Recently there was a bug introduced that only showed when using the hosted hub. Finding this bug would require running a naptha-sdk run module example (naptha run) using LOCAL_HUB=False. I think we probably don't need to run every example. Maybe we can just do a complex one that makes use of sub-deployments (e.g. multiagent chat). Unfortunately, I can't think of an easy way to dynamically change the HUB_URL without restarting node.
As you can see, the testing is often coupled between (a) launching a node, and (b) running a module on that node. Ideally, with each node launch (whether mac, ubuntu or docker) we would run at least one module example. I'm not sure what the best way to do this with GH actions is.
Recently there was a bug introduced that only showed when using the hosted hub. Finding this bug would require running a naptha-sdk run module example (
naptha run) using LOCAL_HUB=False. I think we probably don't need to run every example. Maybe we can just do a complex one that makes use of sub-deployments (e.g. multiagent chat). Unfortunately, I can't think of an easy way to dynamically change the HUB_URL without restarting node.As you can see, the testing is often coupled between (a) launching a node, and (b) running a module on that node. Ideally, with each node launch (whether mac, ubuntu or docker) we would run at least one module example. I'm not sure what the best way to do this with GH actions is.