Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

okay so I am trying to translate a prefect core flow to prefect cloud flow. I want to be able to make changes locally and have it update my cloud flow. Is this possible?

You will need to re-register for the most part for the changes to take effect

And then how do I trigger a cloud run from local after registration?

With Python,
```Client.create_flow_run(...)```
With CLI
```prefect run --project project_name --name flow_name```
will trigger a run on Cloud

So my ultimate goal is to have the prefect flows be a docker container, so I don't want them to register them every single time they run. But only when I run them locally, I want it to re-register. Does that make sense?

Not super understanding, but that seems like it can be solved by a `if ___name___ == ___main___` to register right?

In my file there is this
```    registered_flow_id = flow.register(project_name="Development")
    client.create_flow_run(flow_id=registered_flow_id)```

so if I were to run it now I would force it to register every single time it runs.

Ah ok you can do `flow.register(build=False)` by default I guess?

Or is the issue here you don’t have an id to pass to `create_flow_run`? cuz you can use the task instead which is more flexible:
```create_flow_run.run(project_name=...,flow_name=...)```

Task docs are <https://docs.prefect.io/api/latest/tasks/prefect.html#create-flow-run|here>

Did you mean this one? *prefect.tasks.prefect.flow_run.StartFlowRun.run*

That is also an option, but `create_flow_run` is the newer task

The inconsistency with StartFlowRun is that it changes the return depending if you set `wait=True`. wait=False returns an id and wait=True returns a state.

So it was decoupled into `create_flow_run` and `wait_for_flow_run` but yes any task can be invoked as plain Python anyway with the run method

And there's not a secret way to have the flow detect changes right? lol

You can compare the `flow.serialized_hash()`. So you can try running this twice:
```prefect register -p some_flow.py```
and on the second one, it will not register because no changes were made. This is because the CLI uses `flow.serialized_hash()` as an idempotency key. So if you try to register with the same idempotency key, it will just return the id of the old run

It’s a bit more complicated than that but you can see the source code of the CLI register call <https://github.com/PrefectHQ/prefect/blob/master/src/prefect/cli/build_register.py#L388-L397|here>

So it is not able to do that in-code, just via the cli?

Yeah that code path is only in the CLI so you’d need to replicate it or come up with your own identifier around the serialized hash