fix: track in-flight requests by id; handle llama-swap upsert/remove events

llama-swap emits inflight events with operation 'upsert' (a single request,
re-emitted as elapsed_ms updates) and id-only 'remove' events. The decoder
ignored 'upsert' and dropped every id-only remove, so the tracker only ever
reflected the connect-time snapshot and the count got stuck (e.g. at '1').

The tracker now keys in-flight state by request id (snapshot/add replace by
id; remove deletes by id), and the decoder normalizes 'upsert' to 'add' and
extracts the id from bare removes.

src/lib/event-feed.ts

@@ -9,6 +9,13 @@ export function decodeEvent(msg: SseMessage): FeedEvent | null {
     const inner = JSON.parse(outer.data) as Record<string, unknown>;
 
     if (outer.type === "inflight") {
+      const operation = inner.operation as string;
+      if (operation === "remove") {
+        const id = typeof inner.id === "string" ? inner.id : undefined;
+        if (id) return { type: "inflight", operation: "remove", id };
+        return null;
+      }
+      if (operation !== "snapshot" && operation !== "upsert" && operation !== "add") return null;
       const requests: InflightRequest[] = Array.isArray(inner.requests)
         ? (inner.requests as InflightRequest[])
         : Array.isArray(inner.request)
@@ -18,7 +25,7 @@ export function decodeEvent(msg: SseMessage): FeedEvent | null {
             : [];
       return {
         type: "inflight",
-        operation: inner.operation as "snapshot" | "add" | "remove",
+        operation: operation === "snapshot" ? "snapshot" : "add",
         requests,
       };
     }

src/lib/inflight-tracker.ts

@@ -13,29 +13,25 @@ export interface ModelState {
 }
 
 export type FeedEvent =
-  | { type: "inflight"; operation: "snapshot" | "add" | "remove"; requests: InflightRequest[] }
+  | { type: "inflight"; operation: "snapshot" | "add"; requests: InflightRequest[] }
+  | { type: "inflight"; operation: "remove"; id: string }
   | { type: "modelStatus"; models: ModelState[] };
 
 export type ModelRuntimeState = "stopped" | "loading" | "ready";
 
 export class InflightTracker {
-  private counts = new Map<string, number>();
+  private requests = new Map<string, string>();
   private states = new Map<string, ModelRuntimeState>();
 
   apply(event: FeedEvent): void {
     if (event.type === "inflight") {
       if (event.operation === "snapshot") {
-        const counts = new Map<string, number>();
-        for (const r of event.requests) counts.set(r.model, (counts.get(r.model) ?? 0) + 1);
-        this.counts = counts;
+        this.requests = new Map<string, string>();
+        for (const r of event.requests) if (r.id) this.requests.set(r.id, r.model);
       } else if (event.operation === "add") {
-        for (const r of event.requests) this.counts.set(r.model, (this.counts.get(r.model) ?? 0) + 1);
+        for (const r of event.requests) if (r.id) this.requests.set(r.id, r.model);
       } else if (event.operation === "remove") {
-        for (const r of event.requests) {
-          const c = this.counts.get(r.model) ?? 0;
-          if (c <= 1) this.counts.delete(r.model);
-          else this.counts.set(r.model, c - 1);
-        }
+        this.requests.delete(event.id);
       }
     } else if (event.type === "modelStatus") {
       for (const m of event.models) this.states.set(m.id, normalizeState(m.state));
@@ -43,7 +39,9 @@ export class InflightTracker {
   }
 
   count(modelId: string): number {
-    return this.counts.get(modelId) ?? 0;
+    let n = 0;
+    for (const model of this.requests.values()) if (model === modelId) n++;
+    return n;
   }
 
   state(modelId: string): ModelRuntimeState | undefined {