{
  "schema_version": 1,
  "catalog_sha256": "5efeba987fed1d6a23d42135e47666a2b93d08f38307ee3454aedfcaa0bf55f2",
  "source_repository": "https://github.com/reacher-z/Awesome-CUA",
  "source_catalog": "/downloads/catalog.jsonl",
  "publication_build_info": "/data/build-info.json",
  "count_policy": "Each count retains its own unit and scope. Artifact populations overlap and must not be added together.",
  "fingerprint_policy": "SHA-256 and byte size refer to the exact uncompressed file bytes. The manifest does not hash itself.",
  "artifacts": [
    {
      "id": "catalog-jsonl",
      "path": "/downloads/catalog.jsonl",
      "group": "catalog",
      "title": {
        "en": "Full catalog",
        "zh": "完整目录"
      },
      "description": {
        "en": "Every active source record, including metadata-only discoveries. One JSON object per line.",
        "zh": "全部当前资源及完整来源记录，包含仅核实书目的发现项。每行一个 JSON 对象。"
      },
      "format": "JSONL",
      "media_type": "application/x-ndjson",
      "count": {
        "value": 3251,
        "unit": "resources"
      },
      "scope": "active",
      "bytes": 8465323,
      "sha256": "5efeba987fed1d6a23d42135e47666a2b93d08f38307ee3454aedfcaa0bf55f2"
    },
    {
      "id": "catalog-json",
      "path": "/data/records.json",
      "group": "catalog",
      "title": {
        "en": "Full catalog as a JSON array",
        "zh": "完整目录 JSON 数组"
      },
      "description": {
        "en": "The same complete records in one JSON array for clients that expect a single JSON document.",
        "zh": "同一份完整记录，以单个 JSON 数组提供，适合直接读取 JSON 的客户端。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 3251,
        "unit": "resources"
      },
      "scope": "active",
      "bytes": 8134928,
      "sha256": "d8c1cdaabbfc93c3a7dd1b56186a5106be2a2f0550e01ac4dbe08cb84817d8fb"
    },
    {
      "id": "catalog-sqlite",
      "path": "/downloads/catalog.sqlite",
      "group": "catalog",
      "title": {
        "en": "Queryable database",
        "zh": "可查询数据库"
      },
      "description": {
        "en": "Resource rows, full-text search, topic/platform facets and reviewed relationships.",
        "zh": "资源记录、全文搜索、主题与平台分面，以及已审阅的关联关系。"
      },
      "format": "SQLite",
      "media_type": "application/vnd.sqlite3",
      "count": {
        "value": 3251,
        "unit": "resources"
      },
      "scope": "active",
      "bytes": 16441344,
      "sha256": "a49409afab725ea73c6ce67244f1bf63ab8980f91b8d453af60537b7885c21ea"
    },
    {
      "id": "rag-reviewed",
      "path": "/downloads/rag-documents.jsonl",
      "group": "rag",
      "title": {
        "en": "Reviewed retrieval documents",
        "zh": "已审阅检索文档"
      },
      "description": {
        "en": "Reviewed core and adjacent resources, with text, citations, evidence and license metadata.",
        "zh": "已审阅的核心与相关资源，保留文本、引用、证据和许可元数据。"
      },
      "format": "JSONL",
      "media_type": "application/x-ndjson",
      "count": {
        "value": 1671,
        "unit": "documents"
      },
      "scope": "reviewed",
      "bytes": 7192084,
      "sha256": "80f7a48eda5a48c7b360f50d3bad5e25727773a93b0b8254b3008b019e6a6b95"
    },
    {
      "id": "rag-all",
      "path": "/downloads/rag-all-records.jsonl",
      "group": "rag",
      "title": {
        "en": "All retrieval documents",
        "zh": "全部检索文档"
      },
      "description": {
        "en": "Complete active corpus; preserve the metadata-only review states when retrieving.",
        "zh": "完整当前语料；检索时保留其中仅核实书目的待审状态。"
      },
      "format": "JSONL",
      "media_type": "application/x-ndjson",
      "count": {
        "value": 3251,
        "unit": "documents"
      },
      "scope": "active",
      "bytes": 12502912,
      "sha256": "7ac3c4f77fdd61d6de7815c5c63ddf05e5228d320381c0276296a851710c5702"
    },
    {
      "id": "rag-manifest",
      "path": "/downloads/rag-manifest.json",
      "group": "rag",
      "title": {
        "en": "RAG export manifest",
        "zh": "RAG 导出清单"
      },
      "description": {
        "en": "Document counts, review policy and file fingerprints.",
        "zh": "文档数量、审阅规则和文件指纹。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 2,
        "unit": "exports"
      },
      "scope": "provenance",
      "bytes": 958,
      "sha256": "2edb41a22c8a6416c8fc208af6c2b627744c22a78fb431ebdf4edb7bb0bdfb4a"
    },
    {
      "id": "citations-bib",
      "path": "/downloads/citations.bib",
      "group": "citations",
      "title": {
        "en": "Reviewed paper citations",
        "zh": "已审阅论文引用"
      },
      "description": {
        "en": "Reviewed papers and surveys. Retained arXiv versions and partial dates remain explicit.",
        "zh": "已审阅论文与综述，保留记录的 arXiv 版本与部分日期。"
      },
      "format": "BibTeX",
      "media_type": "application/x-bibtex",
      "count": {
        "value": 1200,
        "unit": "citations"
      },
      "scope": "reviewed",
      "bytes": 905185,
      "sha256": "e1c6aea9134ce5eec9976e6aeb5d64a0d6a3404d944012bea5452cac99831a84"
    },
    {
      "id": "citations-csl-json",
      "path": "/downloads/citations.csl.json",
      "group": "citations",
      "title": {
        "en": "Reviewed paper citations",
        "zh": "已审阅论文引用"
      },
      "description": {
        "en": "Reviewed papers and surveys. Retained arXiv versions and partial dates remain explicit.",
        "zh": "已审阅论文与综述，保留记录的 arXiv 版本与部分日期。"
      },
      "format": "CSL JSON",
      "media_type": "application/vnd.citationstyles.csl+json",
      "count": {
        "value": 1200,
        "unit": "citations"
      },
      "scope": "reviewed",
      "bytes": 2260660,
      "sha256": "5dbd93d13371d8c54f071f3fcae378b10d3f98c69fafd266d9e4a794cb1a607d"
    },
    {
      "id": "citations-all-bib",
      "path": "/downloads/citations-all.bib",
      "group": "citations",
      "title": {
        "en": "All bibliography citations",
        "zh": "全部书目引用"
      },
      "description": {
        "en": "Includes metadata-only papers and surveys with their review state retained.",
        "zh": "包含仅核实书目的论文与综述，并保留审阅状态。"
      },
      "format": "BibTeX",
      "media_type": "application/x-bibtex",
      "count": {
        "value": 2780,
        "unit": "citations"
      },
      "scope": "active",
      "bytes": 2004676,
      "sha256": "17dd398c1e9af54b32a205aedbfa9c14b9d20c43da43a92eb0a32e578b799e88"
    },
    {
      "id": "citations-all-csl-json",
      "path": "/downloads/citations-all.csl.json",
      "group": "citations",
      "title": {
        "en": "All bibliography citations",
        "zh": "全部书目引用"
      },
      "description": {
        "en": "Includes metadata-only papers and surveys with their review state retained.",
        "zh": "包含仅核实书目的论文与综述，并保留审阅状态。"
      },
      "format": "CSL JSON",
      "media_type": "application/vnd.citationstyles.csl+json",
      "count": {
        "value": 2780,
        "unit": "citations"
      },
      "scope": "active",
      "bytes": 5126307,
      "sha256": "fbdbc0b13a721325dd34ed929991ca0a29652de13b58ae9943639c9ea43c23f6"
    },
    {
      "id": "publication-citations-bib",
      "path": "/downloads/publication-citations.bib",
      "group": "citations",
      "title": {
        "en": "Publication citations",
        "zh": "正式出版版引用"
      },
      "description": {
        "en": "The recorded publication edition uses its own authors, title, venue and date precision.",
        "zh": "使用所记录出版版的作者、题名、会议／期刊及日期精度。"
      },
      "format": "BibTeX",
      "media_type": "application/x-bibtex",
      "count": {
        "value": 75,
        "unit": "citations"
      },
      "scope": "reviewed-publication",
      "bytes": 168977,
      "sha256": "cfd4f879c29dd48a05884b95548f5c2cd0477ef10f8198b32a0437a90a8e7287"
    },
    {
      "id": "publication-citations-csl-json",
      "path": "/downloads/publication-citations.csl.json",
      "group": "citations",
      "title": {
        "en": "Publication citations",
        "zh": "正式出版版引用"
      },
      "description": {
        "en": "The recorded publication edition uses its own authors, title, venue and date precision.",
        "zh": "使用所记录出版版的作者、题名、会议／期刊及日期精度。"
      },
      "format": "CSL JSON",
      "media_type": "application/vnd.citationstyles.csl+json",
      "count": {
        "value": 75,
        "unit": "citations"
      },
      "scope": "reviewed-publication",
      "bytes": 262080,
      "sha256": "b22890413656f85868bea03e4b60d8c87b008fc325435693f782224e43ee5be3"
    },
    {
      "id": "citations-manifest",
      "path": "/downloads/citations-manifest.json",
      "group": "citations",
      "title": {
        "en": "Citation export manifest",
        "zh": "引用导出清单"
      },
      "description": {
        "en": "Edition, date and identity policies with exact counts and fingerprints.",
        "zh": "引用版次、日期和身份规则，以及精确数量与指纹。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 6,
        "unit": "exports"
      },
      "scope": "provenance",
      "bytes": 3593,
      "sha256": "1dfd9e6c3c6ca4129a05e894d277d8f63c089d6e433d84e39ab8341e7d741fff"
    },
    {
      "id": "topics",
      "path": "/downloads/topics.json",
      "group": "research",
      "title": {
        "en": "Topic subsets",
        "zh": "主题子集"
      },
      "description": {
        "en": "Exact member IDs and JSONL/citation downloads for every reviewed topic. Memberships overlap.",
        "zh": "每个已审阅主题的精确成员 ID 与 JSONL、引用下载。主题成员可以交叉。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 35,
        "unit": "topics"
      },
      "scope": "reviewed",
      "bytes": 185522,
      "sha256": "889a586aa2b7eaca62477e71aa828c29d762ab0cd6c878a47f5d226038cc434a"
    },
    {
      "id": "benchmarks-profiles",
      "path": "/downloads/benchmarks-profiles.json",
      "group": "research",
      "title": {
        "en": "Benchmark protocols",
        "zh": "基准协议"
      },
      "description": {
        "en": "Source-reviewed interfaces, formats, counting units and scope. Missing fields remain unknown.",
        "zh": "已核读来源的接口、格式、计数单位和范围；未记录字段保持未知。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 61,
        "unit": "profiles"
      },
      "scope": "reviewed",
      "bytes": 237808,
      "sha256": "b0d909acd367f19ea7e7d878881900bae733fc9f883939b7ffe6ae50136a3bbe"
    },
    {
      "id": "datasets-profiles",
      "path": "/downloads/datasets-profiles.json",
      "group": "research",
      "title": {
        "en": "Dataset profiles",
        "zh": "数据集规格"
      },
      "description": {
        "en": "Source-reviewed interfaces, formats, counting units and scope. Missing fields remain unknown.",
        "zh": "已核读来源的接口、格式、计数单位和范围；未记录字段保持未知。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 66,
        "unit": "profiles"
      },
      "scope": "reviewed",
      "bytes": 192576,
      "sha256": "fd6895d2be3850f24eb6b9c1def1573cf01ef787dbb9ad42ef2fc2193c39ea7d"
    },
    {
      "id": "relationships",
      "path": "/downloads/relationships.json",
      "group": "research",
      "title": {
        "en": "Reviewed relationships",
        "zh": "已审阅关系"
      },
      "description": {
        "en": "Explicit source-backed relationships; incidental mentions and shared topics are separate.",
        "zh": "有明确来源的关联关系；偶然提及和相同主题分别处理。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 413,
        "unit": "relations"
      },
      "scope": "reviewed",
      "bytes": 129876,
      "sha256": "a9c14e583fc7d0e99886bbf783f02640a55de0f9a90cc50ee7d242bdc80e4c13"
    },
    {
      "id": "coverage",
      "path": "/downloads/coverage.json",
      "group": "provenance",
      "title": {
        "en": "Coverage ledger",
        "zh": "覆盖台账"
      },
      "description": {
        "en": "Observed source populations, review states and unresolved gaps.",
        "zh": "已观察到的来源范围、审阅状态和待解决缺口。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 1,
        "unit": "reports"
      },
      "scope": "provenance",
      "bytes": 216090,
      "sha256": "45bf4eb62e07338464f500ebec6b6dcd43ecdc90a2ab39626a6e6b32f93da21b"
    },
    {
      "id": "sources",
      "path": "/downloads/sources.json",
      "group": "provenance",
      "title": {
        "en": "Discovery-source registry",
        "zh": "发现来源登记"
      },
      "description": {
        "en": "Community collections and search sources used to discover candidates.",
        "zh": "用于发现候选条目的社区合集和来源文档。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 54,
        "unit": "sources"
      },
      "scope": "provenance",
      "bytes": 29835,
      "sha256": "d11483d5831d09a9221e118f3033abc7f86a3b88536bebd88e6a51fb526a6a77"
    },
    {
      "id": "scope-decisions",
      "path": "/downloads/exclusions.json",
      "group": "provenance",
      "title": {
        "en": "Scope decisions",
        "zh": "范围审阅决定"
      },
      "description": {
        "en": "Exclusion reasons, original evidence and any decisions awaiting revalidation.",
        "zh": "排除理由、原始证据，以及需要重新验证的决定。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 138,
        "unit": "decisions"
      },
      "scope": "provenance",
      "bytes": 184215,
      "sha256": "b2a244fae1c62621cdee42e15ad5225520b5d69fa90761c5136671038cf5c6eb"
    },
    {
      "id": "arxiv-archive",
      "path": "/downloads/arxiv-archive.jsonl",
      "group": "provenance",
      "title": {
        "en": "Raw arXiv metadata archive",
        "zh": "原始 arXiv 元数据档案"
      },
      "description": {
        "en": "Retained ingestion snapshots, including excluded records. This is not the active catalog.",
        "zh": "保留的采集快照，包含已排除记录；其范围不同于当前有效目录。"
      },
      "format": "JSONL",
      "media_type": "application/x-ndjson",
      "count": {
        "value": 2905,
        "unit": "papers"
      },
      "scope": "archive",
      "bytes": 5926334,
      "sha256": "cf5c56b8a50a99e0f0d352dcddc2e30bd4d309dea15e279a899cbad4bc5549e9"
    },
    {
      "id": "publications",
      "path": "/downloads/publications.jsonl",
      "group": "provenance",
      "title": {
        "en": "Publication observations",
        "zh": "出版来源记录"
      },
      "description": {
        "en": "Publication facts attached to canonical paper identities; not a second paper count.",
        "zh": "附加到规范论文身份的出版事实，不作为另一篇论文重复计数。"
      },
      "format": "JSONL",
      "media_type": "application/x-ndjson",
      "count": {
        "value": 77,
        "unit": "editions"
      },
      "scope": "recorded-publication",
      "bytes": 172425,
      "sha256": "b689ea8cab450366d87a4d34d07089e417d6ce2563da21c159f1bf755a033cb6"
    },
    {
      "id": "repository-aliases",
      "path": "/downloads/repository-aliases.json",
      "group": "provenance",
      "title": {
        "en": "Former repository names",
        "zh": "仓库旧地址映射"
      },
      "description": {
        "en": "Former and canonical GitHub or Hugging Face names verified against the same repository ID. They do not add resources.",
        "zh": "通过相同的 GitHub 或 Hugging Face 仓库 ID 核实旧名与规范名称。它们不增加资源记录。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 20,
        "unit": "aliases"
      },
      "scope": "provenance",
      "bytes": 24061,
      "sha256": "f830cd3a152433350805a322d75e5e7beb6ce1e799760c99fe34ab44d8da4c14"
    },
    {
      "id": "github-candidates",
      "path": "/downloads/github-candidates.jsonl",
      "group": "provenance",
      "title": {
        "en": "GitHub candidate repositories",
        "zh": "GitHub 候选仓库"
      },
      "description": {
        "en": "Public API metadata, counted by numeric repository ID. CUA relevance, availability and licenses await review; these observations are separate from and may overlap the active catalog.",
        "zh": "公开 API 元数据，按数字仓库 ID 去重。CUA 相关性、可用性和许可仍待审阅；这些观察记录独立于当前目录且可能与其重叠。"
      },
      "format": "JSONL",
      "media_type": "application/x-ndjson",
      "count": {
        "value": 9310,
        "unit": "repositories"
      },
      "scope": "candidate-observations",
      "bytes": 26396265,
      "sha256": "103c282ddc210c699955dcacf67f49262c4a12a6f0622168cb448b933ae3f4d1"
    },
    {
      "id": "github-candidates-manifest",
      "path": "/downloads/github-candidates-manifest.json",
      "group": "provenance",
      "title": {
        "en": "Candidate observation manifest",
        "zh": "候选观察清单"
      },
      "description": {
        "en": "Independent observation and input fingerprints, query limits, primary receipts, and a separate active-catalog comparison context.",
        "zh": "独立的观察与输入指纹、查询边界、一手收据，以及单独标明的当前目录对照信息。"
      },
      "format": "JSON",
      "media_type": "application/json",
      "count": {
        "value": 1,
        "unit": "reports"
      },
      "scope": "candidate-observations",
      "bytes": 87206,
      "sha256": "50a21bcaa1f6c2280b2d05957b498dd644c8e15bdb145f5f784f8342f028052b"
    },
    {
      "id": "resource-schema-json",
      "path": "/schemas/resource.schema.json",
      "group": "contracts",
      "title": {
        "en": "Resource schema",
        "zh": "资源记录规范"
      },
      "description": {
        "en": "Machine-readable field and format contract.",
        "zh": "可供程序读取的字段与格式约束。"
      },
      "format": "JSON Schema",
      "media_type": "application/schema+json",
      "count": {
        "value": 1,
        "unit": "schemas"
      },
      "scope": "contract",
      "bytes": 22555,
      "sha256": "7bcdd255310cdfbdb29b88db8d465f7d67a2b331b3bc0dd1355502e57ff2862e"
    },
    {
      "id": "scope-decision-schema-json",
      "path": "/schemas/scope-decision.schema.json",
      "group": "contracts",
      "title": {
        "en": "Scope-decision schema",
        "zh": "范围决定规范"
      },
      "description": {
        "en": "Machine-readable field and format contract.",
        "zh": "可供程序读取的字段与格式约束。"
      },
      "format": "JSON Schema",
      "media_type": "application/schema+json",
      "count": {
        "value": 1,
        "unit": "schemas"
      },
      "scope": "contract",
      "bytes": 1705,
      "sha256": "43a82e80518dc844ab491e5029de997d70d75905b2a9042061ef92e4a5eeda09"
    },
    {
      "id": "publication-observation-schema-json",
      "path": "/schemas/publication-observation.schema.json",
      "group": "contracts",
      "title": {
        "en": "Publication-observation schema",
        "zh": "出版来源规范"
      },
      "description": {
        "en": "Machine-readable field and format contract.",
        "zh": "可供程序读取的字段与格式约束。"
      },
      "format": "JSON Schema",
      "media_type": "application/schema+json",
      "count": {
        "value": 1,
        "unit": "schemas"
      },
      "scope": "contract",
      "bytes": 1058,
      "sha256": "20463706986e4fec5f97ce5a5081c3aab6fe9236d28bdad1f3813c7d6c39ca65"
    },
    {
      "id": "download-manifest-schema-json",
      "path": "/schemas/download-manifest.schema.json",
      "group": "contracts",
      "title": {
        "en": "Download-manifest schema",
        "zh": "下载清单规范"
      },
      "description": {
        "en": "Machine-readable field and format contract.",
        "zh": "可供程序读取的字段与格式约束。"
      },
      "format": "JSON Schema",
      "media_type": "application/schema+json",
      "count": {
        "value": 1,
        "unit": "schemas"
      },
      "scope": "contract",
      "bytes": 2724,
      "sha256": "a5f5e04cec54b2472179a45188cb72311745b84850fa92e71ba809a25e392a10"
    },
    {
      "id": "repository-alias-schema-json",
      "path": "/schemas/repository-alias.schema.json",
      "group": "contracts",
      "title": {
        "en": "Repository-alias schema",
        "zh": "仓库地址映射规范"
      },
      "description": {
        "en": "Machine-readable field and format contract.",
        "zh": "可供程序读取的字段与格式约束。"
      },
      "format": "JSON Schema",
      "media_type": "application/schema+json",
      "count": {
        "value": 1,
        "unit": "schemas"
      },
      "scope": "contract",
      "bytes": 3880,
      "sha256": "45f91e0464bf1703a5e2952f2fc92c57bb48fb03828d302012d3f968f248d8b1"
    },
    {
      "id": "github-candidate-observation-schema-json",
      "path": "/schemas/github-candidate-observation.schema.json",
      "group": "contracts",
      "title": {
        "en": "Candidate observation schema",
        "zh": "候选观察记录规范"
      },
      "description": {
        "en": "Candidate status, public field allowlist and independent observation provenance.",
        "zh": "候选状态、公开字段白名单与独立观察溯源。"
      },
      "format": "JSON Schema",
      "media_type": "application/schema+json",
      "count": {
        "value": 1,
        "unit": "schemas"
      },
      "scope": "contract",
      "bytes": 5241,
      "sha256": "1e124948940cd43199d18d1297d1312d26b1890f140728bf87a500fda959fc6d"
    },
    {
      "id": "github-candidate-observation-manifest-schema-json",
      "path": "/schemas/github-candidate-observation-manifest.schema.json",
      "group": "contracts",
      "title": {
        "en": "Candidate manifest schema",
        "zh": "候选清单规范"
      },
      "description": {
        "en": "Candidate status, public field allowlist and independent observation provenance.",
        "zh": "候选状态、公开字段白名单与独立观察溯源。"
      },
      "format": "JSON Schema",
      "media_type": "application/schema+json",
      "count": {
        "value": 1,
        "unit": "schemas"
      },
      "scope": "contract",
      "bytes": 22421,
      "sha256": "0564338b7a6febd690a7887ad2f69379c995dfb6c0cbf1f19d4b9239879975c4"
    }
  ]
}