feat(adapters): port MoviePilot open site framework (7 spiders, 20 user parsers) + MovieClaw v0.27.0 site definitions + mp-sync tools (W1-3c)

- sites/spiders: Haidan/HDDolby/RousiPro/SunnyPT/TNode/TorrentLeech/Yema (+mTorrent wrapper), specialized SiteAdapter, indirect download resolved at download time - sites/user-parsers: SiteParserBase + all 20 MP SiteSchema parsers (lxml XPath via xpath+xmldom), refreshSiteUserData with schema fallback - sites/definitions: zod SiteDefinition schemas, 23 MovieClaw v0.27.0 (MIT) definitions via scripts/mp-sync/movieclaw-import.mjs, registry getSiteDefinition/listSiteDefinitions, NOTICE - scripts/mp-sync: ledger.json (upstream sha256), check.mjs, build-ledger.mjs; docs/porting/mp-sync-ledger.md - license-and-provenance: add copied-mit kind, re-anchor MovieClaw rows to v0.27.0 Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-09-27 18:39:27 +08:00
1 parent cf85481d95
commit d9f2af2f3c
@@ -0,0 +1,114 @@
+/**
+ * SPDX-License-Identifier: GPL-3.0-only
+ * 移植自 MoviePilot(https://github.com/jxxghp/MoviePilot,GPLv3),本文件按 GPLv3 条款提供。
+ * @mp-source app/modules/indexer/__init__.py @ e6820a8
+ *
+ * 站点用户信息解析器注册表与刷新入口。
+ * - `IndexerModule.__init__` 里 `ModuleHelper.load('app.modules.indexer.parser', filter=有 schema)` → `USER_PARSER_CLASSES`
+ *   (按模块文件名字母序,与 pkgutil 遍历顺序一致;回退时按此顺序尝试)
+ * - `IndexerModule.refresh_userdata` → `refreshSiteUserData`:按站点声明的 schema 取解析器;解析后 userid 为空且非公开站时,
+ *   依次尝试其他 schema(兼容 schema 标注错误/变种站),其他 schema 的异常只记录不抛出。
+ *
+ * 与 MP 一致:声明 schema 解析器的异常原样抛出(由调用方记录),只有回退尝试的异常被吞掉。
+ * 差异:结果不含凭据;`fetchMessages` 默认 false(MP 由全局 SITE_MESSAGE 控制)。
+ */
+import type { SiteParserBase, SiteParserClass } from "./base.ts";
+import { BitptSiteUserInfo } from "./bitpt.ts";
+import { extractDomain } from "./common.ts";
+import { DiscuzUserInfo } from "./discuz.ts";
+import { FileListSiteUserInfo } from "./file-list.ts";
+import { GazelleSiteUserInfo } from "./gazelle.ts";
+import { HDDolbySiteUserInfo } from "./hddolby.ts";
+import { IptSiteUserInfo } from "./ipt.ts";
+import { MTorrentSiteUserInfo } from "./mtorrent.ts";
+import { NexusAudiencesSiteUserInfo } from "./nexus-audiences.ts";
+import { NexusHhanclubSiteUserInfo } from "./nexus-hhanclub.ts";
+import { NexusPhpSiteUserInfo } from "./nexus-php.ts";
+import { NexusProjectSiteUserInfo } from "./nexus-project.ts";
+import { NexusRabbitSiteUserInfo } from "./nexus-rabbit.ts";
+import { RousiSiteUserInfo } from "./rousi.ts";
+import { SmallHorseSiteUserInfo } from "./small-horse.ts";
+import { SunnyPTSiteUserInfo } from "./sunnypt.ts";
+import { TNodeSiteUserInfo } from "./tnode.ts";
+import { TorrentLeechSiteUserInfo } from "./torrent-leech.ts";
+import type { SiteSchemaValue, SiteUserInfo, SiteUserParserOptions } from "./types.ts";
+import { Unit3dSiteUserInfo } from "./unit3d.ts";
+import { YemaSiteUserInfo } from "./yema.ts";
+import { ZhixingSiteUserInfo } from "./zhixing.ts";
+
+/** 全部解析器(模块字母序:bitpt, discuz, file_list, gazelle, hddolby, ipt, mtorrent, nexus_*, rousi, …) */
+export const USER_PARSER_CLASSES: readonly SiteParserClass[] = [
+  BitptSiteUserInfo,
+  DiscuzUserInfo,
+  FileListSiteUserInfo,
+  GazelleSiteUserInfo,
+  HDDolbySiteUserInfo,
+  IptSiteUserInfo,
+  MTorrentSiteUserInfo,
+  NexusAudiencesSiteUserInfo,
+  NexusHhanclubSiteUserInfo,
+  NexusPhpSiteUserInfo,
+  NexusProjectSiteUserInfo,
+  NexusRabbitSiteUserInfo,
+  RousiSiteUserInfo,
+  SmallHorseSiteUserInfo,
+  SunnyPTSiteUserInfo,
+  TNodeSiteUserInfo,
+  TorrentLeechSiteUserInfo,
+  Unit3dSiteUserInfo,
+  YemaSiteUserInfo,
+  ZhixingSiteUserInfo,
+] as unknown as SiteParserClass[];
+
+export function getUserParserClass(schema: string | null | undefined): SiteParserClass | undefined {
+  if (!schema) return undefined;
+  return USER_PARSER_CLASSES.find((c) => c.schema === schema);
+}
+
+/** refresh_userdata 的站点入参(MP site dict 子集) */
+export interface SiteUserDataRequest extends Omit<SiteUserParserOptions, "siteName"> {
+  name: string;
+  /** 站点声明的解析 schema(SiteSchema 值) */
+  schema?: string | null;
+  /** 公开站:不报"未找到解析器",也不做 schema 回退 */
+  public?: boolean;
+}
+
+/** MP `refresh_userdata`:未找到解析器返回 null */
+export async function refreshSiteUserData(site: SiteUserDataRequest): Promise<SiteUserInfo | null> {
+  const make = (schema: string | null | undefined): SiteParserBase | null => {
+    const cls = getUserParserClass(schema);
+    if (!cls) return null;
+    const { name, schema: _s, public: _p, ...rest } = site;
+    return new cls({ ...rest, siteName: name });
+  };
+  let obj = make(site.schema);
+  if (!obj) return null;
+  try {
+    await obj.parse();
+    if (!obj.userid && !site.public) {
+      const tried = new Set<string>([String(site.schema)]);
+      for (const cls of USER_PARSER_CLASSES) {
+        const value = cls.schema as SiteSchemaValue | null;
+        if (!value || tried.has(value)) continue;
+        tried.add(value);
+        const alt = make(value);
+        if (!alt) continue;
+        try {
+          await alt.parse();
+        } catch {
+          continue;
+        }
+        if (alt.userid) {
+          obj = alt;
+          break;
+        }
+      }
+    }
+    const info = obj.toUserInfo();
+    info.domain = extractDomain(site.url) || null;
+    return info;
+  } finally {
+    obj.clear();
+  }
+}