UHD_Media_Tool/apps/web/app/components/settings/acquire.tsx

import { ArrowRight } from "lucide-react";
import { useState } from "react";
import { Link } from "react-router";
import { useEnrichStats } from "~/api/identity.ts";
import { useSettings } from "~/api/settings.ts";
import { useSites } from "~/api/sites.ts";
import { DataTable } from "~/components/console/data-table.tsx";
import { consolePath } from "~/components/console/nav.ts";
import { ENRICH_SOURCE_KIND } from "~/components/library/labels.ts";
import { SearchDefaultDialog } from "~/components/sites/search-default-dialog.tsx";
import { Button } from "~/components/ui/button.tsx";
import { num } from "./resource-section.tsx";
import { SettingsBlock } from "./settings-block.tsx";
import { type CardSpec, n, SettingsCard, SettingsGate } from "./settings-card.tsx";

// 获取分区:每日下载预算(budget)、搜索站点(search,编辑沿用站点页的对话框)、资源识别(identity)。
// 订阅、规则、站点本身在各自页面;这里只放它们共用的缺省值。

const GiB = 1024 ** 3;

export const BUDGET_CARD: CardSpec = {
  anchor: "budget",
  title: "每日预算",
  description: "任一项到上限后不再开始新的下载；前两项按自然日计。",
  fields: [
    { name: "budget.maxNewTitlesPerDay", label: "每日新作品数", kind: "number" },
    { name: "budget.maxDownloadGiBPerDay", label: "每日下载量（GiB）", kind: "number" },
    { name: "budget.maxConcurrentDownloads", label: "最大并发下载", kind: "number", hint: "1–100" },
  ],
  fromSettings: (s) => ({
    "budget.maxNewTitlesPerDay": n(s.budget.maxNewTitlesPerDay),
    "budget.maxDownloadGiBPerDay": n(Math.round((s.budget.maxDownloadBytesPerDay / GiB) * 100) / 100),
    "budget.maxConcurrentDownloads": n(s.budget.maxConcurrentDownloads),
  }),
  toPatch: (v) => {
    const gib = num(v["budget.maxDownloadGiBPerDay"]);
    return {
      budget: {
        maxNewTitlesPerDay: num(v["budget.maxNewTitlesPerDay"]),
        maxDownloadBytesPerDay: gib === undefined ? undefined : Math.round(gib * GiB),
        maxConcurrentDownloads: num(v["budget.maxConcurrentDownloads"]),
      },
    };
  },
};

const SEARCH_MODE_LABEL = { default: "默认站点", all: "全部站点", custom: "指定站点" } as const;

/** 搜索站点(settings.search):显示当前范围,修改用站点页同一个对话框 */
function SearchSitesBlock() {
  const settings = useSettings();
  const sites = useSites();
  const [open, setOpen] = useState(false);
  const cfg = settings.data?.search;
  const mode = cfg?.defaultSiteMode ?? "default";
  return (
    <SettingsBlock
      section="acquire"
      anchor="search"
      title="搜索站点"
      description="搜资源时默认查哪些站点，API 不指定站点时也按这个范围。每次搜索前仍可改选。"
      actions={
        <Button size="sm" variant="outline" disabled={!sites.data} onClick={() => setOpen(true)}>
          修改
        </Button>
      }
    >
      <div className="flex flex-wrap items-center justify-between gap-3">
        <p className="text-sm text-foreground" id="f-search.defaultSiteMode">
          {SEARCH_MODE_LABEL[mode]}
          {mode === "custom" ? (
            <span className="ml-2 text-caption text-muted-foreground">{cfg?.siteIds.length ?? 0} 个站点</span>
          ) : null}
        </p>
        <Button asChild size="sm" variant="ghost">
          <Link to={consolePath("sites")}>
            站点页
            <ArrowRight />
          </Link>
        </Button>
      </div>
      {sites.data ? <SearchDefaultDialog open={open} onOpenChange={setOpen} sites={sites.data.items} /> : null}
    </SettingsBlock>
  );
}

// 身份识别(DEC-1 惰性识别,对照 MP:RSS / 列表资源只在匹配订阅或用户搜索时识别)
export const IDENTITY_CARD: CardSpec = {
  anchor: "identity",
  title: "资源识别",
  description: "默认只在资源命中订阅、作品别名或新作品规则，或你手动识别和搜索时查 TMDB；其余放进未识别，不进待处理。",
  fields: [
    {
      name: "identity.eagerTmdbLookup",
      label: "每条新资源都查 TMDB",
      kind: "switch",
      hint: "每条新资源都查一次 TMDB。RSS 量大时每天可能上千次请求，站点独有的短剧和综艺多半查不到",
      wide: true,
    },
    {
      name: "identity.detailIdsPerSiteHour",
      label: "识别时每站每小时最多抓详情页",
      kind: "number",
      hint: "标题认不准时，抓站点详情页取 IMDb、TMDB 或豆瓣 ID 再识别；只在搜索、订阅相关资源、补全和手动重新识别时抓，RSS 收进来的资源不抓。0 = 不抓",
    },
    {
      name: "identity.autoEnrich",
      label: "TMDB 查不到时自动补全元数据",
      kind: "switch",
      hint: "资源命中订阅或新作品规则、TMDB 却查不到时，搜索网页出一份补全建议（每条资源一次，不建作品）。每次最多读 3 个网页，调用一次模型服务",
      wide: true,
    },
    {
      name: "identity.enrichDailyBudget",
      label: "元数据补全每日上限",
      kind: "number",
      hint: "手动和自动合计；7 天内同一标题直接用上次结果，不计次。0 = 停用",
    },
    {
      name: "identity.enrichDomains",
      label: "补全可读取的网站",
      kind: "textarea",
      hint: "每行一个域名，含子域名；只放平台和百科类网站。豆瓣只用搜索摘要，不读页面；robots.txt 不允许的页面不读。已登记站点的域名一律不读",
      wide: true,
    },
    {
      name: "identity.enrichSnippetAllSites",
      label: "也用其它网站的搜索摘要",
      kind: "switch",
      hint: "短剧多半只有聚合站收录。打开后，白名单以外网站的搜索摘要也用来抽年份、集数、主演、简介（只看搜索引擎给的摘要，不打开这些网页）",
      wide: true,
    },
  ],
  fromSettings: (s) => ({
    "identity.eagerTmdbLookup": s.identity.eagerTmdbLookup,
    "identity.autoEnrich": s.identity.autoEnrich,
    "identity.enrichDailyBudget": n(s.identity.enrichDailyBudget),
    "identity.enrichDomains": s.identity.enrichDomains.join("\n"),
    "identity.enrichSnippetAllSites": s.identity.enrichSnippetAllSites ?? true,
    "identity.detailIdsPerSiteHour": n(s.identity.detailIdsPerSiteHour ?? 10),
  }),
  toPatch: (v) => ({
    identity: {
      eagerTmdbLookup: v["identity.eagerTmdbLookup"] === true,
      autoEnrich: v["identity.autoEnrich"] === true,
      enrichDailyBudget: Number(v["identity.enrichDailyBudget"]),
      enrichDomains: String(v["identity.enrichDomains"] ?? "")
        .split(/[\s,，]+/)
        .map((d) => d.trim().toLowerCase())
        .filter(Boolean),
      enrichSnippetAllSites: v["identity.enrichSnippetAllSites"] === true,
      detailIdsPerSiteHour: Number(v["identity.detailIdsPerSiteHour"]),
    },
  }),
  extra: () => <EnrichStatsTable />,
};

type StatRow = { kind: string; withResults: number; hits: number; rate: number | null };

/** 最近 30 天补全命中率(按来源类别,META-2) */
function EnrichStatsTable() {
  const q = useEnrichStats(30);
  const d = q.data;
  return (
    <div className="mb-3 space-y-2">
      <p className="text-meta text-muted-foreground tabular-nums">
        {d
          ? `最近 ${d.days} 天元数据补全 ${d.total} 次，出了建议 ${d.suggested} 次${d.tracked < d.total ? `；有来源明细的 ${d.tracked} 次` : ""}`
          : "读取补全命中率…"}
      </p>
      {d && d.tracked > 0 ? (
        <DataTable<StatRow>
          rows={d.sources}
          rowKey={(r) => r.kind}
          empty={{ title: "还没有补全记录" }}
          columns={[
            { key: "kind", header: "来源", cell: (r) => ENRICH_SOURCE_KIND[r.kind] ?? r.kind },
            {
              key: "withResults",
              header: "有结果",
              align: "right",
              cell: (r) => <span className="tabular-nums">{r.withResults}</span>,
            },
            {
              key: "hits",
              header: "命中",
              align: "right",
              cell: (r) => <span className="tabular-nums">{r.hits}</span>,
            },
            {
              key: "rate",
              header: "命中率",
              align: "right",
              cell: (r) => (
                <span className="tabular-nums">{r.rate === null ? "—" : `${Math.round(r.rate * 100)}%`}</span>
              ),
            },
          ]}
        />
      ) : null}
    </div>
  );
}

export function AcquireSection() {
  return (
    <div className="grid grid-cols-1 gap-8">
      <SettingsGate>{(s) => <SettingsCard section="acquire" spec={BUDGET_CARD} settings={s} />}</SettingsGate>
      <SearchSitesBlock />
      <SettingsGate>{(s) => <SettingsCard section="acquire" spec={IDENTITY_CARD} settings={s} />}</SettingsGate>
    </div>
  );
}