How do I generate a variable sized list from a given list using Hypothesis?

For property-based testing, given a fixed list of values, I need to generate a variable-sized list where order is important and duplicates are allowed. For example, if my fixed list is

texts = ['t1', 't2', 't3', 't4']

I would like to generate different variations, e.g.

['t2']
['t4', 't1'] # Subset and different order
[]
['t3', 't1', 't2'] # Different order
['t4', 't4', 't4', 't1'] # Repetition of t4
['t1', 't2', 't1'] # Repetition but at different location
['t1', 't2']
['t2', 't1'] # different order from the one above and considered different.

What I have managed to use currently is the permutations strategy

from hypothesis import given, strategies as st

@given(st.permutations(texts))
def test_x(some_text):
   ...
   pass

But that does not give me variable size, repetitions

Other requirements:

How can I specify a maximum variable list of 20?

Solution

You are looking for a combination of the lists and the sampled_from strategies:

from hypothesis import strategies as st

texts = ['t1', 't2', 't3', 't4']
lists_from_texts = st.lists(st.sampled_from(texts), max_size=20)

...

@given(lists_from_texts)
def test_x(some_text):
    ...

or if you want to be able to change the source list for different tests:

from typing import List


def lists_from_texts(source: List[str]) -> st.SearchStrategy[List[str]]:
    return st.lists(st.sampled_from(source), max_size=20)

...

@given(lists_from_texts(texts))
def test_x(some_text):
    ...