gen2d/backend/internal/service/inference.go

package service

import (
	"bytes"
	"context"
	"crypto/rand"
	"encoding/base64"
	"encoding/json"
	"fmt"
	"image"
	"image/color"
	"image/png"
	"io"
	"log"
	"mime/multipart"
	"net/http"
	"strings"

	"gen2d/internal/config"
)

// imgCfg 保存文生图配置，由 main 通过 InitImageGenConfig 注入。
var imgCfg config.ImageGenConfig

// InitImageGenConfig 注入文生图配置。
func InitImageGenConfig(cfg config.ImageGenConfig) {
	imgCfg = cfg
}

// ======================== OpenAI 兼容 Images API 类型 ========================

// imageGenRequest OpenAI 兼容文生图请求体。
type imageGenRequest struct {
	Model   string `json:"model"`
	Prompt  string `json:"prompt"`
	N       int    `json:"n,omitempty"`
	Size    string `json:"size,omitempty"`
	Quality string `json:"quality,omitempty"`
}

// imageGenResponse OpenAI 兼容文生图响应体。
type imageGenResponse struct {
	Data []struct {
		URL     string `json:"url"`
		B64JSON string `json:"b64_json"`
	} `json:"data"`
}

// ======================== 文生图 ========================

// GenerateImages 调用 OpenAI 兼容 Images API 生成图片，未配置 key 时回退 mock。
func GenerateImages(ctx context.Context, prompt string, params AssetParams) ([]GeneratedImage, error) {
	count := 1
	if params.Frames.Directions > 0 && params.Frames.FramesPerDirection > 0 {
		count = params.Frames.Directions * params.Frames.FramesPerDirection
	}

	if imgCfg.APIKey != "" {
		width, height := imgCfg.Width, imgCfg.Height
		if params.Resolution > 0 {
			width = params.Resolution
			height = params.Resolution
		}
		log.Println("[inference] calling image gen API")
		return callImageAPI(ctx, prompt, count, width, height)
	}

	size := params.Resolution
	if size <= 0 {
		size = 64
	}
	log.Println("[inference] image API key not configured, using mock")
	return generateMockImages(size, count)
}

// callImageAPI 调用 OpenAI 兼容 Images API。
func callImageAPI(ctx context.Context, prompt string, count, width, height int) ([]GeneratedImage, error) {
	reqBody := imageGenRequest{
		Model:   imgCfg.Model,
		Prompt:  prompt,
		N:       count,
		Size:    fmt.Sprintf("%dx%d", width, height),
		Quality: imgCfg.Quality,
	}

	body, err := json.Marshal(reqBody)
	if err != nil {
		return nil, fmt.Errorf("marshal request: %w", err)
	}

	url := strings.TrimRight(imgCfg.BaseURL, "/") + "/images/generations"
	req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(body))
	if err != nil {
		return nil, fmt.Errorf("create request: %w", err)
	}
	req.Header.Set("Content-Type", "application/json")
	req.Header.Set("Authorization", "Bearer "+imgCfg.APIKey)

	resp, err := http.DefaultClient.Do(req)
	if err != nil {
		return nil, fmt.Errorf("send request: %w", err)
	}
	defer resp.Body.Close()

	if resp.StatusCode != http.StatusOK {
		b, _ := io.ReadAll(io.LimitReader(resp.Body, 4096))
		return nil, fmt.Errorf("image api error %d: %s", resp.StatusCode, string(b))
	}

	return parseImageResponse(ctx, resp.Body, width, height)
}

// parseImageResponse 解析 OpenAI 兼容图片响应（b64_json 或 url）。
func parseImageResponse(ctx context.Context, r io.Reader, width, height int) ([]GeneratedImage, error) {
	var genResp imageGenResponse
	if err := json.NewDecoder(r).Decode(&genResp); err != nil {
		return nil, fmt.Errorf("decode response: %w", err)
	}

	images := make([]GeneratedImage, 0, len(genResp.Data))
	for i, d := range genResp.Data {
		var data []byte
		switch {
		case d.B64JSON != "":
			var err error
			data, err = base64.StdEncoding.DecodeString(d.B64JSON)
			if err != nil {
				return nil, fmt.Errorf("decode base64 image %d: %w", i, err)
			}
		case d.URL != "":
			var err error
			data, err = downloadImage(ctx, d.URL)
			if err != nil {
				return nil, fmt.Errorf("download image %d: %w", i, err)
			}
		default:
			return nil, fmt.Errorf("image %d: no data or url in response", i)
		}
		images = append(images, GeneratedImage{
			Data:   data,
			Width:  width,
			Height: height,
			Format: "png",
		})
	}
	return images, nil
}

// downloadImage 从 URL 下载图片数据。
func downloadImage(ctx context.Context, url string) ([]byte, error) {
	req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
	if err != nil {
		return nil, fmt.Errorf("create download request: %w", err)
	}
	resp, err := http.DefaultClient.Do(req)
	if err != nil {
		return nil, fmt.Errorf("download: %w", err)
	}
	defer resp.Body.Close()
	if resp.StatusCode != http.StatusOK {
		return nil, fmt.Errorf("download status %d", resp.StatusCode)
	}
	return io.ReadAll(resp.Body)
}

// ======================== 图片编辑 ========================

// EditImages 图片编辑接口，调用 OpenAI 兼容 Images Edits API（multipart/form-data）。
func EditImages(ctx context.Context, imageData []byte, prompt string, count int) ([]GeneratedImage, error) {
	if imgCfg.APIKey == "" {
		return nil, fmt.Errorf("image API key not configured")
	}

	var buf bytes.Buffer
	writer := multipart.NewWriter(&buf)

	part, err := writer.CreateFormFile("image", "image.png")
	if err != nil {
		return nil, fmt.Errorf("create form file: %w", err)
	}
	if _, err := part.Write(imageData); err != nil {
		return nil, fmt.Errorf("write image data: %w", err)
	}

	writer.WriteField("prompt", prompt)
	writer.WriteField("model", imgCfg.Model)
	writer.WriteField("n", fmt.Sprintf("%d", count))
	writer.WriteField("size", fmt.Sprintf("%dx%d", imgCfg.Width, imgCfg.Height))

	if err := writer.Close(); err != nil {
		return nil, fmt.Errorf("close multipart writer: %w", err)
	}

	url := strings.TrimRight(imgCfg.BaseURL, "/") + "/images/edits"
	req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, &buf)
	if err != nil {
		return nil, fmt.Errorf("create request: %w", err)
	}
	req.Header.Set("Content-Type", writer.FormDataContentType())
	req.Header.Set("Authorization", "Bearer "+imgCfg.APIKey)

	resp, err := http.DefaultClient.Do(req)
	if err != nil {
		return nil, fmt.Errorf("send request: %w", err)
	}
	defer resp.Body.Close()

	if resp.StatusCode != http.StatusOK {
		b, _ := io.ReadAll(io.LimitReader(resp.Body, 4096))
		return nil, fmt.Errorf("image edit api error %d: %s", resp.StatusCode, string(b))
	}

	return parseImageResponse(ctx, resp.Body, imgCfg.Width, imgCfg.Height)
}

// ======================== 质检 ========================

var QualityChecker = defaultCheckQuality

func CheckQuality(ctx context.Context, images []GeneratedImage, style map[string]string) (bool, string, error) {
	return QualityChecker(ctx, images, style)
}

func defaultCheckQuality(ctx context.Context, images []GeneratedImage, style map[string]string) (bool, string, error) {
	return true, "", nil
}

func NewCountedQualityChecker(passOnRetry int) func(context.Context, []GeneratedImage, map[string]string) (bool, string, error) {
	var callCount int
	return func(_ context.Context, _ []GeneratedImage, _ map[string]string) (bool, string, error) {
		callCount++
		if callCount >= passOnRetry {
			return true, "", nil
		}
		return false, fmt.Sprintf("style inconsistent (attempt %d)", callCount), nil
	}
}

func AlwaysFailQualityChecker() func(context.Context, []GeneratedImage, map[string]string) (bool, string, error) {
	return func(_ context.Context, _ []GeneratedImage, _ map[string]string) (bool, string, error) {
		return false, "style inconsistent", nil
	}
}

// ======================== Mock 回退 ========================

func generateMockImages(size, count int) ([]GeneratedImage, error) {
	images := make([]GeneratedImage, count)
	for i := 0; i < count; i++ {
		data, err := generateMockImage(size, i)
		if err != nil {
			return nil, fmt.Errorf("generate mock image %d: %w", i, err)
		}
		images[i] = GeneratedImage{Data: data, Width: size, Height: size, Format: "png"}
	}
	return images, nil
}

func generateMockImage(size int, seed int) ([]byte, error) {
	img := image.NewRGBA(image.Rect(0, 0, size, size))
	r := uint8((seed*47 + 13) % 256)
	g := uint8((seed*83 + 37) % 256)
	b := uint8((seed*61 + 71) % 256)
	for y := 0; y < size; y++ {
		for x := 0; x < size; x++ {
			img.Set(x, y, color.RGBA{R: r, G: g, B: b, A: 255})
		}
	}
	var buf bytes.Buffer
	if err := png.Encode(&buf, img); err != nil {
		return nil, err
	}
	return buf.Bytes(), nil
}

func generateRandomBytes(n int) ([]byte, error) {
	b := make([]byte, n)
	_, err := rand.Read(b)
	return b, err
}