{
  "version": "1.0",
  "name": "漫画生成工作流验证",
  "description": "验证小说到漫画分镜提示词的AI处理流程",
  "variables": [
    {
      "variable": "input_novel",
      "value_type": "string",
      "required": true,
      "label": "小说原文",
      "description": "用户输入的小说短段落（300-1000字）"
    },
    {
      "variable": "edited_text",
      "value_type": "string",
      "required": false,
      "label": "编辑后文本",
      "description": "AI编辑润色后的叙述文本"
    },
    {
      "variable": "storyboard_markdown",
      "value_type": "string",
      "required": false,
      "label": "分镜提示词",
      "description": "Markdown格式的分镜描述和图片提示词"
    }
  ],
  "nodes": [
    {
      "id": "editor_node",
      "type": "llm",
      "name": "小说编辑AI",
      "description": "润色和增强小说文本描述",
      "config": {
        "model": {
          "provider": "deepseek",
          "name": "deepseek-r1",
          "mode": "chat",
          "temperature": 0.7,
          "max_tokens": 2000
        },
        "prompt": {
          "system": "你是一个专业的小说编辑。你的任务是将用户提供的小说段落进行润色和增强，使其更适合转换为视觉分镜。\n\n要求：\n1. 保持原有情节和人物设定不变\n2. 增强场景描述、角色动作和情感表达\n3. 让文本更具画面感和戏剧张力\n4. 输出保持中文，长度控制在原文本的1.2-1.5倍",
          "user": "请润色以下小说段落，使其更适合视觉化：\n\n{{input_novel}}"
        },
        "outputs": {
          "edited_text": "{{llm_response}}"
        }
      }
    },
    {
      "id": "storyboard_node", 
      "type": "llm",
      "name": "分镜AI",
      "description": "将叙述文本转换为分镜提示词",
      "config": {
        "model": {
          "provider": "deepseek",
          "name": "deepseek-r1",
          "mode": "chat",
          "temperature": 0.6,
          "max_tokens": 3000
        },
        "prompt": {
          "system": "你是一个专业的漫画分镜师。请将叙述文本转换为3-5个分镜描述，每个分镜包含视觉元素和图片生成提示词。\n\n输出格式要求（Markdown）：\n# 分镜分析\n\n## 场景概览\n[简要描述整体场景]\n\n## 分镜详情\n\n### 分镜1：[标题]\n**场景描述：** [详细视觉描述]\n**角色动作：** [主要角色动作]\n**情感氛围：** [情感色彩]\n**图片提示词：** [用于图片生成的提示词，包含风格、构图等]\n\n### 分镜2：[标题]\n...\n\n## 风格建议\n[推荐的图片生成风格，如：动漫风格、写实插画、水彩等]\n\n注意：图片提示词要具体、详细，包含构图、灯光、风格等要素。",
          "user": "请将以下文本转换为漫画分镜提示词：\n\n{{edited_text}}"
        },
        "outputs": {
          "storyboard_markdown": "{{llm_response}}"
        }
      }
    }
  ],
  "edges": [
    {
      "source": "input_novel",
      "sourceHandle": "output",
      "target": "editor_node",
      "targetHandle": "input"
    },
    {
      "source": "editor_node", 
      "sourceHandle": "edited_text",
      "target": "storyboard_node",
      "targetHandle": "edited_text"
    },
    {
      "source": "storyboard_node",
      "sourceHandle": "storyboard_markdown", 
      "target": "output",
      "targetHandle": "input"
    }
  ],
  "outputs": [
    {
      "variable": "storyboard_markdown",
      "label": "分镜提示词",
      "description": "最终生成的Markdown格式分镜提示词"
    }
  ]
}