gen2d/backend/internal/service/nodes.go

package service

import (
	"context"
	"fmt"
	"strings"

	"github.com/cloudwego/eino/compose"
)

// promptOptimizerNode 节点：调用 PromptAgent 生成规范提示词，合并风格与重试信息。
// 输入 PipelineInput，输出最终提示词字符串（直接供 AssetGenerator 消费）。
var promptOptimizerNode = compose.InvokableLambda(func(ctx context.Context, in PipelineInput) (string, error) {
	// 合并风格描述，注入原始 Prompt 中
	styleDesc := buildStyleDescription(in.ProjectStyle, in.TaskStyle)
	if styleDesc != "" {
		if in.Prompt != "" {
			in.Prompt = in.Prompt + "。" + styleDesc
		} else {
			in.Prompt = styleDesc
		}
	}

	// 注入重试原因
	if in.RejectReason != "" {
		if in.Prompt != "" {
			in.Prompt = in.Prompt + "。注意修正以下问题：" + in.RejectReason
		} else {
			in.Prompt = "修正以下问题：" + in.RejectReason
		}
	}

	if len(in.Tags) == 0 && in.Prompt == "" {
		return "", fmt.Errorf("pipeline: Prompt and Tags are both empty")
	}

	// 有标签时调用 PromptAgent 优化提示词
	if len(in.Tags) > 0 {
		agentIn := PromptAgentInput{
			Tags:      in.Tags,
			AssetType: in.AssetType,
			Prompt:    in.Prompt,
			UserNote:  in.UserNote,
		}
		output, err := RunPromptAgent(ctx, agentIn)
		if err != nil {
			return "", fmt.Errorf("prompt agent: %w", err)
		}
		return output.Prompt, nil
	}

	// 无标签时直接使用原始 Prompt，补上技术参数段
	return appendTechNotes(in.Prompt, in.AssetType, in.Params), nil
})

// promptOptimizerPreHandler 首次运行时保存输入到 state；重试时注入 RejectReason。
func promptOptimizerPreHandler(ctx context.Context, in PipelineInput, state *PipelineState) (PipelineInput, error) {
	if state.RetryCount == 0 {
		state.Input = in
		reportProgress(ctx, "prompt_builder", 10)
	} else if state.RejectReason != "" {
		in.RejectReason = state.RejectReason
		reportProgress(ctx, "prompt_builder", 30+state.RetryCount*10)
	}
	return in, nil
}

// promptOptimizerPostHandler 将最终提示词写入全局状态。
func promptOptimizerPostHandler(ctx context.Context, out string, state *PipelineState) (string, error) {
	state.FinalPrompt = out
	reportProgress(ctx, "asset_generator", 35)
	return out, nil
}

// AssetGenerator 节点：调用 AI 推理 API 出图。
var assetGeneratorNode = compose.InvokableLambda(func(ctx context.Context, prompt string) ([]GeneratedImage, error) {
	var params AssetParams
	_ = compose.ProcessState[*PipelineState](ctx, func(_ context.Context, state *PipelineState) error {
		params = state.Input.Params
		return nil
	})
	return GenerateImages(ctx, prompt, params)
})

// assetGeneratorPostHandler 将原始图片写入全局状态。
func assetGeneratorPostHandler(ctx context.Context, out []GeneratedImage, state *PipelineState) ([]GeneratedImage, error) {
	state.RawImages = out
	reportProgress(ctx, "quality_supervisor", 60)
	return out, nil
}

// QualitySupervisor 节点：质检，设置路由目标。
var qualitySupervisorNode = compose.InvokableLambda(func(ctx context.Context, images []GeneratedImage) (PipelineInput, error) {
	var input PipelineInput
	err := compose.ProcessState[*PipelineState](ctx, func(_ context.Context, state *PipelineState) error {
		state.RawImages = images

		style := mergeStyle(state.Input.ProjectStyle, state.Input.TaskStyle)
		pass, reason, checkErr := CheckQuality(ctx, images, style)
		if checkErr != nil {
			return fmt.Errorf("quality check: %w", checkErr)
		}

		state.PassQuality = pass
		if !pass {
			state.RejectReason = reason
		}

		if pass {
			state.NextNode = nodeFormatAdapter
			reportProgress(ctx, "format_adapter", 85)
		} else if state.RetryCount >= 3 {
			state.NextNode = nodeFormatAdapter
			reportProgress(ctx, "format_adapter", 85)
		} else {
			state.RetryCount++
			state.NextNode = nodePromptOptimizer
			reportProgress(ctx, "quality_supervisor", 50+state.RetryCount*10)
		}

		input = state.Input
		return nil
	})
	if err != nil {
		return PipelineInput{}, err
	}
	return input, nil
})

// formatAdapterNode 节点：从 state 读取图片，格式转换，组装输出。
var formatAdapterNode = compose.InvokableLambda(func(ctx context.Context, input PipelineInput) (PipelineOutput, error) {
	var images []GeneratedImage
	_ = compose.ProcessState[*PipelineState](ctx, func(_ context.Context, state *PipelineState) error {
		images = state.RawImages
		return nil
	})

	assets := make([]Asset, len(images))
	for i, img := range images {
		assets[i] = Asset{
			Data:   img.Data,
			Format: img.Format,
			URL:    fmt.Sprintf("output/%d.%s", i, img.Format),
		}
	}

	resolution := input.Params.Resolution
	if resolution <= 0 {
		resolution = 64
	}

	metadata := AssetMetadata{
		FrameWidth:  resolution,
		FrameHeight: resolution,
		FrameCount:  len(images),
		Directions:  input.Params.Frames.Directions,
	}

	return PipelineOutput{
		Assets:   assets,
		Metadata: metadata,
	}, nil
})

// buildStyleDescription 将风格键值对转为自然语言描述，供 PromptAgent 注入。
func buildStyleDescription(projectStyle, taskStyle map[string]string) string {
	merged := mergeStyle(projectStyle, taskStyle)
	if len(merged) == 0 {
		return ""
	}
	var parts []string
	for k, v := range merged {
		parts = append(parts, fmt.Sprintf("%s: %s", k, v))
	}
	return "风格约束：" + strings.Join(parts, "；")
}

// appendTechNotes 在无标签（不走 PromptAgent）时补上技术参数段。
func appendTechNotes(prompt, assetType string, params AssetParams) string {
	var parts []string
	if prompt != "" {
		parts = append(parts, prompt)
	}
	parts = append(parts, fmt.Sprintf("素材类型: %s", assetType))
	if params.Resolution > 0 {
		parts = append(parts, fmt.Sprintf("分辨率: %d", params.Resolution))
	}
	if params.Frames.Directions > 0 {
		parts = append(parts, fmt.Sprintf("方向数: %d", params.Frames.Directions))
	}
	if params.Frames.FramesPerDirection > 0 {
		parts = append(parts, fmt.Sprintf("每方向帧数: %d", params.Frames.FramesPerDirection))
	}
	if params.Format != "" {
		parts = append(parts, fmt.Sprintf("输出格式: %s", params.Format))
	}
	return strings.Join(parts, "；")
}

// mergeStyle 合并工程风格与任务风格覆盖，任务同名键覆盖工程。
func mergeStyle(projectStyle, taskStyle map[string]string) map[string]string {
	result := make(map[string]string)
	for k, v := range projectStyle {
		result[k] = v
	}
	for k, v := range taskStyle {
		result[k] = v
	}
	return result
}