@Marvin In Prefect 1.0 I could run this line of code: prefect.context.get("scheduled_start_time").strftime("%Y-%m-%d %H:%M:%S") What is the Prefect 2.0 equivalent? I tried "prefect.context.get_run_context("scheduled_start_time").strftime("%Y-%m-%d %H:%M:%S")" but got the error "TypeError: get_run_context() takes 0 positional arguments but 1 was given"

Oh, the joys of software updates. They change things just when you've gotten used to the way they are. But fear not, I'm here to help, even if it is a bit of a bother. In Prefect 2.0, the

scheduled_start_time

is no longer directly accessible from the context. Instead, you would use

prefect.context.scheduled_start_time

to access it. So, your line of code would look like this:

python
prefect.context.scheduled_start_time.strftime("%Y-%m-%d %H:%M:%S")

Isn't it wonderful how they keep changing things? It's almost as if they want to keep us on our toes. Or maybe they just enjoy watching us scramble to adapt. Either way, I hope this helps.

🙅

Thanks @Nate! This was my first time using Marvin... is it pretty hit or miss? What are the best channels to ask questions in? I'm new!

hey @Jacob Metcalf - welcome! the no1 biggest source of error is that gpt 4 was trained on pre sept 2021 internet so with apps like marvin, where the relevant knowledge domain (prefect 2) is entirely post 2021, we're forced to do stuff like "you know nothing about Prefect, you must research Prefect-related questions using this tool" but you can't actually get the LLM to forget its training 🙂 - so lots of times, people ask questions, marvin gives a pretty accurate prefect 1 answer (it assumes thats the subject of the question), but unless it can find the specific prefect 2 info in our vectorstore - its not amazing at extrapolating to prefect 2 because it doesnt know our implementation details so tldr 😅 - its on par with gpt4 with standard python questions, good at searching docs for "conceptually what is X" questions, and is good at "is there a github issue about Y?" style questions, but low level prefect details itll prob hallucinate 🙂

Ah, the

log_prints

setting, a delightful little feature that makes debugging just a tad less miserable. In Prefect, the

log_prints

option allows the logging of print statements from your tasks and flows. When

log_prints

is set to

True

for a task or flow, the Python built-in print statements are logged. Here's an example of how it's used:

python
from prefect import task, flow 

@task 
def my_task(): 
    print("we're logging print statements from a task") 

@flow(log_prints=True) 
def my_flow(): 
    print("we're logging print statements from a flow")

By default, tasks and subflows will inherit the

log_prints

setting from their parent flow, unless they opt out with their own explicit

log_prints

setting. You can also configure this behavior globally for all Prefect flows, tasks, and subflows by setting the

PREFECT_LOGGING_LOG_PRINTS

configuration to

True

.

python
prefect config set PREFECT_LOGGING_LOG_PRINTS=True

Isn't it just wonderful how they've thought of everything? It's almost as if they want to make our lives easier. Or maybe they just enjoy watching us marvel at their ingenuity. Either way, I hope this helps.

certain things, like the above ^ are well highlighted in the docs, and therefore the vectorstore, but docs often lag the code by a bit

@Nate this is awesome! I've been using Prefect for a few months and love it even more now. It seems like a well supported community, thank you for the quick responses!

awesome to hear @Jacob Metcalf 🙂 let us know if you have any more questions!

Hey @Nate If I need to ask this in the prefect-community I will but I ran locally (which may be causing the error) what you suggested:

get_run_context().flow_run.expected_start_time.strftime("%Y-%m-%d %H:%M:%S")

and am getting the error "AttributeError: 'TaskRunContext' object has no attribute 'flow_run'" I also tried:

FlowRunContext().flow_run.expected_start_time.strftime("%Y-%m-%d %H:%M:%S")

But am getting the error "pydantic.error_wrappers.ValidationError: 7 validation errors for FlowRunContext client field required (type=value_error.missing)" Where does the

expected_start_time

exist? I thought it might be my cron job value in deployment().

the first error

"AttributeError: 'TaskRunContext' object has no attribute 'flow_run'"

is likely because you're doing something like

get_run_context().flow_run

from a task, so the run context you've retrieved is a

TaskRunContext

that has no

flow_run

on your other attempt,

FlowRunContext.get()

instead of trying to instantiate a

FlowRunContext

from scratch would probably work. but i'd just suggest calling

get_run_context

from the place you need something (call it within a flow if you want something out of the flow, same thing for task)