[FEA] Introducing scorecard for repo basic information auto collector

[edwinjhlee]

Collect information from github

1. star
2. last-commit
3. last-release
4. release number
5. contributor number
6. pr number
7. issue number
8. lang and line count
9. download size, disk usage （ docker linux ）
10. memory consumsion
11. cpu requirement ( high, low, ... )
12. process number and information
13. use network ?
14. config , data file -- cleanup
15. 支持系统

[Zhengqbbb]

卡片的最终类型：InstallCardMeta

export interface InstallCardGithub {
    stars?: number
    lastCommit?: string
    lastRelease?: string
    releases?: number
    contributors?: number
    pullRequests?: number
    issues?: number
}

export interface InstallCardRuntime {
    downloadSize?: string
    diskUsage?: string
    /** MB */
    memory?: { idle?: number, typical?: number, peak?: number }
    cpu?: string
    architectures?: string[]
    processes?: number
    processInfo?: string
    network?: boolean
}

export interface InstallCardFiles {
    config?: string[]
    data?: string[]
    cache?: string[]
    cleanupRequired?: boolean
    safeToRemove?: boolean
}

/** 0 - 100 */
export interface InstallCardRadar {
    performance?: number
    easeOfUse?: number
    features?: number
    portability?: number
    reliability?: number
}

export interface InstallCardMeta {
    version?: string
    tagline?: string
    /** 0 - 5 */
    rating?: number
    tags?: string[]
    github?: InstallCardGithub
    /** language name -> lines */
    languages?: Record<string, number>
    totalLines?: number
    runtime?: InstallCardRuntime
    files?: InstallCardFiles
    platforms?: string[]
    demo?: { cmd?: string, output?: string }
    radar?: InstallCardRadar
}

e.g: jq

{
  "version": "1.7.1",
  "tagline": "Lightweight & Flexible JSON Processor",
  "rating": 5,
  "tags": [
    "Fast",
    "Portable",
    "Zero Dependency",
    "JSON"
  ],
  "github": {
    "stars": 30000,
    "lastCommit": "2026-08-20",
    "lastRelease": "2025-11-01",
    "releases": 32,
    "contributors": 180,
    "pullRequests": 1200,
    "issues": 800
  },
  "languages": {
    "C": 42000,
    "Shell": 3500,
    "Makefile": 1200
  },
  "totalLines": 46700,
  "runtime": {
    "downloadSize": "2.1 MB",
    "diskUsage": "6.8 MB",
    "memory": {
      "idle": 3,
      "typical": 8,
      "peak": 20
    },
    "cpu": "Low",
    "architectures": [
      "amd64",
      "arm64",
      "armv7"
    ],
    "processes": 1,
    "processInfo": "jq (single process)",
    "network": false
  },
  "files": {
    "config": [],
    "data": [],
    "cache": [],
    "cleanupRequired": false
  },
  "platforms": [
    "Linux",
    "macOS",
    "Windows",
    "FreeBSD",
    "OpenBSD",
    "NetBSD"
  ],
  "demo": {
    "cmd": "jq '.foo.bar' input.json",
    "output": "\"value\""
  },
  "radar": {
    "performance": 92,
    "easeOfUse": 78,
    "features": 88,
    "portability": 96,
    "reliability": 90
  }
}

[edwinjhlee]

我应该确定测试系统采用 linux 。 而不是 mac ，或者 windows ， 因为 linux 是 ai 的操作系统。

[Zhengqbbb]

1. 字段总结

卡片字段	来源（原样）
version	stat about.latestVersion
tagline	stat about.description，否则 yml desc
repo	yml 里 GitHub URL 的 owner/repo
github.*	stat popularity / timeline（数字字段仍按数字读）
languages	stat language 前 3 名
totalLines	stat totalLine
tags	footprint runtime.type + yml lang + license（数组就展开，标量当一项）
platforms	源里的 platforms 字符串数组；没有就不写
runtime.diskUsage	footprint disk.size.install.estimate
runtime.memory.idle/peak	footprint memory.*.estimate（解析成 MB 数字，meter 要用）
runtime.cpu	footprint cpu.peak.desc，否则 peak.rating 转成字符串
runtime.processes / processInfo	footprint cpu.processes / workload
runtime.network	footprint network.internet
runtime.architectures	源里的 architectures 字符串数组
files.config/data/cache	footprint disk.paths.*
files.cleanupRequired	上面 paths 任一非空 → true

左上角栏目是「介绍」，

2. 卡片有、两份数据都没有

rating · radar · downloadSize · memory.typical · demo · files.safeToRemove
platforms / architectures：源里目前没有对应数组，所以现在是空的。

stat 有但卡片没有槽：watcher fork created recent.*
footprint 有但卡片没有槽：network.local · filesystem.read/write

3. 外部数据问题（例）

1. stat YAML 非法：379 个 latest.card.yml 标量没加引号，只能先补引号再解析。例：Bark FFmpeg
2. totalLine / language 是字节不是行数：jq totalLine: 668169 实际是字节
3. 同名仓库撞车：stat 目录用上游 repo 名。stat/cli 是 MiniMax 的 cli，不是 GitHub CLI；install gh 对不上
4. 文件名撞车导致错配：install rebound（N-body，无 GitHub）按文件名对上 stat/rebound（shobrook/rebound）；rebound-cli 用 URL 也对上同一份，两张卡同一套 stars

4. 名称不一致

install yml	stat 目录	结果
7-Zip	7zip	去符号后对上
rebound	rebound（另一款软件）	错配
gh	没有 gh；cli 是 MiniMax	未匹配
hermeswebui	hermes-web-ui 与 hermes-webui 并存	去符号碰撞，回退未用

stat 有、install 无对应 yml（16）：Agent-Reach apm ccstatusline ctx fox orca 等。