hi, basic question as i'm having trouble getting set up with dagster cloud. i have the following directory structure, where assets.py depends on

shared

and

lib

(and other libraries within `~/dev`:

~/dev/
├─ shared/
├─ lib/
├─ dagster_project/
│  ├─ dagster_project/
│  │  ├─ assets.py
│  │  ├─ __init__.py
│  ├─ dagster_cloud.yaml
│  ├─ setup.py

when i try to deploy from within

~/dev/dagster_project

, it fails because of the missing dependencies on

shared

and

lib

; when i deploy from

~/dev

, it fails because it can't find the correct

dagster_project

module. what's the correct command or process here?

hey @Vanessa Cai there's a bit of CICD magic to do here, and a few different approaches one approach is to make the shared packages and lib packages available as installable dependencies, this approach is discussed here: https://github.com/dagster-io/dagster/discussions/12340 another approach is to copy those libraries into the dagster_project as part of the CICD process, so that they remain top-level shared libraries locally and in your Git repo, but become sub-libraries from Dagster's perspective during deployment. an example of this approach is the

ml_project

in this repo: https://github.com/slopp/dagteam/tree/main. As documented in the bottom of that repo's readme, the steps are: (a) deploy from ~/dev, but with

dagster_cloud.yaml

in the root

/dev

directory (b) use the

build

argument within

dagster_cloud.yaml

to point at dagster_project (c) have the CICD process copy shared and lib into the dagster_project, and then install those libraries using an appropriate reference in

setup.py

or a

dagster_project/requirements.txt

file

thank you!