ui: improve manual model load and cancel (#847)

- When a model is manually loaded show a cancel buttton and a queued
status
- Implement cancellation in scheduler.Scheduler interface and FIFO
scheduler
- Add cache bust query parameter to bypass browser cache

Fixes #844

internal/router/scheduler/scheduler.go

@@ -47,6 +47,11 @@ type Swapper interface {
 type Scheduler interface {
 	// OnRequest handles one incoming ServeHTTP request.
 	OnRequest(req HandlerReq)
+	// OnCancel handles a request whose client has disconnected before it was
+	// granted. The scheduler must remove the request from its queue and from
+	// any in-flight swap's waiters so it never triggers a model load or grant
+	// for a caller that is no longer there.
+	OnCancel(req HandlerReq)
 	// OnSwapDone handles a swap goroutine reporting completion.
 	OnSwapDone(ev SwapDone)
 	// OnServeDone handles a tracked ServeHTTP finishing (in-flight decrement).