ai-chatbot/lib/agents/workflow-material-appearance.ts

import "server-only";

import { generateText } from "ai";
import { getLanguageModelById } from "@/lib/ai/dynamic-provider";
import { callQwen3Vl } from "@/lib/ai/qwen3-vl";
import { getEnabledAiModelsWithProvider } from "@/lib/db/queries";

const QWEN_VL_MODEL_MATCHER =
  /(qwen|千问|通义).*(vl|vision)|qwen3-vl|qwen-vl/i;

function sanitizeAppearanceSummary(value: unknown, maxLength = 220) {
  return String(value || "").replace(/\s+/g, " ").trim().slice(0, maxLength);
}

function extractJsonObjectText(text: string): string | null {
  const normalized = String(text || "").trim();
  if (!normalized) {
    return null;
  }

  const fencedMatch = normalized.match(/```(?:json)?\s*([\s\S]*?)```/i);
  if (fencedMatch?.[1]) {
    const fencedText = fencedMatch[1].trim();
    if (fencedText.startsWith("{") && fencedText.endsWith("}")) {
      return fencedText;
    }
  }

  if (normalized.startsWith("{") && normalized.endsWith("}")) {
    return normalized;
  }

  const firstBraceIndex = normalized.indexOf("{");
  const lastBraceIndex = normalized.lastIndexOf("}");
  if (firstBraceIndex >= 0 && lastBraceIndex > firstBraceIndex) {
    return normalized.slice(firstBraceIndex, lastBraceIndex + 1);
  }

  return null;
}

function parseAppearanceSummary(text: string) {
  const jsonText = extractJsonObjectText(text);
  if (jsonText) {
    try {
      const parsed = JSON.parse(jsonText) as Record<string, unknown>;
      const summary = sanitizeAppearanceSummary(
        parsed.appearanceSummary || parsed.summary || parsed.description
      );
      if (summary) {
        return summary;
      }
    } catch (error) {
      console.warn(
        "[Workflow Material Appearance] Failed to parse JSON response:",
        error
      );
    }
  }

  return sanitizeAppearanceSummary(
    String(text || "")
      .replace(/```(?:json)?/gi, "")
      .replace(/```/g, "")
  );
}

function buildAppearancePrompt() {
  return [
    "你是电商商品白底图识别器。",
    "请只基于商品本体，识别并描述产品的外观结构和颜色，用于后续生图时保持产品结构不要改变。",
    "描述重点：1. 产品品类与整体轮廓；2. 关键结构件及数量关系，例如层板、格口、门板、抽屉、腿、扶手、靠背、边框、台面、开孔、开口方向；3. 主色、副色与颜色分布。",
    "不要分析或描述材质、纹理、表面工艺、木纹、金属感、布艺感，也不要描述背景、光影、拍摄角度、水印、文案，更不要输出营销形容词。",
    '只输出 JSON：{"appearanceSummary":"..."}',
    "appearanceSummary 必须是中文，长度 60 到 120 字，具体、克制、可直接作为生图约束。",
  ].join("\n");
}

async function analyzeWithConfiguredQwenVisionModel(params: {
  imageUrl: string;
  prompt: string;
}) {
  const configuredModels = await getEnabledAiModelsWithProvider().catch((error) => {
    console.warn(
      "[Workflow Material Appearance] Failed to load configured models:",
      error
    );
    return [];
  });
  const configuredQwenVisionModel = configuredModels.find(
    ({ model }) =>
      model.supportsVision &&
      !model.isSubModel &&
      QWEN_VL_MODEL_MATCHER.test(`${model.displayName} ${model.modelId}`)
  );

  if (!configuredQwenVisionModel) {
    return null;
  }

  try {
    const languageModel = await getLanguageModelById(
      configuredQwenVisionModel.model.id
    );
    if (!languageModel) {
      return null;
    }

    const { text } = await generateText({
      model: languageModel,
      messages: [
        {
          role: "user",
          content: [
            {
              type: "text",
              text: params.prompt,
            },
            {
              type: "image",
              image: params.imageUrl,
            },
          ],
        },
      ],
      temperature: 0,
      maxTokens: 220,
      abortSignal: AbortSignal.timeout(25_000),
    } as any);

    return parseAppearanceSummary(text);
  } catch (error) {
    console.warn(
      "[Workflow Material Appearance] Configured Qwen-VL analysis failed:",
      error
    );
    return null;
  }
}

export async function analyzeWorkflowMaterialAppearanceFromImage(imageUrl: string) {
  const normalizedUrl = String(imageUrl || "").trim();
  if (!/^https?:\/\//i.test(normalizedUrl)) {
    return null;
  }

  const prompt = buildAppearancePrompt();
  const configuredSummary = await analyzeWithConfiguredQwenVisionModel({
    imageUrl: normalizedUrl,
    prompt,
  });
  if (configuredSummary) {
    return configuredSummary;
  }

  try {
    const result = await callQwen3Vl({
      imageUrl: normalizedUrl,
      prompt,
      temperature: 0,
      maxTokens: 220,
      timeoutMs: 25_000,
      retries: 1,
    });
    return parseAppearanceSummary(result.text);
  } catch (error) {
    console.warn(
      "[Workflow Material Appearance] Fallback Qwen3-VL analysis failed:",
      error
    );
    return null;
  }
}