Websocket connection/reconnection cpu spike

[1-leo]

Describe the bug
Ndk takes really long for the first request with very high cpu usage

the problem is caused by

    _stateStreamSubscription = _websocketDS.ws.connection.listen((state) {
      Logger.log.t("${_websocketDS.url} connection state changed to $state");
      if (state is Connected || state is Reconnected) {
        if (!completer.isCompleted) {
          completer.complete();
        }
        if (state is Reconnected && onReconnect != null) {
          onReconnect.call();
        }
      } else if (state is Disconnected) {
        completer = Completer();
        ready = completer.future;
      } else if (state is Connecting || state is Reconnecting) {
        // Do nothing, just waiting for (re)connection to be established
      } else {
        Logger.log.w(
            "${_websocketDS.url} connection state changed to unknown state: $state");
      }
    });

Because a callback is called the issue might be upstream.
Commenting this code resolves the issue but breaks future requests

I suspect a loop when relay is in connecting/reconnecting state causing the cpu spike

[1-leo]

created #319 to experiment

[1-leo]

the mentioned code is not the cause of the problem!
The cause is that as a lot of data steams in, the websocket connections consumes all resources of the main thread => move to isolate
The second big spike is cause by unmarshalling the json => move to isolate

However there are even more issues especally with Nip01Event, we should make it immutable and use copyWith
=> refactor needed
=> move nip13 code to a usecase
=> move calculateId() elswhere.

Because calculateId() is called in the constructor we run it more then we should (sha265 is expensive)

I did a messy implementation in #319 but we should do this properly (with only one isolate for all ws connections)

[frnandu]

Need to pinpoint exactly where we actually use the calculateId() more than we should, apart from db impls.