feat(hub): switch capability provider from MinerU to Aliyun Doc Mind (ADR-0027)

Replace the MinerU client with Alibaba Cloud Document Mind (docmind) as the
backing service for pdf_to_md_bundle. Aliyun docmind covers both PDF→MD
(with LaTeX formula enhancement) and audio/video→text in one provider,
unlike MinerU which only does documents.

Changes:
- Install @alicloud/docmind-api20220711 + @alicloud/credentials + tea-util
- Delete mineruClient.ts; add docmindClient.ts with AliyunDocmindClient
  using the official SDK (SubmitDocParserJobAdvance → poll → GetDocParserResult)
- CapabilitySecretPayload: baseUrl+apiToken → accessKeyId+accessKeySecret+endpoint
- pdfToMdBundle adapter: provider aliyun_docmind, DocmindClientError
- Tests updated: 7/7 green, mock client matches new interface

Pricing (aliyun docmind, 2025-07):
- PDF 增强链路 (含公式 LaTeX): 0.04元/页 ≈ $0.0056/页
- 视频: 0.002元/秒
- 音频: 0.00035元/秒
Cost is derived from page count × unit price (COST_PER_PAGE_USD) until the
service reports an actual cost field; costSource=provider_reported.

The aliyun CLI is NOT used at runtime — Hub calls the SDK directly. The CLI
remains available for operators to manage AccessKeys and test connectivity.
This commit is contained in:
2026-07-18 16:41:14 +08:00
parent b673dd1fe9
commit 7bc9e4f449
8 changed files with 593 additions and 113 deletions
+3
View File
@@ -7,6 +7,9 @@
"node": ">=24"
},
"dependencies": {
"@alicloud/credentials": "^2.4.5",
"@alicloud/docmind-api20220711": "^1.4.15",
"@alicloud/tea-util": "^1.4.11",
"@anthropic-ai/claude-agent-sdk": "^0.3.202",
"@fastify/cookie": "^11.0.2",
"@larksuiteoapi/node-sdk": "^1.70.0",