Admin Replay: re-run the pipeline over past calls in a sandbox (#170)

Correlation has only ever been measured through live AI windows: days of
wall time per change, and the 09-20→22 window was invalidated outright by
unfunded AI accounts (#169). Recordings are kept regardless of AI, so the
traffic to measure against already exists.

- internal/replay.py: runs a time range of real calls through the live
  pipeline code in original order, clock pinned per call, into
  replay_runs/{run_id}/calls|incidents. Modes: audio (re-transcribe),
  transcripts (re-extract), reuse (correlation only from a prior run's
  scenes). Simulates the idle-resolve and orphan-recorrelation sweeps on
  virtual time. No alerts, summaries, vocab, AI-health alerts or pending
  terms. One run at a time, <=5000 calls, <=7 days.
- firestore.py: ContextVar sandbox redirect for calls/incidents.
- clock.py: ContextVar-pinnable now(), used on the correlation path.
- feature_flags.py: ContextVar flag override so replay runs with live AI off.
- upload.py: scene loop extracted to _extract_and_correlate, shared by the
  live pipeline and replay so replay measures the code that runs live.
- resolved_via on every incident resolve, so a real clear can be told
  from the idle timeout — live and in replay.
- routers/replay.py + /admin Replay tab: estimate, start, compare runs,
  drill into incidents with audio.

Reviewed by drb-correlation-review; its leak and fidelity findings are
fixed and covered by tests. c2-core: 456 pass. Frontend typecheck not run
(no Node on the authoring box).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

drb-c2-core/app/internal/feature_flags.py

@@ -6,7 +6,8 @@ in-memory TTL cache so flag reads don't add a Firestore round-trip to every
 call upload.
 """
 import time
-from typing import Any
+from contextvars import ContextVar
+from typing import Any, Optional
 from app.internal.logger import logger
 from app.internal import firestore as fstore
 
@@ -36,6 +37,21 @@ _DEFAULTS: dict[str, bool] = {
     "transcript_correction_enabled": True,
 }
 
+# A replay run (app/internal/replay.py) states exactly which AI steps it runs,
+# independent of the live switches — the whole point is re-running the pipeline
+# while live AI is OFF. ContextVar so the override never reaches a live upload.
+_forced: ContextVar[Optional[dict[str, bool]]] = ContextVar("drb_forced_flags", default=None)
+
+
+def force_flags(flags: Optional[dict[str, bool]]):
+    """Override resolve_flags() for the current context. Returns a reset token."""
+    return _forced.set(flags)
+
+
+def unforce_flags(token) -> None:
+    _forced.reset(token)
+
+
 _cache: dict[str, Any] = {}
 _cache_ts: float = 0.0
 
@@ -211,6 +227,11 @@ async def resolve_flags(system_id: str | None):
     """
     from app.internal import firestore as _fstore
 
+    forced = _forced.get()
+    if forced is not None:
+        full = {k: bool(forced.get(k, False)) for k in _DEFAULTS}
+        return full, lambda name: full.get(name, False)
+
     flags = await get_flags()
 
     system_ai_flags: dict = {}