memind/news-engine-client.mjs

const EMPTY_RANKING = Object.freeze({
  candidateCount: 0,
  eventCount: 0,
  events: [],
  selected: [],
});

export function emptyNewsEngineRanking() {
  return {
    candidateCount: 0,
    eventCount: 0,
    events: [],
    selected: [],
  };
}

export function resolveNewsEngineEndpoint(env = process.env) {
  return String(env.MEMIND_NEWS_ENGINE_URL ?? '').trim().replace(/\/+$/, '');
}

export function isNewsEngineMediaUrl(value, env = process.env) {
  const raw = String(value ?? '').trim();
  if (!raw) return false;
  if (raw.startsWith('/media/')) return true;
  const base = resolveNewsEngineEndpoint(env);
  if (!base) return false;
  try {
    const parsed = new URL(raw);
    const origin = new URL(`${base}/`).origin;
    return parsed.origin === origin && parsed.pathname.startsWith('/media/');
  } catch {
    return false;
  }
}

export function applyRankingImagesToGroups(groups = [], ranking = null, {
  normalizeUrl = (value) => String(value ?? '').trim(),
} = {}) {
  const imageByUrl = new Map();
  for (const event of [...(ranking?.selected ?? []), ...(ranking?.events ?? [])]) {
    const url = normalizeUrl(event?.primaryArticle?.url);
    const image = String(event?.primaryArticle?.image ?? '').trim();
    if (url && image) imageByUrl.set(url, image);
  }
  return (Array.isArray(groups) ? groups : []).map((group) => ({
    ...group,
    results: (Array.isArray(group?.results) ? group.results : []).map((item) => {
      const url = normalizeUrl(item?.url);
      const image = imageByUrl.get(url);
      if (!image) return item;
      return { ...item, image, imageSource: 'news-engine' };
    }),
  }));
}

function envFlag(value, fallback = false) {
  const raw = String(value ?? '').trim().toLowerCase();
  if (!raw) return fallback;
  return ['1', 'true', 'yes', 'on'].includes(raw);
}

export function shouldUseNewsEngineLatestCollection(env = process.env) {
  if (!resolveNewsEngineEndpoint(env)) return false;
  return envFlag(env.MEMIND_NEWS_ENGINE_USE_LATEST_COLLECTION, true);
}

export async function fetchLatestCollectionFromEngine({
  env = process.env,
  fetchImpl = fetch,
  logger = console,
} = {}) {
  const endpoint = resolveNewsEngineEndpoint(env);
  if (!endpoint) return null;
  const timeoutMs = Math.max(
    500,
    Number(env.MEMIND_NEWS_ENGINE_TIMEOUT_MS ?? 5000) || 5000,
  );
  const controller = new AbortController();
  const timer = setTimeout(() => controller.abort(), timeoutMs);
  timer.unref?.();
  try {
    const token = String(env.MEMIND_NEWS_ENGINE_API_TOKEN ?? '').trim();
    const response = await fetchImpl(`${endpoint}/v1/collections/latest`, {
      headers: {
        accept: 'application/json',
        ...(token ? { authorization: `Bearer ${token}` } : {}),
      },
      signal: controller.signal,
    });
    if (response.status === 404) return null;
    if (!response.ok) throw new Error(`HTTP ${response.status}`);
    const collection = await response.json();
    if (!collection || typeof collection !== 'object') return null;
    return collection;
  } catch (error) {
    logger.warn?.(
      '[NewsEngine] latest collection unavailable:',
      error instanceof Error ? error.message : error,
    );
    return null;
  } finally {
    clearTimeout(timer);
  }
}

export function resolveNewsEngineMediaFetchUrl(imageUrl, env = process.env) {
  const raw = String(imageUrl ?? '').trim();
  if (!raw) return '';
  if (raw.startsWith('/media/')) {
    const endpoint = resolveNewsEngineEndpoint(env);
    return endpoint ? `${endpoint}${raw}` : '';
  }
  return isNewsEngineMediaUrl(raw, env) ? raw : '';
}

export async function downloadNewsEngineMediaBytes(imageUrl, {
  env = process.env,
  fetchImpl = fetch,
  timeoutMs = 8000,
} = {}) {
  const fetchUrl = resolveNewsEngineMediaFetchUrl(imageUrl, env);
  if (!fetchUrl) return null;
  try {
    const token = String(env.MEMIND_NEWS_ENGINE_API_TOKEN ?? '').trim();
    const response = await fetchImpl(fetchUrl, {
      redirect: 'follow',
      signal: AbortSignal.timeout(timeoutMs),
      headers: {
        Accept: 'image/avif,image/webp,image/apng,image/*,*/*;q=0.8',
        ...(token ? { authorization: `Bearer ${token}` } : {}),
      },
    });
    if (!response.ok) return null;
    const buffer = Buffer.from(await response.arrayBuffer());
    if (buffer.length > 2_500_000 || buffer.length < 12) return null;
    return buffer;
  } catch {
    return null;
  }
}

export async function rankNewsWithEngine(groups, {
  env = process.env,
  now = Date.now(),
  fetchImpl = fetch,
  logger = console,
} = {}) {
  const endpoint = resolveNewsEngineEndpoint(env);
  if (!endpoint) return emptyNewsEngineRanking();

  const timeoutMs = Math.max(
    500,
    Number(env.MEMIND_NEWS_ENGINE_TIMEOUT_MS ?? 5000) || 5000,
  );
  const controller = new AbortController();
  const timer = setTimeout(() => controller.abort(), timeoutMs);
  timer.unref?.();
  try {
    const token = String(env.MEMIND_NEWS_ENGINE_API_TOKEN ?? '').trim();
    const response = await fetchImpl(`${endpoint}/v1/rank`, {
      method: 'POST',
      headers: {
        'content-type': 'application/json',
        ...(token ? { authorization: `Bearer ${token}` } : {}),
      },
      body: JSON.stringify({
        groups: Array.isArray(groups) ? groups : [],
        options: {
          now,
          limit: Number(env.MEMIND_NEWS_ENGINE_DAILY_LIMIT ?? 15) || 15,
          minimumScore: Number(env.MEMIND_NEWS_ENGINE_MINIMUM_SCORE ?? 0) || 0,
        },
      }),
      signal: controller.signal,
    });
    if (!response.ok) throw new Error(`HTTP ${response.status}`);
    const ranking = await response.json();
    if (!ranking || !Array.isArray(ranking.events) || !Array.isArray(ranking.selected)) {
      throw new Error('invalid ranking response');
    }
    return ranking;
  } catch (error) {
    logger.warn?.(
      '[NewsEngine] ranking unavailable:',
      error instanceof Error ? error.message : error,
    );
    return { ...EMPTY_RANKING, events: [], selected: [] };
  } finally {
    clearTimeout(timer);
  }
}

export function formatRankedNewsCandidates(ranking) {
  const selected = Array.isArray(ranking?.selected) ? ranking.selected : [];
  if (!selected.length) return '';
  const lines = [
    '【News Engine 编辑候选】以下事件已完成聚类、规则评分和栏目配额选择。生成早报时优先采用；must_include=true 的事件必须进入正文。',
  ];
  for (const event of selected) {
    const source = event.primaryArticle?.sourceDomain || 'unknown';
    lines.push(
      `${event.rank}. [${event.category}] ${event.canonicalTitle}`,
      `   score=${event.finalScore} sources=${event.sourceCount} must_include=${event.mustInclude} primary=${source}`,
    );
    if (event.primaryArticle?.url) lines.push(`   ${event.primaryArticle.url}`);
    if (event.primaryArticle?.image) lines.push(`   图：${event.primaryArticle.image}`);
  }
  return lines.join('\n');
}