,
+				"reference_mode":       "frame",
+			},
+			{
+				"key":                  "firefly-ray",
+				"label":                "Luma Ray",
+				"type":                 "video",
+				"provider":             "adobe",
+				"durations":            []string{"5s", "10s"},
+				"ratios":               []string{"21:9", "16:9", "4:3", "1:1", "3:4", "9:16", "9:21"},
+				"resolutions":          []string{"720p"},
+				"max_reference_images": 2,
+				"reference_mode":       "frame",
+			},
+			{
+				"key":                  "firefly-video",
+				"label":                "Firefly Video",
+				"type":                 "video",
+				"provider":             "adobe",
+				"durations":            []string{"5s"},
+				"ratios":               []string{"16:9", "1:1", "9:16"},
+				"resolutions":          []string{"540p", "720p", "1080p"},
+				"max_reference_images": 2,
+				"reference_mode":       "frame",
+			},
+			{
+				"key":                  "runway-gen4-turbo",
+				"label":                "Runway Gen-4 Turbo",
+				"type":                 "video",
+				"provider":             "runway",
+				"durations":            []string{"5s", "10s"},
+				"ratios":               []string{"16:9", "9:16", "1:1", "4:3", "3:4", "21:9"},
+				"resolutions":          []string{"2K"},
+				"max_reference_images": 1,
+				"reference_mode":       "frame",
+				// Runway is strictly image-to-video — a first-frame image is required
+				// (no text2video), so the UI must block submit without one.
+				"requires_reference": true,
+			},
+		},
+	})
+}
+
+func (h *UserGenerationHandler) Catalog(c *gin.Context) {
+	items, err := h.catalogEntries(c)
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"detail": "failed to load catalog"})
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{
+		"data": items,
+	})
+}
+
+func (h *UserGenerationHandler) Models(c *gin.Context) {
+	items, err := h.publicModels()
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"detail": "failed to load models"})
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{"data": items})
+}
+
+func (h *UserGenerationHandler) catalogEntries(c *gin.Context) ([]gin.H, error) {
+	items := []gin.H{
+		{
+			"id":             "gpt-image-2",
+			"provider":       "chatgpt",
+			"type":           "image",
+			// ChatGPT web backend only reliably produces 1K and honors a limited
+			// ratio set; size params are advisory prompt hints. Mirrors the Python
+			// reference (providers/chatgpt/provider.py) — do not offer 2K/4K.
+			"ratios":         []string{"1:1", "16:9", "9:16", "4:3", "3:4"},
+			"resolutions":    []string{"1K"},
+			"image_to_image": true,
+			"max_reference_images": 3,
+			"description":    "ChatGPT image generation",
+		},
+		{
+			"id":                   "firefly-gpt-image-2",
+			"provider":             "adobe",
+			"type":                 "image",
+			"ratios":               []string{"1:1", "5:4", "9:16", "21:9", "16:9", "4:3", "3:2", "4:5", "3:4", "2:3"},
+			"resolutions":          []string{"1K", "2K", "4K"},
+			"image_to_image":       true,
+			"max_reference_images": 6,
+			"description":          "Adobe Firefly GPT Image",
+		},
+		{
+			"id":             "firefly-image-5",
+			"provider":       "adobe",
+			"type":           "image",
+			"ratios":         []string{"1:1", "16:9", "9:16", "4:3", "3:4"},
+			"resolutions":    []string{"1K", "2K"},
+			"image_to_image": true,
+			"description":    "Adobe Firefly Image 5",
+		},
+		{
+			"id":                   "flux-kontext-max",
+			"provider":             "adobe",
+			"type":                 "image",
+			"ratios":               []string{"1:1", "16:9", "9:16", "4:3", "3:4"},
+			"resolutions":          []string{"1K"},
+			"image_to_image":       true,
+			"max_reference_images": 4,
+			"description":          "Adobe Flux Kontext Max",
+		},
+		{
+			"id":             "nano-banana-2",
+			"provider":       "adobe",
+			"type":           "image",
+			"ratios":         []string{"1:1", "16:9", "9:16", "4:3", "3:4"},
+			"resolutions":    []string{"1K"},
+			"image_to_image": true,
+			"description":    "Adobe Gemini Flash Nano Banana",
+		},
+		{
+			"id":                   "gemini-veo31",
+			"provider":             "adobe",
+			"type":                 "video",
+			"ratios":               []string{"16x9", "9x16"},
+			"resolutions":          []string{"720p", "1080p"},
+			"durations":            []string{"4s", "6s", "8s"},
+			"max_reference_images": 2,
+			"reference_mode":       "frame",
+			"description":          "Veo31 video",
+		},
+		{
+			"id":                   "firefly-ray",
+			"provider":             "adobe",
+			"type":                 "video",
+			"ratios":               []string{"21:9", "16:9", "4:3", "1:1", "3:4", "9:16", "9:21"},
+			"resolutions":          []string{"720p"},
+			"durations":            []string{"5s", "10s"},
+			"max_reference_images": 2,
+			"reference_mode":       "frame",
+			"description":          "Luma Ray video",
+		},
+		{
+			"id":                   "firefly-video",
+			"provider":             "adobe",
+			"type":                 "video",
+			"ratios":               []string{"16:9", "1:1", "9:16"},
+			"resolutions":          []string{"540p", "720p", "1080p"},
+			"durations":            []string{"5s"},
+			"max_reference_images": 2,
+			"reference_mode":       "frame",
+			"description":          "Adobe Firefly Video",
+		},
+		{
+			"id":                   "runway-gen4-turbo",
+			"provider":             "runway",
+			"type":                 "video",
+			"ratios":               []string{"16:9", "9:16", "1:1", "4:3", "3:4", "21:9"},
+			"resolutions":          []string{"2K"},
+			"durations":            []string{"5s", "10s"},
+			"max_reference_images": 1,
+			"reference_mode":       "frame",
+			"description":          "Runway Gen-4 Turbo video (图生视频)",
+		},
+		{
+			"id":                   "seedream-4.5",
+			"provider":             "leonardo",
+			"type":                 "image",
+			"ratios":               []string{"2:3", "1:1", "16:9", "4:3", "4:5", "9:16", "2:1"},
+			"resolutions":          []string{"2K", "4K"},
+			"image_to_image":       true,
+			"max_reference_images": 6,
+			"description":          "Leonardo Seedream 4.5 (生图 / 图生图)",
+		},
+		{
+			"id":                   "flux-klein-2",
+			"provider":             "krea",
+			"type":                 "image",
+			"ratios":               []string{"1:1", "4:3", "3:4", "16:9", "9:16"},
+			"resolutions":          []string{"1K", "2K"},
+			"image_to_image":       true,
+			"max_reference_images": 4,
+			"description":          "Krea Flux Klein (生图 / 图生图)",
+		},
+		{
+			"id":                   "imagine-1.5",
+			"provider":             "imagine",
+			"type":                 "image",
+			"ratios":               []string{"1:3", "9:16", "2:3", "3:4", "1:1", "4:3", "3:2", "16:9", "3:1"},
+			"resolutions":          []string{"2K"},
+			"max_reference_images": 0,
+			"description":          "Imagine 1.5 (文生图)",
+		},
+		{
+			"id":                   "imagine-1.5pro",
+			"provider":             "imagine",
+			"type":                 "image",
+			"ratios":               []string{"1:3", "9:16", "2:3", "3:4", "1:1", "4:3", "3:2", "16:9", "3:1"},
+			"resolutions":          []string{"4K"},
+			"max_reference_images": 0,
+			"description":          "Imagine 1.5 Pro (文生图)",
+		},
+	}
+	existing := map[string]bool{}
+	if h.admin != nil {
+		models, err := h.admin.Models(c.Request.Context())
+		if err != nil {
+			return nil, err
+		}
+		for _, item := range models {
+			existing[item.ID] = true
+		}
+	}
+	for i := range items {
+		items[i]["added"] = existing[items[i]["id"].(string)]
+	}
+	return items, nil
+}
+
+func (h *UserGenerationHandler) publicModels() ([]gin.H, error) {
+	items := []gin.H{
+		{
+			"id":          "gpt-image-2",
+			"provider":    "chatgpt",
+			"kind":        "image",
+			// See catalogEntries — ChatGPT only reliably does 1K and a limited
+			// ratio set; matches the Python reference. Keep both lists in sync.
+			"ratios":      []string{"1:1", "16:9", "9:16", "4:3", "3:4"},
+			"resolutions": []string{"1K"},
+			"description": "ChatGPT image generation",
+			"stub":        false,
+		},
+		{
+			"id":          "firefly-gpt-image-2",
+			"provider":    "adobe",
+			"kind":        "image",
+			"ratios":      []string{"1:1", "5:4", "9:16", "21:9", "16:9", "4:3", "3:2", "4:5", "3:4", "2:3"},
+			"resolutions": []string{"1K", "2K", "4K"},
+			"description": "Adobe Firefly GPT Image",
+			"stub":        false,
+		},
+		{
+			"id":          "firefly-image-5",
+			"provider":    "adobe",
+			"kind":        "image",
+			"ratios":      []string{"1:1", "16:9", "9:16", "4:3", "3:4"},
+			"resolutions": []string{"1K", "2K"},
+			"description": "Adobe Firefly Image 5",
+			"stub":        false,
+		},
+		{
+			"id":          "flux-kontext-max",
+			"provider":    "adobe",
+			"kind":        "image",
+			"ratios":      []string{"1:1", "16:9", "9:16", "4:3", "3:4"},
+			"resolutions": []string{"1K"},
+			"description": "Adobe Flux Kontext Max",
+			"stub":        false,
+		},
+		{
+			"id":          "nano-banana-2",
+			"provider":    "adobe",
+			"kind":        "image",
+			"ratios":      []string{"1:1", "16:9", "9:16", "4:3", "3:4"},
+			"resolutions": []string{"1K"},
+			"description": "Adobe Gemini Flash Nano Banana",
+			"stub":        false,
+		},
+		{
+			"id":          "gemini-veo31",
+			"provider":    "adobe",
+			"kind":        "video",
+			"ratios":      []string{"16x9", "9x16"},
+			"resolutions": []string{"720p", "1080p"},
+			"description": "Veo31 video",
+			"stub":        false,
+		},
+		{
+			"id":          "firefly-ray",
+			"provider":    "adobe",
+			"kind":        "video",
+			"ratios":      []string{"21:9", "16:9", "4:3", "1:1", "3:4", "9:16", "9:21"},
+			"resolutions": []string{"720p"},
+			"description": "Luma Ray video",
+			"stub":        false,
+		},
+		{
+			"id":          "firefly-video",
+			"provider":    "adobe",
+			"kind":        "video",
+			"ratios":      []string{"16:9", "1:1", "9:16"},
+			"resolutions": []string{"540p", "720p", "1080p"},
+			"description": "Adobe Firefly Video",
+			"stub":        false,
+		},
+		{
+			"id":          "runway-gen4-turbo",
+			"provider":    "runway",
+			"kind":        "video",
+			"ratios":      []string{"16:9", "9:16", "1:1", "4:3", "3:4", "21:9"},
+			"resolutions": []string{"2K"},
+			"description": "Runway Gen-4 Turbo video",
+			"stub":        false,
+		},
+		{
+			"id":          "seedream-4.5",
+			"provider":    "leonardo",
+			"kind":        "image",
+			"ratios":      []string{"2:3", "1:1", "16:9", "4:3", "4:5", "9:16", "2:1"},
+			"resolutions": []string{"2K", "4K"},
+			"description": "Leonardo Seedream 4.5",
+			"stub":        false,
+		},
+		{
+			"id":          "flux-klein-2",
+			"provider":    "krea",
+			"kind":        "image",
+			"ratios":      []string{"1:1", "4:3", "3:4", "16:9", "9:16"},
+			"resolutions": []string{"1K", "2K"},
+			"description": "Krea Flux Klein",
+			"stub":        false,
+		},
+		{
+			"id":          "imagine-1.5",
+			"provider":    "imagine",
+			"kind":        "image",
+			"ratios":      []string{"1:3", "9:16", "2:3", "3:4", "1:1", "4:3", "3:2", "16:9", "3:1"},
+			"resolutions": []string{"2K"},
+			"description": "Imagine 1.5",
+			"stub":        false,
+		},
+		{
+			"id":          "imagine-1.5pro",
+			"provider":    "imagine",
+			"kind":        "image",
+			"ratios":      []string{"1:3", "9:16", "2:3", "3:4", "1:1", "4:3", "3:2", "16:9", "3:1"},
+			"resolutions": []string{"4K"},
+			"description": "Imagine 1.5 Pro",
+			"stub":        false,
+		},
+	}
+	return items, nil
+}