Best practices for generating classes

[Axel Vestin]

Hello! I'm trying out Hypothesis and PBT for a function that filters a list.

Each item in the list has the following information:

An identifying string

A time where it was issued

A time interval during which it is valid

But it may also have the following extra information:

A reference to another item in the list which should be considered cancelled.

A time interval during which the cancellation is in effect.

I wanted to try to use Hypothesis to make sure that all cases that can occur are handled, which would take many test cases using traditional test methods.

For now I tried writing a very naïve strategy that returns a valid time interval and a cancellation time interval, i've appended it below.

What I'm wondering is; How can I write this code such that I let Hypothesis generate a list where the above holds true? Or am I going about this in a wrong way?

@st.composite

def future_complete_cancellation(

draw,

) -> tuple[dt.datetime, dt.datetime, dt.datetime, dt.datetime, dt.datetime]:

current_time = draw(st_time())

issue_time = draw(st_time(min=current_time + dt.timedelta(minutes=1)))

valid_from = draw(st_time(min=issue_time))

valid_to = draw(st_time(min=valid_from + dt.timedelta(minutes=1)))

cancel_start = draw(st_time(min=issue_time, max=valid_from))

cancel_issue_time = draw(st_time(min=issue_time, max=cancel_start))

cancel_end = draw(st_time(min=valid_to))

return (

current_time,

issue_time,

valid_from,

valid_to,

cancel_issue_time,

cancel_start,

cancel_end,

)

@given(future_complete_cancellation())

@settings(max_examples=5000)

def test_future_cancellation(scenario):

(

current_time,

issue_time,

valid_from,

valid_to,

cancel_issue_time,

cancel_start,

cancel_end,

) = scenario

t01 = get_sigmet(

seq_no="T01", issue_time=issue_time, valid_from=valid_from, valid_to=valid_to

)

t02 = get_cancel_sigmet(

seq_no="T02",

cancel_seq_no=t01.sequence_number,

issue_time=cancel_issue_time,

cancel_from=cancel_start,

cancel_to=cancel_end,

)

output = filter_current_sigmets([t01, t02], current_time)

assert len(output) == 0