siftlode/frontend/src/lib/log.ts

// The client log (E22 S4): the browser half of the ONE event shape `backend/app/logs.py` defines.
//
// Every event is `{ts, level, sev, source: "web", event, msg, attrs, request_id?, tab_id, version}` —
// the same fields and level names as a server line, so a client batch can land in `log_events` beside
// the server's own rows and a request joins across the two by `request_id`.
//
// Events are kept in memory only, in a bounded ring: nothing leaves the browser from here. Sending
// them is `lib/diagnostics.ts`'s job, and only while the user's consented level is raised (E22 S5);
// it and the Settings log viewer follow the ring through `subscribe`.
//
// REDACTION HAPPENS HERE, at capture, not at the call site — the same rules as the server's handler,
// held to them by a fixture both test suites run (`backend/tests/fixtures/log_redaction_cases.json`).
// A secret cannot leak by forgetting to scrub it at one call site, and a new call site is safe by
// default. Never kept, at any level: passwords, tokens, cookies, secret config, E2EE material, email
// addresses. What a level ADDS (filter values at DEBUG, search strings at TRACE) is decided by the
// call site asking `enabled(level)` before it builds the attribute.
//
// Dependency-free on purpose (only the build stamp): the request layer, the notification store and
// the error boundary all log through here, so it must not import any of them.
import { FRONTEND_VERSION } from "./version";

export type Level = "TRACE" | "DEBUG" | "INFO" | "WARNING" | "ERROR" | "CRITICAL";

/** OpenTelemetry severity numbers — the same table as `logs.SEVERITY` (a contract test compares). */
export const SEVERITY: Record<Level, number> = {
  TRACE: 1,
  DEBUG: 5,
  INFO: 9,
  WARNING: 13,
  ERROR: 17,
  CRITICAL: 21,
};

/** How a log viewer groups the six levels into its filter chips — "Saját napló" and the admin
 *  console both read a log through these, so a level lands in the same chip on either screen. */
export type LevelGroup = "all" | "error" | "warning" | "info" | "detail";

export const LEVEL_GROUPS: Record<Exclude<LevelGroup, "all">, readonly Level[]> = {
  error: ["ERROR", "CRITICAL"],
  warning: ["WARNING"],
  info: ["INFO"],
  detail: ["DEBUG", "TRACE"],
};

export interface LogEvent {
  ts: string;
  level: Level;
  sev: number;
  source: "web";
  event: string;
  msg: string;
  attrs?: Record<string, unknown>;
  request_id?: string;
  tab_id: string;
  version: string;
  /** The debug session recording when the event was kept (E16). */
  debug_session_id?: string;
}

/** Events kept. At ~0.5 KB each the ring stays around a megabyte — a long session's worth of
 *  navigation and requests, without growing for as long as a tab stays open. */
export const RING_MAX = 2000;

// --- redaction (mirrors backend/app/logs.py) ------------------------------------------------------
// Substring matching on the KEY: `password`, `current_password`, `plex_server_token`, … — a key we
// have never seen is safe by default.
const SECRET_KEY_PARTS = [
  "password",
  "passphrase",
  "token",
  "secret",
  "cookie",
  "authorization",
  "auth_header",
  "ciphertext",
  "private_key",
  "api_key",
  "apikey",
  "credential",
  "session_key",
];

// Credential-bearing URL parameters (OAuth `code`/`state`, `?token=` links, `#verify=`/`#reset=`).
const SECRET_PARAM_RE =
  /\b((?:access_token|refresh_token|id_token|token|code|state|key|secret|password|pwd|sig|signature|verify|reset|setup)=)([^&\s"']+)/gi;

// Deliberately loose, as on the server: over-matching is harmless, under-matching is not.
const EMAIL_RE = /[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}/g;

const REDACTED = "[redacted]";
const REDACTED_EMAIL = "[email]";

/** Scrub free text: credential-bearing URL params and email addresses. */
function redactText(text: string): string {
  return text.replace(SECRET_PARAM_RE, `$1${REDACTED}`).replace(EMAIL_RE, REDACTED_EMAIL);
}

/** Recursively scrub a structure destined for the log. A value whose KEY names a secret is dropped
 *  whole — a prefix of a token is still a leak. Everything else is scrubbed as text. */
export function redact(value: unknown, key = ""): unknown {
  const lowered = key.toLowerCase();
  if (key && SECRET_KEY_PARTS.some((part) => lowered.includes(part))) return REDACTED;
  if (Array.isArray(value)) return value.map((v) => redact(v, key));
  if (typeof value === "string") return redactText(value);
  if (value && typeof value === "object") {
    return Object.fromEntries(Object.entries(value).map(([k, v]) => [k, redact(v, k)]));
  }
  return value;
}

// --- identity -------------------------------------------------------------------------------------
function randomHex(bytes: number): string {
  const buf = new Uint8Array(bytes);
  crypto.getRandomValues(buf);
  return Array.from(buf, (b) => b.toString(16).padStart(2, "0")).join("");
}

/** An id for one outgoing request, sent as `X-Request-Id`. The server accepts it (it matches
 *  `logs._REQUEST_ID_RE`) and stamps its own line with it, which is what joins the two logs. */
export function newRequestId(): string {
  return randomHex(8);
}

// Per TAB, not per page load: a reload loop is one tab's story, and sessionStorage is exactly that
// scope. Storage that throws (a locked-down browser) just means a fresh id per load.
const TAB_KEY = "siftlode.logTab";
function tabId(): string {
  try {
    const known = sessionStorage.getItem(TAB_KEY);
    if (known) return known;
    const id = randomHex(6);
    sessionStorage.setItem(TAB_KEY, id);
    return id;
  } catch {
    return randomHex(6);
  }
}
const TAB_ID = tabId();

// --- the logger -----------------------------------------------------------------------------------
let threshold: Level = "INFO";
let ring: LogEvent[] = [];
// The debug session this tab is recording into (E16), or null. Held HERE, the lowest layer, because
// two things below everything else need it: every kept event carries it, and `api/core` puts it on
// every request as `X-Debug-Session` — which is how the server tags its own lines of that request.
let debugSession: string | null = null;

/** Record into `id` from now on (null = stop). `lib/diagnostics` sets it from `me.debug_session`. */
export function setDebugSession(id: string | null): void {
  debugSession = id;
}

/** The debug session recording right now, if any. */
export function currentDebugSession(): string | null {
  return debugSession;
}

/** The capture level. INFO is the always-on "Alap"; DEBUG and TRACE come from the user's consented
 *  diagnostics level (E22 S5). */
export function setLevel(level: Level): void {
  threshold = level;
}

/** Would an event at `level` be kept? Ask before building an attribute that costs something or that
 *  only a higher level may carry (a search string is TRACE-only). */
export function enabled(level: Level): boolean {
  return SEVERITY[level] >= SEVERITY[threshold];
}

/** The kept events, oldest first (a copy). */
export function events(): LogEvent[] {
  return ring.slice();
}

/** Drop every kept event (the viewer's "clear", tests). */
export function clear(): void {
  ring = [];
  notify(null);
}

// Told about every kept event as it is kept (`null` = the ring was cleared). A listener that throws
// is its own problem: it must not stop the others, nor the code that logged.
type Listener = (event: LogEvent | null) => void;
const listeners = new Set<Listener>();

/** Follow the ring: called with each event once it is kept, and with `null` when it is cleared.
 *  Returns the unsubscribe. */
export function subscribe(listener: Listener): () => void {
  listeners.add(listener);
  return () => {
    listeners.delete(listener);
  };
}

function notify(event: LogEvent | null): void {
  for (const listener of listeners) {
    try {
      listener(event);
    } catch {
      // see Listener
    }
  }
}

function emit(
  level: Level,
  event: string,
  msg: string,
  attrs?: Record<string, unknown>,
  requestId?: string,
): void {
  if (!enabled(level)) return;
  try {
    const entry: LogEvent = {
      ts: new Date().toISOString(),
      level,
      sev: SEVERITY[level],
      source: "web",
      event,
      msg: redactText(msg),
      tab_id: TAB_ID,
      version: FRONTEND_VERSION,
    };
    if (attrs && Object.keys(attrs).length) entry.attrs = redact(attrs) as Record<string, unknown>;
    if (requestId) entry.request_id = requestId;
    if (debugSession) entry.debug_session_id = debugSession;
    ring.push(entry);
    if (ring.length > RING_MAX) ring.splice(0, ring.length - RING_MAX);
    notify(entry);
  } catch {
    // Logging must never break the code that logs (an attr with a throwing getter, a cycle).
  }
}

type Emit = (event: string, msg: string, attrs?: Record<string, unknown>) => void;

/** `log.info("nav.view", "page opened", {page})` — a stable dotted event code, a short English
 *  message for developers, structured attributes. */
export const log: Record<"trace" | "debug" | "info" | "warn" | "error" | "fatal", Emit> & {
  at: (level: Level, event: string, msg: string, attrs?: Record<string, unknown>) => void;
  request: (level: Level, msg: string, attrs: Record<string, unknown>, requestId: string) => void;
} = {
  trace: (e, m, a) => emit("TRACE", e, m, a),
  debug: (e, m, a) => emit("DEBUG", e, m, a),
  info: (e, m, a) => emit("INFO", e, m, a),
  warn: (e, m, a) => emit("WARNING", e, m, a),
  error: (e, m, a) => emit("ERROR", e, m, a),
  fatal: (e, m, a) => emit("CRITICAL", e, m, a),
  /** At a level decided at run time (a notification's, a status code's). */
  at: (level, e, m, a) => emit(level, e, m, a),
  /** One `http.request` event, carrying the request id as the correlation field. */
  request: (level, msg, attrs, requestId) => emit(level, "http.request", msg, attrs, requestId),
};

/** An error as loggable attributes: its kind, message and stack — redacted like everything else. */
export function errorAttrs(err: unknown): Record<string, unknown> {
  if (err instanceof Error) {
    return { error: err.name, message: err.message, ...(err.stack ? { stack: err.stack } : {}) };
  }
  return { error: typeof err, message: String(err) };
}