diff --git a/scripts/src/hwlab-node-help.ts b/scripts/src/hwlab-node-help.ts index 2552cf40..62117d82 100644 --- a/scripts/src/hwlab-node-help.ts +++ b/scripts/src/hwlab-node-help.ts @@ -67,15 +67,23 @@ export function hwlabNodeWebProbeHelp(): Record { "bun scripts/cli.ts hwlab nodes web-probe run --node D601 --lane v03 --wait-messages-ms 1000", "bun scripts/cli.ts hwlab nodes web-probe run --node D601 --lane v03 --url https://hwlab.pikapython.com --fresh-session --message 'ping'", "bun scripts/cli.ts hwlab nodes web-probe script --node D601 --lane v03 --script-file .state/probes/workbench.mjs", + "bun scripts/cli.ts hwlab nodes web-probe observe start --node D601 --lane v03 --target-path /workbench --sample-interval-ms 5000", + "bun scripts/cli.ts hwlab nodes web-probe observe command --node D601 --lane v03 --state-dir .state/web-observe/D601/v03/YYYY/MM/DD/ --type sendPrompt --text 'ping'", + "bun scripts/cli.ts hwlab nodes web-probe observe status --node D601 --lane v03 --state-dir .state/web-observe/D601/v03/YYYY/MM/DD/", + "bun scripts/cli.ts hwlab nodes web-probe observe analyze --node D601 --lane v03 --state-dir .state/web-observe/D601/v03/YYYY/MM/DD/", "bun scripts/cli.ts hwlab nodes web-probe script --node D601 --lane v03 <<'JS'\nexport default async ({ waitWorkbenchReady, fetchJson, fetchApiMatrix, recordStep, collectText, safeEvaluate, screenshot }) => {\n const ready = await waitWorkbenchReady();\n const workspace = await fetchJson('/v1/workbench/workspace?projectId=prj_hwpod_workbench');\n const apiMatrix = await fetchApiMatrix(['/v1/workbench/workspace?projectId=prj_hwpod_workbench', '/auth/session']);\n const workspaceText = await collectText('#workspace');\n const evaluated = await safeEvaluate(({ a, b }) => ({ sum: a + b }), { a: 1, b: 2 });\n await screenshot('workbench.png');\n recordStep('workbench-summary', { finalUrl: ready.finalUrl, workspaceOk: workspace.ok, apiMatrixOk: apiMatrix.ok });\n return { finalUrl: ready.finalUrl, workspaceOk: workspace.ok, workspaceText, evaluated };\n};\nJS", ], actions: { run: "Run the repo-owned scripts/web-live-dom-probe.mjs helper.", script: "Run caller-provided Playwright JS after CLI-managed /auth/login; the script receives authenticated browser/context/page plus gotoStable/reloadStable/gotoCurrentStable/safeReload/fetchJson/safeFetchJson/fetchApiMatrix/recordStep/collectText/safeEvaluate/waitWorkbenchReady/screenshotOnError/summarizeWorkspace/summarizeConversation helpers and must not handle secrets itself.", + observe: "Start, inspect, control, stop, collect, and analyze a pure-client long-running Workbench observer on the target host. The observer has one Playwright page authority, receives commands through stateDir/commands files, writes JSONL artifacts, and does not expose any inbound service API.", }, notes: [ "Prefer --script-file for reusable probes; stdin heredocs remain supported for one-off probes.", "Issue-ready evidence is available under issueEvidence and summary.issueEvidence; full script report is persisted under probe.reportPath with a SHA-256 fingerprint.", + "observe sampling is passive by default: it records DOM summaries and natural page request/response/requestfailed events with observerInitiated=false; it does not actively fetch Workbench APIs, reload, switch sessions, route/intercept, or call repair helpers.", + "observe command actions are explicit user/control actions and are appended to control.jsonl; use --type sendPrompt/goto/screenshot/mark/stop and keep prompt text out of issue comments by citing textHash/textBytes.", + "observe analyze is offline-only: it reads artifact JSONL and writes analysis/report.md plus analysis/report.json without accessing Workbench APIs or driving the browser.", "Use recordStep(name, data) or fetchApiMatrix(paths) to keep structured partial evidence when a later step fails.", "Use reloadStable(), gotoCurrentStable(), or safeReload() for bounded retries around page reload/current-URL navigation jitter such as ERR_NETWORK_CHANGED.", "Playwright page.evaluate accepts one serializable argument; use page.evaluate(({ a, b }) => ..., { a, b }) or safeEvaluate(fn, { a, b }).", diff --git a/scripts/src/hwlab-node-impl.ts b/scripts/src/hwlab-node-impl.ts index 97ca7c45..1d6f04a7 100644 --- a/scripts/src/hwlab-node-impl.ts +++ b/scripts/src/hwlab-node-impl.ts @@ -10,6 +10,7 @@ import { classifySshTcpPoolFailure } from "./ssh"; import { HWLAB_NODE_CONTROL_PLANE_CONFIG_PATH, hwlabNodeControlPlaneInfraHelp, runHwlabNodeControlPlaneInfra } from "./hwlab-node-control-plane"; import { hwlabRuntimeLaneConfigPath, hwlabRuntimeLaneSpec, hwlabRuntimeLaneSpecForNode, isHwlabRuntimeLane, type HwlabRuntimeLane, type HwlabRuntimeLaneSpec, type HwlabRuntimeObservabilityRecordingRuleSpec, type HwlabRuntimeObservabilitySpec, type HwlabRuntimeObservabilityWarningAlertSpec, type HwlabRuntimePublicExposureSpec } from "./hwlab-node-lanes"; import { nodeWebProbeScriptRunnerSource } from "./hwlab-node-web-probe-runner-source"; +import { nodeWebObserveAnalyzerSource, nodeWebObserveRunnerSource } from "./hwlab-node-web-observe-runner-source"; import { hwlabNodeHelp, hwlabNodeObservabilityHelp, hwlabNodeWebProbeHelp } from "./hwlab-node-help"; import { compactWebProbeResult, compactWebProbeScriptResult } from "./hwlab-node-web-probe-summary"; import { nodeObservabilityRecordingRuleExpression, nodeObservabilityRecordingRuleSummaries, nodeObservabilityWarningAlertExpression, nodeObservabilityWarningAlertSummaries } from "./hwlab-node-observability-promql"; @@ -56,7 +57,35 @@ interface NodeWebProbeScriptOptions { }; } -type NodeWebProbeOptions = NodeWebProbeRunOptions | NodeWebProbeScriptOptions; +type NodeWebProbeObserveAction = "start" | "status" | "command" | "stop" | "collect" | "analyze"; +type NodeWebProbeObserveCommandType = "login" | "preflight" | "goto" | "sendPrompt" | "clickSession" | "screenshot" | "mark" | "stop"; + +interface NodeWebProbeObserveOptions { + action: "observe"; + observeAction: NodeWebProbeObserveAction; + node: string; + lane: string; + url: string; + targetPath: string; + viewport: string; + sampleIntervalMs: number; + screenshotIntervalMs: number; + maxSamples: number; + commandTimeoutSeconds: number; + waitMs: number; + tailLines: number; + maxFiles: number; + stateDir: string | null; + jobId: string | null; + force: boolean; + commandType: NodeWebProbeObserveCommandType | null; + commandText: string | null; + commandPath: string | null; + commandLabel: string | null; + commandSessionId: string | null; +} + +type NodeWebProbeOptions = NodeWebProbeRunOptions | NodeWebProbeScriptOptions | NodeWebProbeObserveOptions; type NodeObservabilityAction = "plan" | "apply" | "status" | "workbench-summary" | "performance-summary"; @@ -5896,13 +5925,14 @@ function rewriteDelegatedNodeString(value: string, scoped: ReturnType const lane = options.lane; if (!isHwlabRuntimeLane(lane)) throw new Error(`web-probe only supports HWLAB runtime lanes, got ${lane}`); const spec = hwlabRuntimeLaneSpecForNode(lane, options.node); + if (options.action === "observe" && options.observeAction !== "start") return runNodeWebProbeObserve(options, spec, null, null, null); const secretSpec = runtimeSecretSpec({ node: options.node, lane }); const material = readBootstrapAdminPasswordMaterial(secretSpec); const credential = webProbeCredential(secretSpec, material); @@ -6046,7 +6159,9 @@ function runNodeWebProbe(options: NodeWebProbeOptions): Record return { ok: false, status: "blocked", - command: `hwlab nodes web-probe ${options.action} --node ${options.node} --lane ${options.lane}`, + command: options.action === "observe" + ? `hwlab nodes web-probe observe ${options.observeAction} --node ${options.node} --lane ${options.lane}` + : `hwlab nodes web-probe ${options.action} --node ${options.node} --lane ${options.lane}`, node: options.node, lane: options.lane, workspace: spec.workspace, @@ -6056,6 +6171,7 @@ function runNodeWebProbe(options: NodeWebProbeOptions): Record next: { secretStatus: `bun scripts/cli.ts hwlab nodes secret status --node ${options.node} --lane ${options.lane} --name ${secretSpec.bootstrapAdminSecret}` }, }; } + if (options.action === "observe") return runNodeWebProbeObserve(options, spec, secretSpec, material, credential); if (options.action === "script") return runNodeWebProbeScript(options, spec, secretSpec, material, credential); if (options.commandTimeoutSeconds > 55) return runNodeWebProbeAsync(options, spec, secretSpec, material, credential); const probeArgs = nodeWebProbeRunArgs(options, "run"); @@ -6279,6 +6395,311 @@ function webProbeCredential(secretSpec: RuntimeSecretSpec, material: BootstrapAd }; } +function runNodeWebProbeObserve( + options: NodeWebProbeObserveOptions, + spec: HwlabRuntimeLaneSpec, + secretSpec: RuntimeSecretSpec | null, + material: BootstrapAdminPasswordMaterial | null, + credential: Record | null, +): Record { + if (options.observeAction === "start") { + if (secretSpec === null || material === null || credential === null || material.password === null) throw new Error("web-probe observe start requires bootstrap admin credential material"); + return runNodeWebProbeObserveStart(options, spec, secretSpec, material, credential); + } + if (options.observeAction === "status") return runNodeWebProbeObserveStatus(options, spec); + if (options.observeAction === "command") return runNodeWebProbeObserveCommand(options, spec, false); + if (options.observeAction === "stop") return runNodeWebProbeObserveCommand({ ...options, commandType: "stop" }, spec, true); + if (options.observeAction === "collect") return runNodeWebProbeObserveCollect(options, spec); + return runNodeWebProbeObserveAnalyze(options, spec); +} + +function runNodeWebProbeObserveStart( + options: NodeWebProbeObserveOptions, + spec: HwlabRuntimeLaneSpec, + secretSpec: RuntimeSecretSpec, + material: BootstrapAdminPasswordMaterial, + credential: Record, +): Record { + const jobId = `webobs-${Date.now().toString(36)}-${randomBytes(3).toString("hex")}`; + const timestamp = new Date().toISOString().replace(/[-:]/gu, "").replace(/[.]\d{3}Z$/u, "Z"); + const day = timestamp.slice(0, 8); + const stateDir = `.state/web-observe/${safeWebObserveSegment(options.node)}/${safeWebObserveSegment(options.lane)}/${day.slice(0, 4)}/${day.slice(4, 6)}/${day.slice(6, 8)}/${timestamp}_${safeWebObserveTargetSegment(options.targetPath)}_${jobId}`; + const runnerB64 = Buffer.from(nodeWebObserveRunnerSource(), "utf8").toString("base64"); + const script = [ + "set -eu", + `state_dir=${shellQuote(stateDir)}`, + "mkdir -p \"$state_dir\"", + "chmod 700 \"$state_dir\"", + "runner=\"$state_dir/observer-runner.mjs\"", + `node -e "require('fs').writeFileSync(process.argv[1], Buffer.from(process.argv[2], 'base64'))" "$runner" ${shellQuote(runnerB64)}`, + "chmod 700 \"$runner\"", + [ + `HWLAB_WEB_BASE_URL=${shellQuote(options.url)}`, + `HWLAB_WEB_USER=${shellQuote(secretSpec.bootstrapAdminUsername)}`, + `HWLAB_WEB_PASS=${shellQuote(material.password)}`, + `UNIDESK_WEB_OBSERVE_STATE_DIR=${shellQuote(stateDir)}`, + `UNIDESK_WEB_OBSERVE_JOB_ID=${shellQuote(jobId)}`, + `UNIDESK_WEB_OBSERVE_TARGET_PATH=${shellQuote(options.targetPath)}`, + `UNIDESK_WEB_OBSERVE_SAMPLE_INTERVAL_MS=${shellQuote(String(options.sampleIntervalMs))}`, + `UNIDESK_WEB_OBSERVE_SCREENSHOT_INTERVAL_MS=${shellQuote(String(options.screenshotIntervalMs))}`, + `UNIDESK_WEB_OBSERVE_MAX_SAMPLES=${shellQuote(String(options.maxSamples))}`, + `UNIDESK_WEB_OBSERVE_VIEWPORT=${shellQuote(options.viewport)}`, + "nohup node \"$runner\" >\"$state_dir/stdout.log\" 2>\"$state_dir/stderr.log\" \"$state_dir/pid\"", + "sleep 1", + `node -e ${shellQuote("const fs=require('fs'); const dir=process.argv[1]; const read=(n)=>{try{return JSON.parse(fs.readFileSync(dir+'/'+n,'utf8'))}catch{return null}}; const pid=fs.existsSync(dir+'/pid')?fs.readFileSync(dir+'/pid','utf8').trim():null; console.log(JSON.stringify({ok:true,command:'web-probe-observe start',jobId:process.argv[2],stateDir:dir,pid:Number(pid)||null,manifestPath:dir+'/manifest.json',heartbeat:read('heartbeat.json'),manifest:read('manifest.json'),statusCommand:'bun scripts/cli.ts hwlab nodes web-probe observe status --node '+process.argv[3]+' --lane '+process.argv[4]+' --state-dir '+dir,stopCommand:'bun scripts/cli.ts hwlab nodes web-probe observe stop --node '+process.argv[3]+' --lane '+process.argv[4]+' --state-dir '+dir,valuesRedacted:true},null,2))")} "$state_dir" ${shellQuote(jobId)} ${shellQuote(options.node)} ${shellQuote(options.lane)}`, + ].join("\n"); + const result = runTransWorkspaceStdinScript(options.node, spec.workspace, script, options.commandTimeoutSeconds); + const started = parseJsonObject(result.stdout); + return { + ok: result.exitCode === 0 && started?.ok === true, + status: result.exitCode === 0 && started?.ok === true ? "started" : "blocked", + command: `hwlab nodes web-probe observe start --node ${options.node} --lane ${options.lane}`, + node: options.node, + lane: options.lane, + workspace: spec.workspace, + url: options.url, + targetPath: options.targetPath, + credential, + observer: started, + result: compactCommandResultRedacted(result, [material.password ?? ""]), + valuesRedacted: true, + }; +} + +function runNodeWebProbeObserveStatus(options: NodeWebProbeObserveOptions, spec: HwlabRuntimeLaneSpec): Record { + const script = [ + "set -eu", + nodeWebObserveResolveStateDirShell(options), + nodeWebObserveStatusNodeScript(options.tailLines, options.node, options.lane), + ].join("\n"); + const result = runTransWorkspaceStdinScript(options.node, spec.workspace, script, options.commandTimeoutSeconds); + const status = parseJsonObject(result.stdout); + return { + ok: result.exitCode === 0 && status?.ok !== false, + status: result.exitCode === 0 ? "observed" : "blocked", + command: `hwlab nodes web-probe observe status --node ${options.node} --lane ${options.lane}`, + node: options.node, + lane: options.lane, + workspace: spec.workspace, + observer: status, + result: compactCommandResult(result), + valuesRedacted: true, + }; +} + +function runNodeWebProbeObserveCommand(options: NodeWebProbeObserveOptions, spec: HwlabRuntimeLaneSpec, stopCommand: boolean): Record { + const type = options.commandType ?? (stopCommand ? "stop" : null); + if (type === null) throw new Error("web-probe observe command requires --type"); + const commandId = `cmd-${Date.now().toString(36)}-${randomBytes(3).toString("hex")}`; + const payload = { + id: commandId, + type, + createdAt: new Date().toISOString(), + source: "cli", + path: options.commandPath, + text: options.commandText, + label: options.commandLabel, + sessionId: options.commandSessionId, + }; + const payloadB64 = Buffer.from(JSON.stringify(payload), "utf8").toString("base64"); + const script = [ + "set -eu", + nodeWebObserveResolveStateDirShell(options), + "mkdir -p \"$state_dir/commands/pending\"", + `node -e "const fs=require('fs'),path=require('path'); const dir=process.argv[1], id=process.argv[2], payload=Buffer.from(process.argv[3], 'base64').toString('utf8'); fs.writeFileSync(path.join(dir,'commands','pending',id+'.json'), payload+'\\n', {mode:0o600});" "$state_dir" ${shellQuote(commandId)} ${shellQuote(payloadB64)}`, + nodeWebObserveWaitCommandShell(commandId, options.waitMs), + ].join("\n"); + const result = runTransWorkspaceStdinScript(options.node, spec.workspace, script, options.commandTimeoutSeconds); + const commandResult = parseJsonObject(result.stdout); + if (options.force && stopCommand && result.exitCode !== 0) { + const killResult = runTransWorkspaceStdinScript(options.node, spec.workspace, [ + "set -eu", + nodeWebObserveResolveStateDirShell(options), + "if [ -f \"$state_dir/pid\" ]; then kill \"$(cat \"$state_dir/pid\")\" >/dev/null 2>&1 || true; fi", + "printf '{\"ok\":true,\"forced\":true,\"stateDir\":\"%s\"}\\n' \"$state_dir\"", + ].join("\n"), 55); + return { + ok: killResult.exitCode === 0, + status: killResult.exitCode === 0 ? "forced-stop-requested" : "blocked", + command: `hwlab nodes web-probe observe stop --node ${options.node} --lane ${options.lane}`, + node: options.node, + lane: options.lane, + workspace: spec.workspace, + commandId, + observerCommand: commandSummaryForOutput(payload), + gracefulResult: compactCommandResult(result), + forceResult: compactCommandResult(killResult), + valuesRedacted: true, + }; + } + return { + ok: result.exitCode === 0 && commandResult?.ok !== false, + status: result.exitCode === 0 ? (options.waitMs > 0 ? "completed-or-queued" : "queued") : "blocked", + command: stopCommand + ? `hwlab nodes web-probe observe stop --node ${options.node} --lane ${options.lane}` + : `hwlab nodes web-probe observe command --node ${options.node} --lane ${options.lane}`, + node: options.node, + lane: options.lane, + workspace: spec.workspace, + commandId, + observerCommand: commandSummaryForOutput(payload), + observer: commandResult, + result: compactCommandResult(result), + valuesRedacted: true, + }; +} + +function runNodeWebProbeObserveCollect(options: NodeWebProbeObserveOptions, spec: HwlabRuntimeLaneSpec): Record { + const script = [ + "set -eu", + nodeWebObserveResolveStateDirShell(options), + nodeWebObserveCollectNodeScript(options.maxFiles), + ].join("\n"); + const result = runTransWorkspaceStdinScript(options.node, spec.workspace, script, options.commandTimeoutSeconds); + const collect = parseJsonObject(result.stdout); + return { + ok: result.exitCode === 0 && collect?.ok !== false, + status: result.exitCode === 0 ? "collected" : "blocked", + command: `hwlab nodes web-probe observe collect --node ${options.node} --lane ${options.lane}`, + node: options.node, + lane: options.lane, + workspace: spec.workspace, + collect, + result: compactCommandResult(result), + valuesRedacted: true, + }; +} + +function runNodeWebProbeObserveAnalyze(options: NodeWebProbeObserveOptions, spec: HwlabRuntimeLaneSpec): Record { + const analyzerB64 = Buffer.from(nodeWebObserveAnalyzerSource(), "utf8").toString("base64"); + const script = [ + "set -eu", + nodeWebObserveResolveStateDirShell(options), + "analyzer=\"$state_dir/observer-analyzer.mjs\"", + `node -e "require('fs').writeFileSync(process.argv[1], Buffer.from(process.argv[2], 'base64'))" "$analyzer" ${shellQuote(analyzerB64)}`, + "chmod 700 \"$analyzer\"", + "UNIDESK_WEB_OBSERVE_STATE_DIR=\"$state_dir\" node \"$analyzer\"", + ].join("\n"); + const result = runTransWorkspaceStdinScript(options.node, spec.workspace, script, options.commandTimeoutSeconds); + const analysis = parseJsonObject(result.stdout); + return { + ok: result.exitCode === 0 && analysis?.ok === true, + status: result.exitCode === 0 && analysis?.ok === true ? "analyzed" : "blocked", + command: `hwlab nodes web-probe observe analyze --node ${options.node} --lane ${options.lane}`, + node: options.node, + lane: options.lane, + workspace: spec.workspace, + analysis, + result: compactCommandResult(result), + valuesRedacted: true, + }; +} + +function nodeWebObserveResolveStateDirShell(options: Pick): string { + if (options.stateDir !== null) { + return [ + `state_dir=${shellQuote(options.stateDir)}`, + "test -d \"$state_dir\" || { printf '{\"ok\":false,\"error\":\"state-dir-missing\",\"stateDir\":\"%s\"}\\n' \"$state_dir\"; exit 2; }", + ].join("\n"); + } + if (options.jobId === null) throw new Error("web-probe observe requires --state-dir or --job-id"); + const root = `.state/web-observe/${safeWebObserveSegment(options.node)}/${safeWebObserveSegment(options.lane)}`; + return [ + `observe_root=${shellQuote(root)}`, + `job_id=${shellQuote(options.jobId)}`, + "state_dir=$(find \"$observe_root\" -type d -name \"*_${job_id}\" 2>/dev/null | sort | tail -n 1 || true)", + "test -n \"$state_dir\" || { printf '{\"ok\":false,\"error\":\"job-id-not-found\",\"jobId\":\"%s\",\"root\":\"%s\"}\\n' \"$job_id\" \"$observe_root\"; exit 2; }", + ].join("\n"); +} + +function nodeWebObserveStatusNodeScript(tailLines: number, node: string, lane: string): string { + return `node -e ${shellQuote(` +const fs=require('fs'),path=require('path'); +const dir=process.env.state_dir||process.argv[1]; +const node=process.argv[2]; +const lane=process.argv[3]; +const tailN=${tailLines}; +const readJson=(name)=>{try{return JSON.parse(fs.readFileSync(path.join(dir,name),'utf8'))}catch{return null}}; +const tailJsonl=(name)=>{try{return fs.readFileSync(path.join(dir,name),'utf8').trim().split(/\\r?\\n/).filter(Boolean).slice(-tailN).map(line=>{try{return JSON.parse(line)}catch{return {parseError:true, rawTail:line.slice(-500)}}})}catch{return []}}; +const short=(value)=>String(value||'').slice(0,160); +const compactManifest=(item)=>item?{jobId:item.jobId,status:item.status,specRef:item.specRef,baseUrl:item.baseUrl,targetPath:item.targetPath,pageAuthority:item.pageAuthority,sampling:item.sampling,safety:item.safety,startedAt:item.startedAt,completedAt:item.completedAt}:null; +const compactControl=(item)=>({ts:item.ts,seq:item.seq,phase:item.phase,type:item.type,commandId:item.commandId,source:item.source,input:item.input,afterUrl:item.afterUrl,detail:item.detail&&typeof item.detail==='object'?{reason:item.detail.reason||null,mark:item.detail.mark||null,stopping:item.detail.stopping===true,pageId:item.detail.pageId||null}:null}); +const compactSample=(item)=>({seq:item.seq,ts:item.ts,reason:item.reason,url:item.url,path:item.path,routeSessionId:item.routeSessionId||null,activeSessionId:item.activeSessionId||null,observerInitiated:item.observerInitiated===true,scroll:item.scroll||null,messageCount:Array.isArray(item.messages)?item.messages.length:0,traceRowCount:Array.isArray(item.traceRows)?item.traceRows.length:0,messages:Array.isArray(item.messages)?item.messages.slice(-3).map((row)=>({index:row.index,tag:row.tag,status:row.status||null,textHash:row.textHash||null,textPreview:short(row.textPreview),textBytes:row.textBytes||0})):[],traceRows:Array.isArray(item.traceRows)?item.traceRows.slice(-3).map((row)=>({index:row.index,tag:row.tag,status:row.status||null,textHash:row.textHash||null,textPreview:short(row.textPreview),textBytes:row.textBytes||0})):[],performanceCount:Array.isArray(item.performance)?item.performance.length:0}); +const compactNetwork=(item)=>({ts:item.ts,type:item.type,method:item.method,url:item.url,resourceType:item.resourceType,status:item.status||null,observerInitiated:item.observerInitiated===true,bodyRead:item.bodyRead===true,commandId:item.commandId||null,failure:item.failure||null}); +const pidText=fs.existsSync(path.join(dir,'pid'))?fs.readFileSync(path.join(dir,'pid'),'utf8').trim():null; +let alive=false; if(pidText){try{process.kill(Number(pidText),0); alive=true}catch{}} +console.log(JSON.stringify({ok:true,command:'web-probe-observe status',stateDir:dir,pid:pidText?Number(pidText):null,processAlive:alive,manifest:compactManifest(readJson('manifest.json')),heartbeat:readJson('heartbeat.json'),tails:{control:tailJsonl('control.jsonl').map(compactControl),samples:tailJsonl('samples.jsonl').map(compactSample),network:tailJsonl('network.jsonl').map(compactNetwork),errors:tailJsonl('errors.jsonl'),artifacts:tailJsonl('artifacts.jsonl')},next:{command:'bun scripts/cli.ts hwlab nodes web-probe observe command --node '+node+' --lane '+lane+' --state-dir '+dir+' --type mark --label checkpoint',stop:'bun scripts/cli.ts hwlab nodes web-probe observe stop --node '+node+' --lane '+lane+' --state-dir '+dir,analyze:'bun scripts/cli.ts hwlab nodes web-probe observe analyze --node '+node+' --lane '+lane+' --state-dir '+dir},valuesRedacted:true},null,2)); +`)} "$state_dir" ${shellQuote(node)} ${shellQuote(lane)}`; +} + +function nodeWebObserveCollectNodeScript(maxFiles: number): string { + return `node -e ${shellQuote(` +const fs=require('fs'),path=require('path'),crypto=require('crypto'); +const dir=process.argv[1]; const maxFiles=${maxFiles}; +const out=[]; const walk=(p)=>{for(const ent of fs.readdirSync(p,{withFileTypes:true})){const full=path.join(p,ent.name); if(ent.isDirectory()) walk(full); else out.push(full); if(out.length>=maxFiles) return;}}; +walk(dir); +const files=out.slice(0,maxFiles).map(file=>{const st=fs.statSync(file); const hash=crypto.createHash('sha256').update(fs.readFileSync(file)).digest('hex'); return {path:file,relative:path.relative(dir,file),byteCount:st.size,sha256:'sha256:'+hash}}); +const totalBytes=files.reduce((sum,item)=>sum+item.byteCount,0); +console.log(JSON.stringify({ok:true,command:'web-probe-observe collect',stateDir:dir,fileCount:files.length,totalBytes,files,valuesRedacted:true},null,2)); +`)} "$state_dir"`; +} + +function nodeWebObserveWaitCommandShell(commandId: string, waitMs: number): string { + if (waitMs <= 0) { + return [ + `printf '{"ok":true,"queued":true,"commandId":%s,"stateDir":%s}\\n' ${shellQuote(JSON.stringify(commandId))} "$(node -e "console.log(JSON.stringify(process.argv[1]))" "$state_dir")"`, + ].join("\n"); + } + const waitSeconds = Math.max(1, Math.ceil(waitMs / 1000)); + return [ + `command_id=${shellQuote(commandId)}`, + `deadline=$(( $(date +%s) + ${waitSeconds} ))`, + "while [ \"$(date +%s)\" -le \"$deadline\" ]; do", + " if [ -f \"$state_dir/commands/done/${command_id}.json\" ]; then cat \"$state_dir/commands/done/${command_id}.json\"; exit 0; fi", + " if [ -f \"$state_dir/commands/failed/${command_id}.json\" ]; then cat \"$state_dir/commands/failed/${command_id}.json\"; exit 2; fi", + " sleep 1", + "done", + "printf '{\"ok\":true,\"queued\":true,\"waitTimedOut\":true,\"commandId\":\"%s\",\"stateDir\":\"%s\"}\\n' \"$command_id\" \"$state_dir\"", + ].join("\n"); +} + +function commandSummaryForOutput(payload: Record): Record { + const text = typeof payload.text === "string" ? payload.text : null; + return { + id: payload.id, + type: payload.type, + path: payload.path ?? null, + label: payload.label ?? null, + sessionId: payload.sessionId ?? null, + textHash: text === null ? null : `sha256:${createHash("sha256").update(text).digest("hex")}`, + textBytes: text === null ? null : Buffer.byteLength(text), + valuesRedacted: true, + }; +} + +function isSafeWebObserveStateDir(value: string): boolean { + return value.length > 0 + && !value.includes("\0") + && !value.includes("..") + && /^\.state\/web-observe\/[A-Za-z0-9_.-]+\/[A-Za-z0-9_.-]+\/\d{4}\/\d{2}\/\d{2}\/[A-Za-z0-9_.TZ-]+_[A-Za-z0-9_.-]+_webobs-[A-Za-z0-9_.-]+$/u.test(value); +} + +function isSafeWebObserveJobId(value: string): boolean { + return /^webobs-[A-Za-z0-9_.-]+$/u.test(value); +} + +function safeWebObserveSegment(value: string): string { + return value.replace(/[^A-Za-z0-9_.-]+/gu, "-").replace(/^-+|-+$/gu, "") || "item"; +} + +function safeWebObserveTargetSegment(value: string): string { + const segment = value.replace(/^https?:\/\//u, "").replace(/[^A-Za-z0-9_.-]+/gu, "-").replace(/^-+|-+$/gu, ""); + return segment.slice(0, 48) || "workbench"; +} + function runNodeWebProbeScript( options: NodeWebProbeScriptOptions, spec: HwlabRuntimeLaneSpec, diff --git a/scripts/src/hwlab-node-web-observe-runner-source.ts b/scripts/src/hwlab-node-web-observe-runner-source.ts new file mode 100644 index 00000000..837c377a --- /dev/null +++ b/scripts/src/hwlab-node-web-observe-runner-source.ts @@ -0,0 +1,716 @@ +// SPEC: PJ2026-01040111 长程观测 draft-2026-06-20-p0-passive-web-probe-observer. +// Responsibility: Source strings for the pure-client HWLAB web-probe observer and offline analyzer. + +export function nodeWebObserveRunnerSource(): string { + return String.raw`#!/usr/bin/env node +import { createHash, randomBytes } from "node:crypto"; +import { appendFile, mkdir, readFile, readdir, rename, stat, unlink, writeFile } from "node:fs/promises"; +import path from "node:path"; +import { pathToFileURL } from "node:url"; + +const specRef = "PJ2026-01040111 长程观测 draft-2026-06-20-p0-passive-web-probe-observer"; +const startedAtMs = Date.now(); +const startedAt = new Date(startedAtMs).toISOString(); +const baseUrl = normalizeBaseUrl(process.env.HWLAB_WEB_BASE_URL); +const username = process.env.HWLAB_WEB_USER || "admin"; +const password = process.env.HWLAB_WEB_PASS || ""; +const stateDir = path.resolve(process.env.UNIDESK_WEB_OBSERVE_STATE_DIR || ".state/web-observe/manual"); +const jobId = safeId(process.env.UNIDESK_WEB_OBSERVE_JOB_ID || "webobs-" + Date.now().toString(36) + "-" + randomBytes(3).toString("hex")); +const targetPath = process.env.UNIDESK_WEB_OBSERVE_TARGET_PATH || "/workbench"; +const sampleIntervalMs = positiveInteger(process.env.UNIDESK_WEB_OBSERVE_SAMPLE_INTERVAL_MS, 5000); +const screenshotIntervalMs = positiveInteger(process.env.UNIDESK_WEB_OBSERVE_SCREENSHOT_INTERVAL_MS, 300000); +const maxSamples = positiveInteger(process.env.UNIDESK_WEB_OBSERVE_MAX_SAMPLES, 0); +const viewport = parseViewport(process.env.UNIDESK_WEB_OBSERVE_VIEWPORT || "1440x900"); +const pageId = "page-" + randomBytes(4).toString("hex"); +const dirs = { + commandsPending: path.join(stateDir, "commands", "pending"), + commandsProcessing: path.join(stateDir, "commands", "processing"), + commandsDone: path.join(stateDir, "commands", "done"), + commandsFailed: path.join(stateDir, "commands", "failed"), + screenshots: path.join(stateDir, "screenshots"), + analysis: path.join(stateDir, "analysis"), +}; +const files = { + manifest: path.join(stateDir, "manifest.json"), + heartbeat: path.join(stateDir, "heartbeat.json"), + control: path.join(stateDir, "control.jsonl"), + samples: path.join(stateDir, "samples.jsonl"), + network: path.join(stateDir, "network.jsonl"), + console: path.join(stateDir, "console.jsonl"), + errors: path.join(stateDir, "errors.jsonl"), + artifacts: path.join(stateDir, "artifacts.jsonl"), +}; + +let browser; +let context; +let page; +let sampleSeq = 0; +let commandSeq = 0; +let artifactSeq = 0; +let activeCommandId = null; +let stopping = false; +let terminalStatus = "starting"; +let lastScreenshotAtMs = 0; +let auth = null; + +try { + if (!password) throw new Error("missing HWLAB_WEB_PASS"); + await prepareDirs(); + await writeManifest({ status: "starting" }); + await writeHeartbeat({ status: "starting" }); + const launcher = await import(pathToFileURL(path.resolve("scripts/src/browser-launcher.mjs")).href); + const { chromium } = await launcher.importPlaywright(); + browser = await launcher.launchChromium(chromium); + context = await browser.newContext({ viewport }); + auth = await runControlCommand({ id: "startup-login", type: "login", createdAt: startedAt, source: "startup" }, async () => authenticate(context)); + page = await context.newPage(); + attachPassiveListeners(page); + await runControlCommand({ id: "startup-goto", type: "goto", path: targetPath, createdAt: new Date().toISOString(), source: "startup" }, async () => gotoTarget(targetPath)); + terminalStatus = "running"; + await writeManifest({ status: "running", auth: publicAuth(auth) }); + await writeHeartbeat({ status: "running" }); + while (!stopping) { + await drainOneCommand(); + await samplePage("interval"); + if (maxSamples > 0 && sampleSeq >= maxSamples) { + await appendJsonl(files.control, controlRecord({ id: "max-samples", type: "stop", source: "sampler" }, "completed", { reason: "max-samples", maxSamples })); + break; + } + await sleep(sampleIntervalMs); + } + terminalStatus = "completed"; + await writeHeartbeat({ status: "completed" }); + await writeManifest({ status: "completed", completedAt: new Date().toISOString() }); + process.exitCode = 0; +} catch (error) { + terminalStatus = "failed"; + await appendJsonl(files.errors, eventRecord("runner-error", { error: errorSummary(error) })).catch(() => {}); + await writeHeartbeat({ status: "failed", error: errorSummary(error) }).catch(() => {}); + await writeManifest({ status: "failed", error: errorSummary(error) }).catch(() => {}); + process.exitCode = 2; +} finally { + if (browser) await browser.close().catch(() => {}); +} + +async function prepareDirs() { + await mkdir(stateDir, { recursive: true, mode: 0o700 }); + await Promise.all(Object.values(dirs).map((dir) => mkdir(dir, { recursive: true, mode: 0o700 }))); +} + +async function writeManifest(extra = {}) { + const manifest = { + ok: extra.status !== "failed", + command: "web-probe-observe", + specRef, + jobId, + pid: process.pid, + stateDir, + baseUrl, + targetPath, + pageAuthority: { browser: "chromium", context: "single", pageId, continuityBreaksRecorded: true }, + sampling: { mode: "passive", sampleIntervalMs, screenshotIntervalMs, maxSamples, observerInitiatedDefault: false, responseBodyReadDefault: false }, + commandDirs: dirs, + artifacts: files, + safety: { pureClient: true, inboundApi: false, database: false, queueConsumer: false, k8sWorkload: false, valuesRedacted: true, secretValuesPrinted: false }, + startedAt, + ...extra, + }; + await writeFile(files.manifest, JSON.stringify(manifest, null, 2) + "\n", { mode: 0o600 }); +} + +async function writeHeartbeat(extra = {}) { + const heartbeat = { + ok: terminalStatus !== "failed", + jobId, + pid: process.pid, + stateDir, + status: terminalStatus, + pageId, + baseUrl, + currentUrl: currentPageUrl(), + sampleSeq, + commandSeq, + activeCommandId, + updatedAt: new Date().toISOString(), + uptimeMs: Date.now() - startedAtMs, + ...extra, + }; + await writeFile(files.heartbeat, JSON.stringify(heartbeat, null, 2) + "\n", { mode: 0o600 }); +} + +function attachPassiveListeners(targetPage) { + targetPage.on("request", (request) => { + void appendJsonl(files.network, eventRecord("request", { + observerInitiated: false, + commandId: activeCommandId, + method: request.method(), + url: safeUrl(request.url()), + resourceType: request.resourceType(), + frameUrl: safeFrameUrl(request.frame()), + })); + }); + targetPage.on("response", (response) => { + const request = response.request(); + void appendJsonl(files.network, eventRecord("response", { + observerInitiated: false, + commandId: activeCommandId, + method: request.method(), + url: safeUrl(response.url()), + resourceType: request.resourceType(), + status: response.status(), + statusText: response.statusText(), + fromServiceWorker: response.fromServiceWorker(), + bodyRead: false, + })); + }); + targetPage.on("requestfailed", (request) => { + void appendJsonl(files.network, eventRecord("requestfailed", { + observerInitiated: false, + commandId: activeCommandId, + method: request.method(), + url: safeUrl(request.url()), + resourceType: request.resourceType(), + failure: request.failure()?.errorText ?? null, + })); + }); + targetPage.on("console", (message) => { + void appendJsonl(files.console, eventRecord("console", { type: message.type(), text: truncate(message.text(), 1000), location: message.location() })); + }); + targetPage.on("pageerror", (error) => { + void appendJsonl(files.errors, eventRecord("pageerror", { error: errorSummary(error) })); + }); + targetPage.on("crash", () => { + void appendJsonl(files.errors, eventRecord("page-crash", { pageId })); + }); + targetPage.on("close", () => { + void appendJsonl(files.control, eventRecord("continuity-break", { pageId, reason: "page-closed" })); + }); +} + +async function drainOneCommand() { + const entries = (await readdir(dirs.commandsPending).catch(() => [])).filter((name) => name.endsWith(".json")).sort(); + const name = entries[0]; + if (!name) return; + const pending = path.join(dirs.commandsPending, name); + const processing = path.join(dirs.commandsProcessing, name); + await rename(pending, processing).catch(() => null); + const raw = await readFile(processing, "utf8"); + const command = JSON.parse(raw); + const id = safeId(command.id || name.replace(/[.]json$/u, "")); + command.id = id; + try { + const result = await processCommand(command); + const done = { ok: true, commandId: id, type: command.type, completedAt: new Date().toISOString(), result: sanitize(result) }; + await writeFile(path.join(dirs.commandsDone, id + ".json"), JSON.stringify(done, null, 2) + "\n", { mode: 0o600 }); + await appendJsonl(files.control, controlRecord(command, "completed", done.result)); + await unlink(processing).catch(() => {}); + } catch (error) { + const failed = { ok: false, commandId: id, type: command.type, failedAt: new Date().toISOString(), error: errorSummary(error) }; + await writeFile(path.join(dirs.commandsFailed, id + ".json"), JSON.stringify(failed, null, 2) + "\n", { mode: 0o600 }); + await appendJsonl(files.control, controlRecord(command, "failed", failed.error)); + await unlink(processing).catch(() => {}); + } finally { + activeCommandId = null; + await writeHeartbeat({ status: terminalStatus }); + } +} + +async function processCommand(command) { + commandSeq += 1; + activeCommandId = command.id; + await writeHeartbeat({ status: "running", activeCommandId }); + await appendJsonl(files.control, controlRecord(command, "started", commandInputSummary(command))); + switch (command.type) { + case "login": return authenticate(context); + case "preflight": return preflightSummary(); + case "goto": return gotoTarget(command.path || command.url || targetPath); + case "sendPrompt": return sendPrompt(String(command.text || "")); + case "clickSession": return clickSession(String(command.sessionId || command.value || "")); + case "screenshot": return captureScreenshot(command.reason || "manual", command.imageType || "png"); + case "mark": return { mark: truncate(command.label || command.text || "mark", 200), currentUrl: currentPageUrl(), pageId }; + case "stop": stopping = true; return { stopping: true, currentUrl: currentPageUrl(), pageId }; + default: throw new Error("unsupported observer command type: " + command.type); + } +} + +async function runControlCommand(command, fn) { + activeCommandId = command.id; + commandSeq += 1; + const beforeUrl = currentPageUrl(); + const started = Date.now(); + await appendJsonl(files.control, controlRecord(command, "started", { beforeUrl, input: commandInputSummary(command) })); + try { + const result = await fn(); + await appendJsonl(files.control, controlRecord(command, "completed", { beforeUrl, afterUrl: currentPageUrl(), durationMs: Date.now() - started, result: sanitize(result) })); + return result; + } catch (error) { + await appendJsonl(files.control, controlRecord(command, "failed", { beforeUrl, afterUrl: currentPageUrl(), durationMs: Date.now() - started, error: errorSummary(error) })); + throw error; + } finally { + activeCommandId = null; + } +} + +async function authenticate(browserContext) { + const loginUrl = new URL("/auth/login", baseUrl).toString(); + const response = await browserContext.request.post(loginUrl, { + data: { username, password }, + headers: { accept: "application/json", "content-type": "application/json" }, + timeout: 15000, + }); + const cookies = await browserContext.cookies(baseUrl); + const cookieNames = cookies.map((cookie) => cookie.name).filter((name) => /session|auth|token/iu.test(name)); + return { + ok: response.ok() && cookieNames.length > 0, + method: "api", + loginPath: new URL(loginUrl).pathname, + status: response.status(), + statusText: response.statusText(), + cookiePresent: cookieNames.length > 0, + cookieNames, + valuesRedacted: true, + }; +} + +function publicAuth(value) { + if (!value) return null; + return { ok: value.ok === true, method: value.method, status: value.status, cookiePresent: value.cookiePresent === true, cookieNames: value.cookieNames || [], valuesRedacted: true }; +} + +async function gotoTarget(rawTarget) { + const target = new URL(String(rawTarget || targetPath), baseUrl).toString(); + const beforeUrl = currentPageUrl(); + const response = await page.goto(target, { waitUntil: "domcontentloaded", timeout: 30000 }); + await page.waitForTimeout(1000).catch(() => {}); + return { beforeUrl, afterUrl: currentPageUrl(), httpStatus: response ? response.status() : null, pageId }; +} + +async function sendPrompt(text) { + if (text.trim().length === 0) throw new Error("sendPrompt requires non-empty text"); + const beforeUrl = currentPageUrl(); + const editor = page.locator('textarea, [role="textbox"], [contenteditable="true"], input[type="text"]').last(); + await editor.waitFor({ state: "visible", timeout: 15000 }); + const tag = await editor.evaluate((element) => element.tagName.toLowerCase()).catch(() => ""); + const editable = await editor.evaluate((element) => element.getAttribute("contenteditable") === "true").catch(() => false); + if (tag === "textarea" || tag === "input") await editor.fill(text); + else if (editable) { + await editor.click(); + await page.keyboard.insertText(text); + } else { + await editor.click(); + await page.keyboard.insertText(text); + } + const submit = page.locator([ + 'button[type="submit"]', + 'button:has-text("发送")', + 'button:has-text("Send")', + '[data-testid*="send" i]', + '[aria-label*="send" i]', + '[aria-label*="发送"]' + ].join(", ")).last(); + await submit.waitFor({ state: "visible", timeout: 15000 }); + await submit.click(); + await page.waitForTimeout(1500); + return { beforeUrl, afterUrl: currentPageUrl(), textHash: sha256Text(text), textBytes: Buffer.byteLength(text), pageId }; +} + +async function clickSession(sessionId) { + if (!sessionId) throw new Error("clickSession requires --session-id or --value"); + const beforeUrl = currentPageUrl(); + const escaped = cssEscape(sessionId); + const candidate = page.locator("[data-session-id=\"" + escaped + "\"], [href*=\"" + escaped + "\"], text=" + sessionId).first(); + await candidate.waitFor({ state: "visible", timeout: 15000 }); + await candidate.click(); + await page.waitForTimeout(1000); + return { beforeUrl, afterUrl: currentPageUrl(), sessionId, pageId }; +} + +async function preflightSummary() { + return { currentUrl: currentPageUrl(), title: await page.title().catch(() => null), pageId, auth: publicAuth(auth) }; +} + +async function samplePage(reason) { + if (!page || page.isClosed()) return; + sampleSeq += 1; + const dom = await page.evaluate(() => { + const trim = (value, limit = 500) => String(value || "").replace(/\s+/g, " ").trim().slice(0, limit); + const visible = (element) => { + const rect = element.getBoundingClientRect(); + const style = window.getComputedStyle(element); + return rect.width > 0 && rect.height > 0 && style.visibility !== "hidden" && style.display !== "none"; + }; + const textHashInput = (element) => trim(element.textContent || "", 800); + const summarize = (selector, limit) => Array.from(document.querySelectorAll(selector)).filter(visible).slice(-limit).map((element, index) => { + const rect = element.getBoundingClientRect(); + return { + index, + tag: element.tagName.toLowerCase(), + testId: element.getAttribute("data-testid"), + role: element.getAttribute("role"), + status: element.getAttribute("data-status") || element.getAttribute("aria-busy") || null, + text: textHashInput(element), + rect: { x: Math.round(rect.x), y: Math.round(rect.y), width: Math.round(rect.width), height: Math.round(rect.height) }, + }; + }); + const url = location.href; + const routeSessionMatch = url.match(/\/workbench\/sessions\/([^/?#]+)/u); + const activeSession = document.querySelector('[data-active="true"][data-session-id], [aria-selected="true"][data-session-id], .active[data-session-id]'); + const activeSessionId = activeSession ? activeSession.getAttribute("data-session-id") : null; + const messageSelector = '[data-testid*="message" i], [class*="message" i], article, [role="article"]'; + const traceSelector = '[data-testid*="trace" i], [class*="trace" i], [data-trace-id], [data-testid*="event" i]'; + const messages = summarize(messageSelector, 20); + const traceRows = summarize(traceSelector, 30); + const active = document.activeElement; + return { + url, + path: location.pathname, + routeSessionId: routeSessionMatch ? decodeURIComponent(routeSessionMatch[1]) : null, + activeSessionId, + title: document.title, + focus: active ? { tag: active.tagName.toLowerCase(), testId: active.getAttribute("data-testid"), role: active.getAttribute("role") } : null, + viewport: { width: window.innerWidth, height: window.innerHeight, devicePixelRatio: window.devicePixelRatio }, + scroll: { x: Math.round(window.scrollX), y: Math.round(window.scrollY), height: Math.round(document.documentElement.scrollHeight), width: Math.round(document.documentElement.scrollWidth) }, + messages, + traceRows, + performance: performance.getEntriesByType("resource").slice(-30).map((entry) => ({ name: entry.name.split(/[?#]/u)[0].slice(0, 240), initiatorType: entry.initiatorType, startTime: Math.round(entry.startTime), duration: Math.round(entry.duration) })), + }; + }).catch((error) => ({ error: errorSummary(error), url: currentPageUrl() })); + const sample = { + seq: sampleSeq, + ts: new Date().toISOString(), + reason, + pageId, + commandId: activeCommandId, + observerInitiated: false, + ...digestDom(dom), + }; + await appendJsonl(files.samples, sample); + if (screenshotIntervalMs > 0 && Date.now() - lastScreenshotAtMs >= screenshotIntervalMs) { + await captureScreenshot("checkpoint", "jpeg").catch((error) => appendJsonl(files.errors, eventRecord("screenshot-error", { error: errorSummary(error) }))); + } + await writeHeartbeat({ status: terminalStatus }); +} + +function digestDom(dom) { + if (dom && dom.error) return dom; + const messages = Array.isArray(dom.messages) ? dom.messages.map((item) => ({ ...item, textHash: sha256Text(item.text || ""), textPreview: truncate(item.text || "", 160), textBytes: Buffer.byteLength(item.text || "") })) : []; + const traceRows = Array.isArray(dom.traceRows) ? dom.traceRows.map((item) => ({ ...item, textHash: sha256Text(item.text || ""), textPreview: truncate(item.text || "", 160), textBytes: Buffer.byteLength(item.text || "") })) : []; + return { ...dom, messages, traceRows }; +} + +async function captureScreenshot(reason, imageType = "png") { + if (!page || page.isClosed()) throw new Error("page is not available for screenshot"); + artifactSeq += 1; + const safeReason = safeId(String(reason || "manual")).slice(0, 40) || "manual"; + const type = imageType === "jpeg" || imageType === "jpg" ? "jpeg" : "png"; + const ext = type === "jpeg" ? "jpg" : "png"; + const file = path.join(dirs.screenshots, String(sampleSeq).padStart(6, "0") + "_" + String(artifactSeq).padStart(4, "0") + "_" + safeReason + "." + ext); + const options = type === "jpeg" ? { path: file, type: "jpeg", quality: 70, fullPage: false } : { path: file, type: "png", fullPage: false }; + await page.screenshot(options); + const meta = await fileMeta(file); + const artifact = { seq: artifactSeq, sampleSeq, ts: new Date().toISOString(), kind: "screenshot", reason, path: file, type, byteCount: meta.byteCount, sha256: meta.sha256, pageId, currentUrl: currentPageUrl() }; + await appendJsonl(files.artifacts, artifact); + lastScreenshotAtMs = Date.now(); + return artifact; +} + +function eventRecord(type, data) { + return { ts: new Date().toISOString(), type, jobId, pageId, sampleSeq, commandId: activeCommandId, ...sanitize(data) }; +} + +function controlRecord(command, phase, detail) { + return { + ts: new Date().toISOString(), + seq: commandSeq, + phase, + commandId: command.id, + type: command.type, + source: command.source || "file", + input: commandInputSummary(command), + beforeUrl: command.beforeUrl || null, + afterUrl: currentPageUrl(), + pageId, + detail: sanitize(detail), + }; +} + +function commandInputSummary(command) { + const text = typeof command.text === "string" ? command.text : null; + return { + type: command.type, + path: command.path || null, + url: command.url ? safeUrl(command.url) : null, + sessionId: command.sessionId || command.value || null, + label: command.label ? truncate(command.label, 200) : null, + textHash: text === null ? null : sha256Text(text), + textBytes: text === null ? null : Buffer.byteLength(text), + textPreview: null, + valuesRedacted: true, + }; +} + +async function appendJsonl(file, value) { + await appendFile(file, JSON.stringify(sanitize(value)) + "\n", { mode: 0o600 }); +} + +async function fileMeta(file) { + const [buffer, stats] = await Promise.all([readFile(file), stat(file)]); + return { byteCount: stats.size, sha256: "sha256:" + createHash("sha256").update(buffer).digest("hex") }; +} + +function currentPageUrl() { + try { return page && !page.isClosed() ? page.url() : null; } catch { return null; } +} + +function safeFrameUrl(frame) { + try { return frame ? safeUrl(frame.url()) : null; } catch { return null; } +} + +function safeUrl(value) { + try { + const url = new URL(String(value), baseUrl); + for (const key of Array.from(url.searchParams.keys())) { + if (/token|key|secret|password|auth|cookie/iu.test(key)) url.searchParams.set(key, "[redacted]"); + } + return url.toString(); + } catch { + return truncate(String(value || ""), 300); + } +} + +function normalizeBaseUrl(value) { + const raw = value || "http://127.0.0.1:3000"; + const url = new URL(raw); + return url.origin; +} + +function parseViewport(value) { + const match = String(value).match(/^(\d{3,5})x(\d{3,5})$/u); + return match ? { width: Number(match[1]), height: Number(match[2]) } : { width: 1440, height: 900 }; +} + +function positiveInteger(value, fallback) { + const parsed = Number(value); + return Number.isFinite(parsed) && parsed >= 0 ? Math.floor(parsed) : fallback; +} + +function sha256Text(value) { + return "sha256:" + createHash("sha256").update(String(value)).digest("hex"); +} + +function safeId(value) { + return String(value || "").replace(/[^A-Za-z0-9_.-]+/gu, "-").replace(/^-+|-+$/gu, "").slice(0, 120) || "item"; +} + +function cssEscape(value) { + return String(value).replace(/\\/gu, "\\\\").replace(/"/gu, "\\\""); +} + +function truncate(value, limit) { + const text = String(value || ""); + return text.length > limit ? text.slice(0, limit) + "..." : text; +} + +function sanitize(value) { + if (value === null || value === undefined) return value; + if (typeof value === "string") return value === password ? "[redacted]" : value.replaceAll(password, "[redacted]"); + if (typeof value === "number" || typeof value === "boolean") return value; + if (Array.isArray(value)) return value.map(sanitize); + if (typeof value === "object") return Object.fromEntries(Object.entries(value).map(([key, item]) => /password|cookie|authorization|token|secret/iu.test(key) ? [key, "[redacted]"] : [key, sanitize(item)])); + return String(value); +} + +function errorSummary(error) { + return { name: error && error.name ? error.name : "Error", message: error && error.message ? truncate(error.message, 1000) : truncate(String(error), 1000), stackTail: error && error.stack ? truncate(error.stack, 2000) : null }; +} + +function sleep(ms) { + return new Promise((resolve) => setTimeout(resolve, Math.max(0, ms))); +} +`; +} + +export function nodeWebObserveAnalyzerSource(): string { + return String.raw`#!/usr/bin/env node +import { createHash } from "node:crypto"; +import { mkdir, readFile, readdir, stat, writeFile } from "node:fs/promises"; +import path from "node:path"; + +const stateDir = path.resolve(process.env.UNIDESK_WEB_OBSERVE_STATE_DIR || process.argv[2] || ".state/web-observe/manual"); +const analysisDir = path.join(stateDir, "analysis"); +const reportJsonPath = path.join(analysisDir, "report.json"); +const reportMdPath = path.join(analysisDir, "report.md"); +const samples = await readJsonl(path.join(stateDir, "samples.jsonl")); +const control = await readJsonl(path.join(stateDir, "control.jsonl")); +const network = await readJsonl(path.join(stateDir, "network.jsonl")); +const errors = await readJsonl(path.join(stateDir, "errors.jsonl")); +const artifacts = await readJsonl(path.join(stateDir, "artifacts.jsonl")); +const manifest = await readJson(path.join(stateDir, "manifest.json")); +const heartbeat = await readJson(path.join(stateDir, "heartbeat.json")); + +await mkdir(analysisDir, { recursive: true, mode: 0o700 }); +const findings = buildFindings(samples, control, network, errors); +const transitions = buildTransitions(samples); +const commandTimeline = control.filter((item) => item.phase === "completed" || item.phase === "failed").map((item) => ({ ts: item.ts, phase: item.phase, commandId: item.commandId, type: item.type, input: item.input, afterUrl: item.afterUrl })); +const report = { + ok: findings.filter((item) => item.severity === "red").length === 0, + command: "web-probe-observe analyze", + generatedAt: new Date().toISOString(), + stateDir, + manifest: compactManifest(manifest), + heartbeat: compactHeartbeat(heartbeat), + counts: { samples: samples.length, control: control.length, network: network.length, errors: errors.length, artifacts: artifacts.length }, + commandTimeline, + transitions, + findings, + artifactSummary: await artifactSummary(artifacts), + safety: { offlineOnly: true, browserDriven: false, apiFetch: false, valuesRedacted: true }, +}; +await writeFile(reportJsonPath, JSON.stringify(report, null, 2) + "\n", { mode: 0o600 }); +await writeFile(reportMdPath, renderMarkdown(report), { mode: 0o600 }); +const [jsonMeta, mdMeta] = await Promise.all([fileMeta(reportJsonPath), fileMeta(reportMdPath)]); +console.log(JSON.stringify({ ok: true, command: "web-probe-observe analyze", stateDir, reportJsonPath, reportJsonSha256: jsonMeta.sha256, reportMdPath, reportMdSha256: mdMeta.sha256, counts: report.counts, findings: findings.slice(0, 20), valuesRedacted: true }, null, 2)); + +async function readJson(file) { + try { return JSON.parse(await readFile(file, "utf8")); } catch { return null; } +} + +async function readJsonl(file) { + try { + const text = await readFile(file, "utf8"); + return text.split(/\r?\n/u).map((line) => line.trim()).filter(Boolean).map((line) => { + try { return JSON.parse(line); } catch { return { parseError: true, rawHash: sha256(line) }; } + }); + } catch { return []; } +} + +function buildFindings(samples, control, network, errors) { + const findings = []; + const commandTimes = control.filter((item) => item.phase === "completed" || item.phase === "started").map((item) => Date.parse(item.ts)).filter(Number.isFinite); + const routeSessions = new Set(samples.map((item) => item.routeSessionId).filter(Boolean)); + const activeSessions = new Set(samples.map((item) => item.activeSessionId).filter(Boolean)); + if (routeSessions.size > 1) findings.push({ id: "session-route-changed", severity: "amber", summary: "route session changed during observation", routeSessionCount: routeSessions.size, samples: sampleRefs(samples, (item) => item.routeSessionId) }); + if (activeSessions.size > 1) findings.push({ id: "active-session-changed", severity: "amber", summary: "active session changed during observation", activeSessionCount: activeSessions.size, samples: sampleRefs(samples, (item) => item.activeSessionId) }); + const mismatches = samples.filter((item) => item.routeSessionId && item.activeSessionId && item.routeSessionId !== item.activeSessionId); + if (mismatches.length > 0) findings.push({ id: "route-active-session-mismatch", severity: "red", summary: "routeSessionId and activeSessionId diverged", count: mismatches.length, samples: mismatches.slice(0, 10).map(ref) }); + const uncommandedChanges = []; + for (let i = 1; i < samples.length; i += 1) { + const prev = digestSample(samples[i - 1]); + const next = digestSample(samples[i]); + if (prev !== next && !nearCommand(samples[i], commandTimes, 10000)) uncommandedChanges.push(ref(samples[i])); + } + if (uncommandedChanges.length > 0) findings.push({ id: "uncommanded-visible-state-change", severity: "amber", summary: "visible message/trace digest changed without a nearby command", count: uncommandedChanges.length, samples: uncommandedChanges.slice(0, 20) }); + const finalFlicker = detectFinalFlicker(samples); + if (finalFlicker.length > 0) findings.push({ id: "final-response-flicker", severity: "red", summary: "message text digest disappeared or switched to diagnostic-like text after non-empty final text", count: finalFlicker.length, samples: finalFlicker.slice(0, 20) }); + const scrollJumps = []; + for (let i = 1; i < samples.length; i += 1) { + const prevY = Number(samples[i - 1]?.scroll?.y ?? 0); + const nextY = Number(samples[i]?.scroll?.y ?? 0); + if (prevY > 250 && nextY < 40 && !nearCommand(samples[i], commandTimes, 8000)) scrollJumps.push({ from: ref(samples[i - 1]), to: ref(samples[i]) }); + } + if (scrollJumps.length > 0) findings.push({ id: "scroll-jump-top", severity: "amber", summary: "scroll position jumped near top without nearby command", count: scrollJumps.length, samples: scrollJumps.slice(0, 10) }); + const traceTerminal = samples.some((item) => Array.isArray(item.traceRows) && item.traceRows.some((row) => /completed|failed|canceled|terminal|done/iu.test((row.status || "") + " " + (row.textPreview || "")))); + const traceSeen = samples.some((item) => Array.isArray(item.traceRows) && item.traceRows.length > 0); + if (traceSeen && !traceTerminal) findings.push({ id: "trace-without-terminal", severity: "amber", summary: "trace rows were visible but no terminal status was sampled", firstTraceSample: ref(samples.find((item) => Array.isArray(item.traceRows) && item.traceRows.length > 0)) }); + const naturalApi = network.filter((item) => item.observerInitiated === false && item.type === "response" && /\/v1\/|\/auth\//u.test(String(item.url || ""))); + findings.push({ id: "natural-api-dom-lag-baseline", severity: "info", summary: "natural API responses and DOM samples are available for manual lag correlation", naturalApiResponses: naturalApi.length, sampleCount: samples.length }); + if (errors.length > 0) findings.push({ id: "browser-console-or-page-errors", severity: "amber", summary: "pageerror/runner errors were captured", count: errors.length, first: errors.slice(0, 5) }); + if (samples.length === 0) findings.push({ id: "no-samples", severity: "red", summary: "observer produced no samples" }); + return findings; +} + +function buildTransitions(samples) { + const rows = []; + let last = null; + for (const sample of samples) { + const digest = digestSample(sample); + if (digest !== last) { + rows.push({ seq: sample.seq, ts: sample.ts, url: sample.url, routeSessionId: sample.routeSessionId || null, activeSessionId: sample.activeSessionId || null, messageCount: Array.isArray(sample.messages) ? sample.messages.length : 0, traceRowCount: Array.isArray(sample.traceRows) ? sample.traceRows.length : 0, digest }); + last = digest; + } + } + return rows.slice(0, 200); +} + +function detectFinalFlicker(samples) { + const flickers = []; + let lastNonEmpty = null; + for (const sample of samples) { + const messageText = Array.isArray(sample.messages) ? sample.messages.map((item) => item.textPreview || "").join("\n") : ""; + const nonEmpty = messageText.trim().length > 0; + if (nonEmpty && !/temporarily|timeout|无法连接|暂时|error|failed|超时/iu.test(messageText)) lastNonEmpty = { sample, messageText }; + if (lastNonEmpty && nonEmpty && /temporarily|timeout|无法连接|暂时|error|failed|超时/iu.test(messageText)) flickers.push({ from: ref(lastNonEmpty.sample), to: ref(sample) }); + if (lastNonEmpty && !nonEmpty) flickers.push({ from: ref(lastNonEmpty.sample), to: ref(sample), reason: "text-disappeared" }); + } + return flickers; +} + +function digestSample(sample) { + const messages = Array.isArray(sample.messages) ? sample.messages.map((item) => item.textHash || item.textPreview || "").join("|") : ""; + const trace = Array.isArray(sample.traceRows) ? sample.traceRows.map((item) => (item.status || "") + ":" + (item.textHash || item.textPreview || "")).join("|") : ""; + return sha256((sample.routeSessionId || "") + "|" + (sample.activeSessionId || "") + "|" + messages + "|" + trace); +} + +function nearCommand(sample, commandTimes, windowMs) { + const ts = Date.parse(sample.ts); + return Number.isFinite(ts) && commandTimes.some((item) => Math.abs(ts - item) <= windowMs); +} + +function sampleRefs(samples, pick) { + const seen = new Set(); + const refs = []; + for (const sample of samples) { + const value = pick(sample); + if (!value || seen.has(value)) continue; + seen.add(value); + refs.push({ ...ref(sample), value }); + } + return refs.slice(0, 20); +} + +function ref(sample) { + if (!sample) return null; + return { seq: sample.seq ?? null, ts: sample.ts ?? null, url: sample.url ?? null, routeSessionId: sample.routeSessionId ?? null, activeSessionId: sample.activeSessionId ?? null }; +} + +async function artifactSummary(artifacts) { + const items = artifacts.slice(-30).map((item) => ({ kind: item.kind, reason: item.reason, sampleSeq: item.sampleSeq, path: item.path, sha256: item.sha256, byteCount: item.byteCount })); + return { count: artifacts.length, latest: items }; +} + +function compactManifest(value) { + if (!value) return null; + return { jobId: value.jobId, stateDir: value.stateDir, baseUrl: value.baseUrl, targetPath: value.targetPath, startedAt: value.startedAt, status: value.status, sampling: value.sampling, safety: value.safety }; +} + +function compactHeartbeat(value) { + if (!value) return null; + return { jobId: value.jobId, pid: value.pid, status: value.status, sampleSeq: value.sampleSeq, commandSeq: value.commandSeq, currentUrl: value.currentUrl, updatedAt: value.updatedAt, uptimeMs: value.uptimeMs }; +} + +function renderMarkdown(report) { + const findingLines = report.findings.length === 0 ? "- 无红灯项。" : report.findings.map((item) => "- " + item.severity + ": " + item.id + " - " + item.summary).join("\n"); + const commandLines = report.commandTimeline.length === 0 ? "- 无控制命令。" : report.commandTimeline.map((item) => "- " + item.ts + " " + item.phase + " " + item.type + " " + item.commandId + " " + (item.afterUrl || "")).join("\n"); + const transitionLines = report.transitions.length === 0 ? "- 无状态变化。" : report.transitions.slice(0, 80).map((item) => "- #" + item.seq + " " + item.ts + " messages=" + item.messageCount + " traceRows=" + item.traceRowCount + " route=" + (item.routeSessionId || "-") + " active=" + (item.activeSessionId || "-")).join("\n"); + return "# web-probe observe analysis\n\n" + + "- stateDir: " + report.stateDir + "\n" + + "- generatedAt: " + report.generatedAt + "\n" + + "- samples: " + report.counts.samples + "\n" + + "- control: " + report.counts.control + "\n" + + "- network: " + report.counts.network + "\n" + + "- errors: " + report.counts.errors + "\n\n" + + "## Findings\n\n" + findingLines + "\n\n" + + "## Command timeline\n\n" + commandLines + "\n\n" + + "## State transitions\n\n" + transitionLines + "\n"; +} + +async function fileMeta(file) { + const [buffer, stats] = await Promise.all([readFile(file), stat(file)]); + return { byteCount: stats.size, sha256: "sha256:" + createHash("sha256").update(buffer).digest("hex") }; +} + +function sha256(value) { + return "sha256:" + createHash("sha256").update(String(value)).digest("hex"); +} +`; +}