Hello All, Thanks in advance . I am getting an error while using flow.from_source with Git repo. Module not found.

Flow could not be retrieved from deployment.
Traceback (most recent call last):
  File "<frozen importlib._bootstrap_external>", line 940, in exec_module
  File "<frozen importlib._bootstrap>", line 241, in _call_with_frames_removed
  File "/tmp/tmpwp9f0k88prefect/branch/src/flows/simpletest_git_p2/simpletest_git_p2.py", line 12, in <module>
    from botocore.exceptions import ClientError
ModuleNotFoundError: No module named 'botocore'

Its running fine once I commented out the package botocore . I saw an option env, but i don't know how / where to write. Dockerfile is also in the git repo. I do have a docker image with all required pip list. Below is the code used .

if __name__ == "__main__":
    simpletest_git_p2.from_source(
        source=GitRepository(
        url="<https://github.cms.gov/gitrepo.git>",
        branch="branch",
        credentials={
            "access_token": Secret.load("github-access-token")
        }
    ),
    entrypoint=f"src/flows/{flow_name}/{flow_name}.py:{flow_name}",
    ).deploy(
        name="private-git-storage-deployment",
        work_pool_name="workpool",
        build=False 
    )

Copy code

build=False

if you're not building an image that has your dependencies, where is your image that's used at runtime coming from?

i have an image

its gonna be something like

from prefect.deployments import DeploymentImage

if __name__ == "__main__":
    simpletest_git_p2.from_source(
        source=GitRepository(
        url="<https://github.cms.gov/gitrepo.git>",
        branch="branch",
        credentials={
            "access_token": Secret.load("github-access-token")
        }
    ),
    entrypoint=f"src/flows/{flow_name}/{flow_name}.py:{flow_name}",
    ).deploy(
        name="private-git-storage-deployment",
        work_pool_name="workpool",
        build=True,
        image=DeploymentImage(dockerfile="Dockerfile.mine", image="repo/image_name:tag")
    )

ok let me try

sorry, in the above I meant

build=True,

ok... let me try

yeah build False means we dont do anything to build or push the image, so at runtime you probably still had the prefect base image

@Kamal I did not solved at all

for me issue resolved. by configuring workpool and work space properly . Thanks to my Devops team

Can you please share how you managed to do that? Thanks

we are using Kubernetes type of work pool, 1. gave the image name under image 2. gave the name space name under namespace 3. gave service account name 4. We also had a prefect agent workpool, which we turned off/deleted i am not sure my Devops created the workspace , but please refer the below document for reference. https://docs.prefect.io/latest/concepts/work-pools/#:~:text=the%20PREFECT_WORKER_HEARTBEAT_SECONDS%20setting.-,Starting%20a%20worker,%2D%2Dtype%20flag%20is%20used.