Lumi/plugins/lumi_transcription/backend/logs/jsonl_log.js

69 lines
3.0 KiB
JavaScript

const fs = require("fs");
const path = require("path");
class JsonlDiagnosticLog {
constructor(directory, options = {}) {
this.directory = directory;
this.retentionMs = (options.retentionDays || 7) * 86400000;
this.maxBytes = options.maxBytes || 256 * 1024 * 1024;
this.includeCaptionText = options.includeCaptionText !== false;
fs.mkdirSync(directory, { recursive: true });
}
append(entry) {
const safe = sanitize({ timestamp: new Date().toISOString(), ...entry }, this.includeCaptionText);
fs.appendFileSync(this.fileForToday(), `${JSON.stringify(safe)}\n`, { encoding: "utf8", mode: 0o600 });
}
cleanup(now = Date.now()) {
const files = this.files();
let removed = 0;
for (const file of files) {
if (now - file.mtimeMs > this.retentionMs) { fs.rmSync(file.path, { force: true }); removed += 1; }
}
const retained = this.files();
let total = retained.reduce((sum, file) => sum + file.size, 0);
for (const file of retained) {
if (total <= this.maxBytes) break;
fs.rmSync(file.path, { force: true });
total -= file.size;
removed += 1;
}
return { removed, bytes: total };
}
files() {
return fs.readdirSync(this.directory).filter((name) => /^transcription-\d{4}-\d{2}-\d{2}\.jsonl$/.test(name)).map((name) => {
const target = path.join(this.directory, name);
const stat = fs.statSync(target);
return { path: target, name, size: stat.size, mtimeMs: stat.mtimeMs };
}).sort((a, b) => a.mtimeMs - b.mtimeMs);
}
fileForToday() { return path.join(this.directory, `transcription-${new Date().toISOString().slice(0, 10)}.jsonl`); }
}
function sanitize(value, includeCaptionText) {
if (Buffer.isBuffer(value)) return "[binary omitted]";
if (Array.isArray(value)) return value.map((entry) => sanitize(entry, includeCaptionText));
if (typeof value === "string") return sanitizeText(value);
if (!value || typeof value !== "object") return value;
const result = {};
for (const [key, child] of Object.entries(value)) {
if (isSensitiveKey(key)) result[key] = "[redacted]";
else if (!includeCaptionText && /(?:stable|uncertain|caption)_text/i.test(key)) result[key] = "[caption text disabled]";
else result[key] = sanitize(child, includeCaptionText);
}
return result;
}
function isSensitiveKey(key) {
return /(?:audio|pcm|credential|(?:^|[_-])(?:authorization|cookie|password|passwd|secret|signature|token|api[_-]?key|private[_-]?key)(?:$|[_-]))/i.test(String(key));
}
function sanitizeText(value) {
return String(value || "")
.replace(/\b(Bearer|Basic|LumiDevice)\s+[A-Za-z0-9._~+/=-]+/gi, "$1 [redacted]")
.replace(/\b((?:set-)?cookie\s*:\s*)[^\r\n]+/gi, "$1[redacted]")
.replace(/([?&](?:token|key|secret|password|authorization)=)[^&#\s]+/gi, "$1[redacted]")
.replace(/\b((?:api[_-]?key|access[_-]?token|refresh[_-]?token|client[_-]?secret|password|passwd|secret|signature|authorization)\s*[:=]\s*)[^\s,;}]+/gi, "$1[redacted]");
}
module.exports = { JsonlDiagnosticLog, sanitize, sanitizeText };