ai-chatbot/lib/ai/qwen3-vl.ts

import "server-only";

import { getEnabledAiProviders } from "@/lib/db/queries";

const QWEN3_VL_MODELS = [
  "qwen3-vl-plus",
  "qwen-vl-plus",
] as const;

const QWEN_COMPAT_BASE_URL = "https://dashscope.aliyuncs.com/compatible-mode/v1";

function parseAssistantText(payload: any): string {
  const choices = Array.isArray(payload?.choices) ? payload.choices : [];
  const content = choices[0]?.message?.content;

  if (typeof content === "string") {
    return content.trim();
  }

  if (Array.isArray(content)) {
    const chunks = content
      .map((part: any) => (typeof part?.text === "string" ? part.text.trim() : ""))
      .filter(Boolean);
    return chunks.join("\n").trim();
  }

  return "";
}

async function getQwenVisionProviderApiKey() {
  const providers = await getEnabledAiProviders();
  const hasApiKey = (item: (typeof providers)[number]) =>
    typeof item.apiKey === "string" && Boolean(item.apiKey.trim());
  const provider =
    providers.find((item) => item.type === "alibaba" && hasApiKey(item)) ||
    providers.find((item) => item.type === "alibaba-image" && hasApiKey(item));

  if (!provider?.apiKey) {
    throw new Error("未配置可用的阿里云多模态 Provider API Key");
  }

  return provider.apiKey.trim();
}

export async function callQwen3Vl(params: {
  imageUrl?: string;
  imageUrls?: string[];
  model?: string;
  prompt: string;
  temperature?: number;
  maxTokens?: number;
  timeoutMs?: number;
  retries?: number;
}) {
  const apiKey = await getQwenVisionProviderApiKey();
  const endpoint = `${QWEN_COMPAT_BASE_URL}/chat/completions`;
  const timeoutMs = Math.max(10000, Math.min(params.timeoutMs ?? 45000, 120000));
  const retries = Math.max(0, Math.min(params.retries ?? 1, 2));
  const imageUrls = [
    ...(Array.isArray(params.imageUrls) ? params.imageUrls : []),
    params.imageUrl || "",
  ]
    .map((item) => String(item || "").trim())
    .filter(Boolean);

  if (imageUrls.length === 0) {
    throw new Error("Qwen3-VL 调用缺少图片输入");
  }

  const preferredModel = String(params.model || "").trim();
  const models = Array.from(
    new Set([
      ...(/^qwen.*(?:vl|vision)/i.test(preferredModel) ? [preferredModel] : []),
      ...QWEN3_VL_MODELS,
    ])
  );

  let lastError: Error | null = null;

  for (const model of models) {
    for (let attempt = 0; attempt <= retries; attempt += 1) {
      try {
        const controller = new AbortController();
        const timeout = setTimeout(() => controller.abort(), timeoutMs);

        try {
          const response = await fetch(endpoint, {
            method: "POST",
            signal: controller.signal,
            headers: {
              "Content-Type": "application/json",
              Authorization: `Bearer ${apiKey}`,
            },
            body: JSON.stringify({
              model,
              temperature: params.temperature ?? 0.1,
              max_tokens: params.maxTokens ?? 300,
              messages: [
                {
                  role: "user",
                  content: [
                    { type: "text", text: params.prompt },
                    ...imageUrls.map((url) => ({
                      type: "image_url",
                      image_url: {
                        url,
                      },
                    })),
                  ],
                },
              ],
            }),
          });

          const payload = await response.json().catch(() => ({}));
          if (!response.ok) {
            const message =
              payload?.error?.message || payload?.message || `HTTP ${response.status}`;
            throw new Error(message);
          }

          const text = parseAssistantText(payload);
          if (!text) {
            throw new Error("Qwen3-VL 未返回文本内容");
          }

          return { text, model };
        } finally {
          clearTimeout(timeout);
        }
      } catch (error) {
        lastError = error instanceof Error ? error : new Error(String(error));
        const errorText = String(lastError.message || "").toLowerCase();
        const isModelNotFound =
          /model.*not.*found|invalid model|not support/i.test(errorText);
        const isRetriable =
          /abort|timeout|timed out|ecconnreset|econnreset|429|rate limit|503/.test(errorText);

        if (isModelNotFound) {
          break;
        }

        if (isRetriable && attempt < retries) {
          await new Promise((resolve) => setTimeout(resolve, 400 * (attempt + 1)));
          continue;
        }

        break;
      }
    }
  }

  throw lastError || new Error("Qwen3-VL 调用失败");
}