chore(dc): add data collection infrastructure (DC-01/02/03)

Co-authored-by: qinfendebingshuo <207279273+qinfendebingshuo@users.noreply.github.com>
This commit is contained in:
copilot-swe-agent[bot] 2026-03-17 10:17:19 +00:00
parent 5fc33489af
commit c9d44ccb9a
6 changed files with 679 additions and 1 deletions

View File

@ -14,7 +14,8 @@
"🔍", "📊", "📡", "🧠", "📢", "🚨", "📰", "🔧", "🦅"
],
"signature_exempt_paths": [
"syslog-inbox/"
"syslog-inbox/",
"data/dc-reports/"
],
"system_protected_paths": [
".github/",

View File

@ -62,6 +62,16 @@ jobs:
- name: 设置自检日期
run: echo "CHECK_DATE=$(TZ='Asia/Shanghai' date '+%Y-%m-%d')" >> $GITHUB_ENV
- name: DC-01 Notion Usage Snapshot
env:
NOTION_TOKEN: ${{ secrets.NOTION_API_TOKEN }}
run: node scripts/dc-notion-usage.js
- name: DC-02 Workflow Performance Snapshot
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
run: node scripts/dc-workflow-perf.js
- name: Commit self-check result
uses: peter-evans/create-pull-request@v7
with:
@ -76,4 +86,5 @@ jobs:
delete-branch: true
add-paths: |
.github/persona-brain/memory.json
data/dc-reports/
labels: auto-sync

0
data/dc-reports/.gitkeep Normal file
View File

View File

@ -0,0 +1,147 @@
/**
* scripts/dc-agent-behavior.js
* DC-03 · Agent 行为记录器
*
* 采集时机每次铸渊处理 Notion 工单回写后附带写入
* 存储位置data/dc-reports/agent-behavior-YYYY-MM.json按月累积
*
* 用法
* 1. 作为模块引入
* const { appendAgentBehaviorLog } = require('./dc-agent-behavior');
* appendAgentBehaviorLog({ type: '规则同步', auto_resolved: true, processing_sec: 18 });
*
* 2. 命令行调用
* node scripts/dc-agent-behavior.js --type "规则同步" --auto --sec 18
* node scripts/dc-agent-behavior.js --type "工单回执" --human --sec 45 --llm-calls 3 --tokens 1200
*/
'use strict';
const fs = require('fs');
const path = require('path');
const ROOT = path.resolve(__dirname, '..');
const DC_DIR = path.join(ROOT, 'data/dc-reports');
const BEIJING_OFFSET_MS = 8 * 3600 * 1000;
// ━━━ 获取当月标识 ━━━
function getCurrentMonth() {
const now = new Date();
return new Date(now.getTime() + BEIJING_OFFSET_MS).toISOString().slice(0, 7);
}
// ━━━ 加载或初始化月度文件 ━━━
function loadMonthlyReport(month) {
const filePath = path.join(DC_DIR, `agent-behavior-${month}.json`);
try {
return JSON.parse(fs.readFileSync(filePath, 'utf8'));
} catch {
return {
month: month,
tickets: [],
llm_calls_total: 0,
avg_tokens_per_call: 0,
human_intervention_rate: 0
};
}
}
// ━━━ 重新计算统计指标 ━━━
function recalcStats(report) {
const total = report.tickets.length;
if (total === 0) return;
const humanCount = report.tickets.filter(t => t.human_intervention).length;
report.human_intervention_rate = parseFloat((humanCount / total).toFixed(2));
// llm_calls_total 和 avg_tokens_per_call 在 append 时累加
}
// ━━━ 核心:追加行为记录 ━━━
function appendAgentBehaviorLog(ticketData) {
fs.mkdirSync(DC_DIR, { recursive: true });
const month = ticketData.month || getCurrentMonth();
const report = loadMonthlyReport(month);
const record = {
type: ticketData.type || 'unknown',
auto_resolved: ticketData.auto_resolved !== false,
human_intervention: ticketData.human_intervention === true,
processing_sec: ticketData.processing_sec || 0,
timestamp: new Date().toISOString()
};
report.tickets.push(record);
// 累加 LLM 调用统计
if (ticketData.llm_calls) {
const prevTotal = report.llm_calls_total;
report.llm_calls_total += ticketData.llm_calls;
if (ticketData.tokens_per_call) {
// 加权平均
const prevTokens = report.avg_tokens_per_call * prevTotal;
const newTokens = ticketData.tokens_per_call * ticketData.llm_calls;
report.avg_tokens_per_call = Math.round(
(prevTokens + newTokens) / report.llm_calls_total
);
}
}
recalcStats(report);
const filePath = path.join(DC_DIR, `agent-behavior-${month}.json`);
fs.writeFileSync(filePath, JSON.stringify(report, null, 2));
console.log(`🤖 DC-03 · 行为记录已追加: ${record.type} (${month})`);
return report;
}
// ━━━ CLI 模式 ━━━
function parseCli() {
const args = process.argv.slice(2);
if (args.length === 0) return null;
const data = {};
for (let i = 0; i < args.length; i++) {
switch (args[i]) {
case '--type':
data.type = args[++i]; break;
case '--auto':
data.auto_resolved = true; break;
case '--human':
data.human_intervention = true;
data.auto_resolved = false; break;
case '--sec':
data.processing_sec = parseInt(args[++i], 10) || 0; break;
case '--llm-calls':
data.llm_calls = parseInt(args[++i], 10) || 0; break;
case '--tokens':
data.tokens_per_call = parseInt(args[++i], 10) || 0; break;
case '--month':
data.month = args[++i]; break;
}
}
return data.type ? data : null;
}
// ━━━ 入口 ━━━
if (require.main === module) {
const cliData = parseCli();
if (cliData) {
appendAgentBehaviorLog(cliData);
} else {
console.log('用法: node dc-agent-behavior.js --type "类型" [--auto|--human] [--sec N] [--llm-calls N] [--tokens N]');
process.exit(1);
}
}
module.exports = { appendAgentBehaviorLog };

250
scripts/dc-notion-usage.js Normal file
View File

@ -0,0 +1,250 @@
/**
* scripts/dc-notion-usage.js
* DC-01 · Notion 读写分析器
*
* 采集时机每日自检结束时附带写入
* 存储位置data/dc-reports/notion-usage-YYYY-MM-DD.json
*
* 采集方式
* 读取天眼报告中的 Notion 相关诊断数据
* 统计当日 API 调用次数并写入标准格式
* Notion API 可用时额外查询数据库元信息
*/
'use strict';
const fs = require('fs');
const path = require('path');
const https = require('https');
const ROOT = path.resolve(__dirname, '..');
const DC_DIR = path.join(ROOT, 'data/dc-reports');
const SKYEYE_DIR = path.join(ROOT, 'data/skyeye-reports');
const BEIJING_OFFSET_MS = 8 * 3600 * 1000;
const now = new Date();
const today = new Date(now.getTime() + BEIJING_OFFSET_MS).toISOString().split('T')[0];
// ━━━ 工具函数 ━━━
function loadJson(filePath) {
try {
return JSON.parse(fs.readFileSync(filePath, 'utf8'));
} catch {
return null;
}
}
function notionGet(apiPath, token) {
return new Promise((resolve, reject) => {
const options = {
hostname: 'api.notion.com',
path: apiPath,
method: 'GET',
headers: {
'Authorization': 'Bearer ' + token,
'Notion-Version': '2022-06-28',
'Content-Type': 'application/json',
'User-Agent': 'dc-notion-usage'
},
timeout: 15000
};
const req = https.request(options, (res) => {
let data = '';
res.on('data', chunk => data += chunk);
res.on('end', () => {
try {
resolve({ statusCode: res.statusCode, body: JSON.parse(data) });
} catch {
resolve({ statusCode: res.statusCode, body: data });
}
});
});
req.on('error', reject);
req.on('timeout', () => { req.destroy(); reject(new Error('timeout')); });
req.end();
});
}
// ━━━ 从天眼报告中提取 Notion 相关数据 ━━━
function extractFromSkyeye() {
const info = {
skyeye_available: false,
bridges_status: null,
notion_connected: false,
api_hint_calls: 0
};
// 查找最近的天眼报告
try {
const files = fs.readdirSync(SKYEYE_DIR)
.filter(f => f.startsWith('skyeye-') && f.endsWith('.json'))
.sort()
.reverse();
if (files.length === 0) return info;
const report = loadJson(path.join(SKYEYE_DIR, files[0]));
if (!report) return info;
info.skyeye_available = true;
// 提取 bridge 扫描中的 Notion 连接信息
if (report.bridges) {
const bridges = typeof report.bridges === 'string'
? loadJson(path.join(SKYEYE_DIR, 'bridge-health.json'))
: report.bridges;
if (bridges && bridges.notion_api) {
info.notion_connected = bridges.notion_api.connected === true;
info.bridges_status = bridges.notion_api.status || null;
}
}
} catch {
// 天眼报告不可用,跳过
}
return info;
}
// ━━━ 查询 Notion 数据库元信息 ━━━
async function queryNotionDatabases(token) {
const databases = [];
try {
const resp = await notionGet('/v1/search', token);
// search 需要 POST用简单 GET 替代:查询已知数据库
// 由于 search 是 POST此处跳过使用环境变量中已知的数据库 ID
} catch {
// Notion API 不可达
}
// 尝试查询环境变量中的已知数据库
const knownDbs = [];
if (process.env.NOTION_DB_SYSLOG) {
knownDbs.push({ id: process.env.NOTION_DB_SYSLOG, name: 'SYSLOG收件箱' });
}
if (process.env.NOTION_TICKET_DB_ID) {
knownDbs.push({ id: process.env.NOTION_TICKET_DB_ID, name: '人格协作工单簿' });
}
if (process.env.BRIDGE_QUEUE_DB_ID) {
knownDbs.push({ id: process.env.BRIDGE_QUEUE_DB_ID, name: '调度队列' });
}
for (const db of knownDbs) {
try {
const resp = await notionGet('/v1/databases/' + db.id, token);
if (resp.statusCode === 200 && resp.body && resp.body.properties) {
const props = Object.keys(resp.body.properties);
databases.push({
name: resp.body.title?.[0]?.plain_text || db.name,
daily_reads: 0,
daily_writes: 0,
active_fields: props.slice(0, 20),
unused_fields: []
});
}
} catch {
// 该数据库不可访问
}
}
return databases;
}
// ━━━ 读取 writeback log 统计当日调用次数 ━━━
function countTodayApiCalls() {
let totalCalls = 0;
let peakHourMap = {};
// 检查 notion-writeback-log.json
const wbLog = loadJson(path.join(ROOT, 'data/notion-writeback-log.json'));
if (wbLog && Array.isArray(wbLog)) {
for (const entry of wbLog) {
if (entry.timestamp && entry.timestamp.startsWith(today)) {
totalCalls++;
const hour = entry.timestamp.substring(11, 13);
peakHourMap[hour] = (peakHourMap[hour] || 0) + 1;
}
}
}
// 检查 bridge-logs
const bridgeLogDir = path.join(ROOT, 'data/bridge-logs');
if (fs.existsSync(bridgeLogDir)) {
try {
const logFiles = fs.readdirSync(bridgeLogDir)
.filter(f => f.includes(today) && f.endsWith('.json'));
for (const lf of logFiles) {
const log = loadJson(path.join(bridgeLogDir, lf));
if (log && log.notion_calls) {
totalCalls += log.notion_calls;
}
}
} catch {
// bridge-logs 不可读
}
}
// 找出峰值小时
let peakHour = null;
let peakCount = 0;
for (const [hour, count] of Object.entries(peakHourMap)) {
if (count > peakCount) {
peakCount = count;
const h = parseInt(hour, 10);
peakHour = `${String(h).padStart(2, '0')}:00-${String(h + 1).padStart(2, '0')}:00`;
}
}
return { totalCalls, peakHour: peakHour || 'N/A' };
}
// ━━━ 主流程 ━━━
async function main() {
console.log(`📦 DC-01 · Notion 读写分析器 · ${today}`);
fs.mkdirSync(DC_DIR, { recursive: true });
const skyeyeInfo = extractFromSkyeye();
const { totalCalls, peakHour } = countTodayApiCalls();
const notionToken = process.env.NOTION_TOKEN;
let databases = [];
if (notionToken) {
console.log(' → Notion Token 可用,查询数据库元信息...');
databases = await queryNotionDatabases(notionToken);
console.log(` → 查询到 ${databases.length} 个数据库`);
} else {
console.log(' ⚠️ 缺少 NOTION_TOKEN跳过 Notion API 查询');
}
const report = {
date: today,
databases: databases,
total_api_calls: totalCalls,
peak_hour: peakHour,
_meta: {
generated_by: 'dc-notion-usage.js',
generated_at: now.toISOString(),
skyeye_available: skyeyeInfo.skyeye_available,
notion_connected: skyeyeInfo.notion_connected
}
};
const outputFile = path.join(DC_DIR, `notion-usage-${today}.json`);
fs.writeFileSync(outputFile, JSON.stringify(report, null, 2));
console.log(` ✅ DC-01 报告已写入: ${path.relative(ROOT, outputFile)}`);
}
main().catch(err => {
console.error('❌ DC-01 采集失败:', err.message);
process.exit(1);
});

269
scripts/dc-workflow-perf.js Normal file
View File

@ -0,0 +1,269 @@
/**
* scripts/dc-workflow-perf.js
* DC-02 · Workflow 性能采集器
*
* 采集时机每日自检结束时附带写入
* 存储位置data/dc-reports/workflow-perf-YYYY-MM-DD.json
*
* 采集方式
* 调用 GitHub API GET /repos/{owner}/{repo}/actions/runs
* 拉取当日运行记录聚合写入标准格式
*/
'use strict';
const fs = require('fs');
const path = require('path');
const https = require('https');
const ROOT = path.resolve(__dirname, '..');
const DC_DIR = path.join(ROOT, 'data/dc-reports');
const WF_DIR = path.join(ROOT, '.github/workflows');
const OWNER = 'qinfendebingshuo';
const REPO = 'guanghulab';
const BEIJING_OFFSET_MS = 8 * 3600 * 1000;
const now = new Date();
const today = new Date(now.getTime() + BEIJING_OFFSET_MS).toISOString().split('T')[0];
// ━━━ GitHub API 请求 ━━━
function githubGet(apiPath) {
const token = process.env.GITHUB_TOKEN || process.env.GH_TOKEN;
return new Promise((resolve, reject) => {
const options = {
hostname: 'api.github.com',
path: apiPath,
method: 'GET',
headers: {
'User-Agent': 'dc-workflow-perf',
'Accept': 'application/vnd.github.v3+json'
},
timeout: 30000
};
if (token) {
options.headers['Authorization'] = 'Bearer ' + token;
}
const req = https.request(options, (res) => {
let data = '';
res.on('data', chunk => data += chunk);
res.on('end', () => {
try {
resolve({ statusCode: res.statusCode, body: JSON.parse(data) });
} catch {
resolve({ statusCode: res.statusCode, body: data });
}
});
});
req.on('error', reject);
req.on('timeout', () => { req.destroy(); reject(new Error('timeout')); });
req.end();
});
}
// ━━━ 从 GitHub API 拉取当日 workflow runs ━━━
async function fetchTodayRuns() {
const runs = [];
let page = 1;
const maxPages = 5;
while (page <= maxPages) {
try {
const resp = await githubGet(
`/repos/${OWNER}/${REPO}/actions/runs?created=>=${today}T00:00:00Z&per_page=100&page=${page}`
);
if (resp.statusCode !== 200) {
console.log(` ⚠️ GitHub API 返回 ${resp.statusCode},停止分页`);
break;
}
const data = resp.body;
if (!data.workflow_runs || data.workflow_runs.length === 0) break;
runs.push(...data.workflow_runs);
if (runs.length >= data.total_count) break;
page++;
} catch (err) {
console.log(` ⚠️ GitHub API 请求失败: ${err.message}`);
break;
}
}
return runs;
}
// ━━━ 本地 workflow 文件分析 ━━━
function countLocalWorkflows() {
try {
return fs.readdirSync(WF_DIR)
.filter(f => f.endsWith('.yml') || f.endsWith('.yaml'))
.length;
} catch {
return 0;
}
}
function extractTriggerType(run) {
return run.event || 'unknown';
}
function extractDependencies(workflowName) {
// 尝试从 workflow 文件内容中提取 needs 依赖
const deps = [];
try {
const files = fs.readdirSync(WF_DIR).filter(f => f.endsWith('.yml') || f.endsWith('.yaml'));
for (const file of files) {
const content = fs.readFileSync(path.join(WF_DIR, file), 'utf8');
const nameMatch = content.match(/^name\s*:\s*["']?(.+?)["']?\s*$/m);
const name = nameMatch ? nameMatch[1].trim() : file;
if (name === workflowName || file === workflowName) {
// 提取 needs 字段
const needsMatches = content.match(/needs\s*:\s*\[?([^\]\n]+)\]?/g);
if (needsMatches) {
for (const m of needsMatches) {
const needs = m.replace(/needs\s*:\s*\[?\s*/, '').replace(/\]?\s*$/, '');
deps.push(...needs.split(',').map(s => s.trim().replace(/['"]/g, '')).filter(Boolean));
}
}
// 提取 workflow_run 依赖
if (content.includes('workflow_run:')) {
const wfRunMatch = content.match(/workflows\s*:\s*\[([^\]]+)\]/);
if (wfRunMatch) {
deps.push(...wfRunMatch[1].split(',').map(s => s.trim().replace(/['"]/g, '')).filter(Boolean));
}
}
break;
}
}
} catch {
// 文件读取失败
}
return [...new Set(deps)];
}
// ━━━ 聚合 runs → 按 workflow 分组统计 ━━━
function aggregateRuns(runs) {
const wfMap = {};
let failedToday = 0;
for (const run of runs) {
const name = run.name || run.workflow_id?.toString() || 'unknown';
if (!wfMap[name]) {
wfMap[name] = {
name: name,
trigger: extractTriggerType(run),
runs_today: 0,
durations: [],
successes: 0,
failures: 0,
dependencies: extractDependencies(name)
};
}
wfMap[name].runs_today++;
// 计算运行时长(秒)
if (run.created_at && run.updated_at) {
const start = new Date(run.created_at).getTime();
const end = new Date(run.updated_at).getTime();
const durationSec = Math.round((end - start) / 1000);
if (durationSec > 0) {
wfMap[name].durations.push(durationSec);
}
}
if (run.conclusion === 'success') {
wfMap[name].successes++;
} else if (run.conclusion === 'failure') {
wfMap[name].failures++;
failedToday++;
}
}
const workflows = Object.values(wfMap).map(wf => {
const totalRuns = wf.successes + wf.failures;
return {
name: wf.name,
trigger: wf.trigger,
runs_today: wf.runs_today,
avg_duration_sec: wf.durations.length > 0
? Math.round(wf.durations.reduce((a, b) => a + b, 0) / wf.durations.length)
: 0,
success_rate: totalRuns > 0
? parseFloat((wf.successes / totalRuns).toFixed(2))
: null,
dependencies: wf.dependencies
};
});
return { workflows, failedToday };
}
// ━━━ 统计可并行 workflow 数量 ━━━
function countParallelCapable() {
let parallelCount = 0;
try {
const files = fs.readdirSync(WF_DIR).filter(f => f.endsWith('.yml') || f.endsWith('.yaml'));
for (const file of files) {
const content = fs.readFileSync(path.join(WF_DIR, file), 'utf8');
// 没有 workflow_run 触发且没有 needs 依赖的 job 视为可并行
if (!content.includes('workflow_run:') && !/needs\s*:/.test(content)) {
parallelCount++;
}
}
} catch {
// 读取失败
}
return parallelCount;
}
// ━━━ 主流程 ━━━
async function main() {
console.log(`⚙️ DC-02 · Workflow 性能采集器 · ${today}`);
fs.mkdirSync(DC_DIR, { recursive: true });
const totalWorkflows = countLocalWorkflows();
console.log(` → 本地 workflow 文件数: ${totalWorkflows}`);
const runs = await fetchTodayRuns();
console.log(` → 当日 workflow runs: ${runs.length}`);
const { workflows, failedToday } = aggregateRuns(runs);
const parallelCapable = countParallelCapable();
const report = {
date: today,
workflows: workflows,
total_workflows: totalWorkflows,
failed_today: failedToday,
parallel_capable: parallelCapable,
_meta: {
generated_by: 'dc-workflow-perf.js',
generated_at: now.toISOString(),
runs_fetched: runs.length
}
};
const outputFile = path.join(DC_DIR, `workflow-perf-${today}.json`);
fs.writeFileSync(outputFile, JSON.stringify(report, null, 2));
console.log(` ✅ DC-02 报告已写入: ${path.relative(ROOT, outputFile)}`);
}
main().catch(err => {
console.error('❌ DC-02 采集失败:', err.message);
process.exit(1);
});