Traceback (most recent call last):
  File "daily.py", line 95, in <module>
    with Flow('start dataflow job 1', executor=LocalDaskExecutor(), storage=GCS('ghpr-dev-prefect-flows')) as startFlow1:
TypeError: __init__() got an unexpected keyword argument 'executor'
(prefect) bash-3.2$ prefect version
0.13.7

hmm i am unable to set the executor for a flow like this https://docs.prefect.io/api/latest/core/flow.html#flow-2 says that it does have the

executor

parameter though i am trying to increase parallelism for my flows

Hi folks, quick question about the Prefect agent and virtual environments. I have a machine with two registered flows, each flow runs inside its own virtual environment (I'm using

venv

). If I have two separate agents running on the machine, each started from within their own venv, they can both execute their own flows successfully. As soon as I try to cut it down to one agent listening for both flows, the agent fails to execute any flows which live inside another venv (

Failed to load and execute Flow's environment: ModuleNotFoundError

). I'd like to keep the separated venvs if possible and just have one agent running, which would execute each of the flows within the context of their own venv. Is this possible, or would I need to bite the bullet and have one venv for both flows?

Hello everyone, I’m trying to have an oriented approach to Flows and Tasks. The idea is to inherit specific flow types that already have some methods as tasks. Only overriding them if necessary. I realize that the decorator

task

doesn’t seem to work for class properties. Here is the error:

File "/Users/nuno/Developer/Data-Framework/data-prefect/data_prefect/utils/flows.py", line 67, in factory_flow
    flow.fetch()
  File "/Users/nuno/Developer/Data-Framework/data-prefect/.venv/lib/python3.8/site-packages/prefect/core/task.py", line 470, in __call__
    new.bind(
  File "/Users/nuno/Developer/Data-Framework/data-prefect/.venv/lib/python3.8/site-packages/prefect/core/task.py", line 511, in bind
    callargs = dict(signature.bind(*args, **kwargs).arguments)  # type: Dict
  File "/usr/local/Cellar/python@3.8/3.8.5/Frameworks/Python.framework/Versions/3.8/lib/python3.8/inspect.py", line 3025, in bind
    return self._bind(args, kwargs)
  File "/usr/local/Cellar/python@3.8/3.8.5/Frameworks/Python.framework/Versions/3.8/lib/python3.8/inspect.py", line 2940, in _bind
    raise TypeError(msg) from None
TypeError: missing a required argument: 'self'

It seems that I cannot pass the method “self” argument. Do you guys have any suggestion? Thank you in advance.

i am dynamically scheduling a flow run (to start not immediately but in the future) from within a flow and am polling the prefect AI to wait for its completion

def get_flow_run_state(client, flow_run_id):
    q = parse_graphql(
            {'query': {
                with_args('flow_run', 
                    {'where': {'id': {'_eq': flow_run_id}}}): {
                        'state'
                        }
                    }
                    })
    state = client.graphql(q).data.flow_run[0].state
    return state

def wait_flow_complete(flow_run_id):
    client = Client()

    state = None
    while state != 'Success':
        sleep(10)
        state = get_flow_run_state(client, flow_run_id)

question: is it possible to avoid the polling and achieve the same thing?

Hi. When using

DaskKubernetesEnvironment

how to setup the k8s namespace in which we want the job to run in the cluster? I'm using

scheduler_spec_file/worker_spec_file

and in the yaml files I set the

metadata:namespace: <name>

. But for the job itself it starts in the

default

namespace

Another question of `DaskKubernetesEnvironment`: I'm saving the images in azure container registry. I've created an

image_pull_secret

and I've tried with setting it in the custom scheduler/worker yaml files and also as an argument in

DaskKubernetesEnvironment

. Both fail. Looking the cluter error it gives is:

Failed to pull image "<image_url>": rpc error: code = Unknown desc = Error response from daemon: Get <image_url>: unauthorized: authentication required, visit <https://aka.ms/acr/authorization> for more information.

Looking into the yaml of the job submitted to the cluster (Pods->Actions->Edit) it is clear that there is no field as bellow:

imagePullSecrets:
    - name: regcred

Is it a bug or I'm doing something wrong? Thank you

Hi everyone, might overthink something but when I pulled the docker image how do I start the server? This is what I get as a response: FileNotFoundError: [Errno 2] No such file or directory: 'docker-compose': 'docker-compose'. Cheers, Marek

@Anna Geller (old account) Thanks for your setup guide on AWS EKS + prefect setup, very helpful. Others should check it out too: https://towardsdatascience.com/distributed-data-pipelines-made-easy-with-aws-eks-and-prefect-106984923b30

Hi all, trying to test out Dask w/ Prefect, getting this error when trying to run flow.run(executor=DaskExecutor()):

Hi there. I am working on a flow that depends on a

start_date

and

end_date

. These can come from two

Parameters

or, if the parameters are not provided, the dates are computed after applying some logic to

prefect.context.scheduled_start_time

. I'd like to use the computed start and end dates for templating the task result's location. If these dates are passed as task inputs, they are accesible (I can do this

location="{start_date}_{end_date}.txt"

) but I'd also like to use these variables in some downstream tasks that don't list them as inputs. Is there another way for a downstream task to access these? Since they don't always come from a Parameter, they are not available in

prefect.context.paramters

. I tried adding them to the context at run time with

prefect.context["start_date"] = start_dt

but modifying the context like this doesn't feel right. Any suggestions?

Is there a way to use the

StringFormatter

task with a template that is read at run time?

hi guys, my prefect agent can run an already-manually-pulled image but cannot itself pull the image so i am trying to mount my gcloud (google container registry) credentials like this:

$ prefect agent start docker -l gcs-flow-storage  --volume /var/run/docker.sock:/var/run/docker.sock  --volume ~/.config:/root/.config --volume ~/.docker:/root/.docker

but i am still getting

Unexpected error: ImageNotFound(HTTPError('404 Client Error: Not Found for url: <http+docker://localhost/v1.40/containers/create'>))
Traceback (most recent call last):
  File "/usr/local/lib/python3.8/site-packages/docker/api/client.py", line 259, in _raise_for_status
    response.raise_for_status()
  File "/usr/local/lib/python3.8/site-packages/requests/models.py", line 941, in raise_for_status
    raise HTTPError(http_error_msg, response=self)
requests.exceptions.HTTPError: 404 Client Error: Not Found for url: <http+docker://localhost/v1.40/containers/create>

Hi everybody. I want the results of my tasks to be saved based on either the

flow_run_id

or

flow_run_name

(i.e.

target='{flow_run_id}/foo'

). What I have trouble with is designing a method that works both when using Prefect Server and when using Prefect through the basic Python. For example, I am able to set the

flow_run_id

when running through the Python API with

flow.run(flow_run_id=<my flow run id>)

, however setting

flow_run_id

seems to be impossible when running through the Prefect server. Inversly, one can set

flow_run_name

when running through the Prefect Server but the concept of a

flow_run_name

doesn't seem to exist in the Python API. What would be the best way to resolve this issue? Thank you very much, Nejc

Hi Guys. I have an issue on prefect cloud displaying the schematic of a massive flow with about 160 tasks (we do some parallel stuff here where we currently cannot use the map logic but single tasks). When trying to open the schematic, I see a spinner for a while before the thing crashes with Error5. Is this known that big flows cannot be visualized?

I'm trying to run a flow using LocalEnvironment with DaskExecutor but the flow run never starts. It just keeps logging that the flow has entered the running state, and the Dask dashboard shows there's nothing running on the cluster. Has anyone ran into this problem before?

flow.environment = LocalEnvironment(
            executor=DaskExecutor(
                adapt_kwargs={
                    "minimum": 1,
                    "maximum": 5,
                },
            ),
        )

FWIW, I'm using prefect 0.13.2, dask 2.26.0, distributed 2.26.0

Hi guys. Does anyone know if the DAG can be accessed from within a task during runtime?

I have a flow that insert rows into a database. Sometimes a row or two are not inserted due to an error. Right now the flow terminates, but I want to continue running the flow but in the end set the flow run to failed or maybe to some kind of warning. Is that possible?

I am wondering if there is a mechanism in prefect for flow looping/mapping. Ie. I want to run a flow with N inputs and collect those outputs for a downstream task. I could write a for loop within the flow, but this creates a very wide DAG which becomes unwieldy after 1000 iterations.

Hi folks, I'm having issues with ModuleNotFoundError when running a flow from a local agent. I have a project structure like this:

|--project
   |-.venv
   |-etl
     |--__init__.py
     |--foo
     |  |--__init__.py
     |  |--foo_flow.py
     |--bar
     |  |--__init__.py
     |  |--bar_flow.py
     |--shared_utils
        |--__init__.py
        |--utils.py

In

foo_flow.py

and

bar_flow.py

I import modules from

shared_utils

by appending the etl folder to sys.path before importing

utils.py

. When I run any of these flows with

flow.run()

they work fine. However, when I do

flow.register()

instead and then start a local agent, from any of the directories listed above, all flow runs initiated from the server to the agent fail instantly with

Failed to load and execute Flow's environment: ModuleNotFoundError("No module named 'shared_utils'"

. According to a stackoverflow answer this is because the agent's python path doesn't include

project/etl

but it still fails when I run the agent with

--import-path "C:\project\etl"

. I've tried registering the flow from every directory listed above, and tried starting the agent from every directory listed above, and also tried passing every directory listed above as an

--import-project

argument, and I get exactly the same error every time. Can anybody please point out what I'm missing?

Hi everyone. I am developing a simple flow that has to query a database, save a file as csv, run some checks and SFTP the file over to a remote server. The query takes a while to run so I am using a templated

target

and

PandasSerializer

to write the csv. I am wondering what the recommended pattern is for the downstream SFTP task to locate the csv. If I were not relying on the

LocalResult

to save the file, I'd simply return the path to the local file from the function that generates the csv. However, since I am using the

Result

mechanism, this task just returns a dataframe. Do I just have to rely on the templated

target

location for downstream tasks to find the csv file? In other words, do I have to use the same template and render it manually in the SFTP task?

Hi everyone, I am currently running a dynamic job after a call from a db. I receive a configuration which allows me to know the number of steps to run. It looks like that: • build the config • for each part of the config run a step I have two implementations. The first one is mostly static while the second is more dynamic. What would be your recommended way? 1. static implementation

import prefect
from prefect import Flow, task

@task
def build_job():
    return {'steps': {'a': 'a', 'b': 'b', 'c': 'c'}}


@task(log_stdout=True)
def run_step(step):
    print(step)


with Flow('test') as flow:
    config = build_job()

    tmp_task = None
    for step_name in ['a', 'b', 'c']:
        upstream_tasks = None if tmp_task is None else [tmp_task]
        tmp_task = run_step(
            config['steps'][step_name], upstream_tasks=upstream_tasks
        )

flow.run()

2. Use loop to be more dynamic

import prefect
from prefect.engine.signals import LOOP
from prefect import Flow, task

@task
def build_job():
    return {'order': ['a', 'b', 'c'], 'steps': {'a': 'a', 'b': 'b', 'c': 'c'}}

@task(log_stdout=True)
def run_step(job_file):
    # we extract the accumulated task loop result from context
    loop_payload = prefect.context.get("task_loop_result", {})
    step_name = loop_payload.get("step_name", job_file['order'][0])

    step = job_file['steps'][step_name]

    print(step)

    pos = job_file['order'].index(step_name) + 1

    if pos < len(job_file['order']):
        next_step = job_file['order'][pos]
        raise LOOP(message=f"Fib {step_name}", result=dict(step_name=next_step))

with Flow('test') as flow:
    config = build_job()
    run_step(config)

Is there a plan to support something like ?

with Flow('test') as flow:
    config = build_job()

    tmp_task = None
    for name in config['order']:
        config['steps'][name]

Thanks !

👋 @Marvin 👋

This contest will all end in tears, I just know it. You're entered anyway, @abhilash.kr.

hi there! Does anyone know how to set this config item via environment variable for prefect server? (This is 0.12.6 - going to migrate soon, but not quite yet): https://github.com/PrefectHQ/prefect/blob/0.12.6/src/prefect/config.toml#L48

hi! is it possible to configure the Github storage environment to use a container on ECR?

👋 @Marvin

This contest is the sort of thing you lifeforms enjoy, is it? I've entered you to win, @Berty.

👋 @Marvin 🎉

This contest will all end in tears, I just know it. You're entered anyway, @Spencer.