fix(turn-metrics): exact per-turn placement via turnId, decode-only TPS, step-level retention

Placement
- Chunks carry their producing turn: StoredChunk.turnId (backend [email protected]) for
  committed rows (persisted in the IDB cache; old cached rows are upgraded on
  sync) and the live event's turnId for provisional ones (the optimistic echo is
  tagged on turn-start).
- interleaveTurnMetrics attributes every group to its turn (turnId, else the
  turn owning its stepId) and puts the turn total after the turn's LAST visible
  group. No user-message counting: sending, streaming, steering, images, errored
  turns and chat-limit unloading can no longer shift metrics between turns — at
  any point during a turn, not just after it ends. A turn cut in half by the
  chat limit keeps its total under its visible tail.
- Legacy rows (no turnId) fall back to an order-preserving, count-guarded match
  that leaves a turn blank rather than show another turn's numbers.
- A turn that errors re-syncs metrics (the backend now persists them).

TPS
- Step TPS = outputTokens / decodeMs only (no genTotalMs fallback — it includes
  time-to-first-token). Turn TPS sums tokens and decode time over timed steps
  only. Tool execution is excluded by the backend.

Display
- Retention = this turn's first-step cache reads / previous turn's last-step
  prompt (step-level; turn sums made it pin at 100%).
- Turn row shows output tokens + input processed (with cached), instead of a
  misleading 'total tok' that summed every step's full prompt.

2 files changed, 17 insertions, 1 deletions

diff --git a/src/adapters/idb/index.test.ts b/src/adapters/idb/index.test.ts
index c05c605..04e2177 100644
--- a/src/adapters/idb/index.test.ts
+++ b/src/adapters/idb/index.test.ts
@@ -30,6 +30,16 @@ describe("createIdbChunkStore", () => {
     expect(loaded[0]?.chunk).toEqual(textChunk("a"));
   });
 
+  it("round-trips a chunk's turnId (and upserts it onto an older row)", async () => {
+    const store = createIdbChunkStore({ indexedDB: new IDBFactory() });
+    await store.append("conv1", [makeChunk(1, "a"), makeChunk(2, "b")]);
+    expect((await store.load("conv1")).map((c) => c.turnId)).toEqual([undefined, undefined]);
+    await store.append("conv1", [{ ...makeChunk(1, "a"), turnId: "t1" }]);
+    const loaded = await store.load("conv1");
+    expect(loaded.map((c) => c.turnId)).toEqual(["t1", undefined]);
+    expect("turnId" in (loaded[1] ?? {})).toBe(false);
+  });
+
   it("append out-of-order still loads seq-ordered", async () => {
     const store = createIdbChunkStore({ indexedDB: new IDBFactory() });
     const chunks = [makeChunk(3, "c"), makeChunk(1, "a"), makeChunk(2, "b")];
diff --git a/src/adapters/idb/index.ts b/src/adapters/idb/index.ts
index 96b2cbc..3dffef9 100644
--- a/src/adapters/idb/index.ts
+++ b/src/adapters/idb/index.ts
@@ -14,6 +14,7 @@ interface ChunkRecord {
   seq: number;
   role: StoredChunk["role"];
   chunk: StoredChunk["chunk"];
+  turnId?: string;
 }
 
 interface MetaRecord {
@@ -70,7 +71,11 @@ function keyRangeFor(conversationId: string): IDBKeyRange {
 }
 
 function chunksToStoredChunks(records: ChunkRecord[]): StoredChunk[] {
-  return records.map((r) => ({ seq: r.seq, role: r.role, chunk: r.chunk }));
+  return records.map((r) =>
+    typeof r.turnId === "string"
+      ? { seq: r.seq, role: r.role, chunk: r.chunk, turnId: r.turnId }
+      : { seq: r.seq, role: r.role, chunk: r.chunk },
+  );
 }
 
 export function createIdbChunkStore(opts?: CreateIdbChunkStoreOptions): ConversationChunkStore {
@@ -113,6 +118,7 @@ export function createIdbChunkStore(opts?: CreateIdbChunkStoreOptions): Conversa
           seq: c.seq,
           role: c.role,
           chunk: c.chunk,
+          ...(c.turnId !== undefined ? { turnId: c.turnId } : {}),
         } satisfies ChunkRecord);
       }