I would like to use Hops instead of clusters. One of the advantages of working with clusters is that I can troubleshoot a cluster by checking the ‘live’ data inside of it. In other words, I can see exactly in the cluster where the script fails by checking the data of the in/outputs of the blocks. The same method of using the live data can be used to easily expand functionality of an exisiting cluster.
Is there a similar way of debugging/developing with live data inside a Hops definition?
They are not mutually exclusive. Hops has more latency than clusters (but it’s great!).
As to debugging, I would say not with live data. However, debugging can be done with static data, either with internalized geometry or using Data Input. Of course, Data Output must then be used in the GH model calling Hops, but only for the static testing.
Hops does have more latency than clusters for single operations, but it does have the potential to beat clusters for cases where solutions can be split into many parallel calls.
You can also right click on a hops component and save the current data being fed into it as json. You can then read that json using a tool like jswan into the other definition to trace through what is happening for a specific set of input.