Following a tutorial on Hypothesis I found this problem. I have to do a roundtrip test with a run_length encode/decode (explained in docstrings):
from hypothesis import given, strategies as st
from itertools import groupby
from typing import List, Union
def run_length_encoder(in_string: str) -> List[Union[str, int]]:
"""
>>> run_length_encoder("aaaaabbcbc")
['a', 'a', 5, 'b', 'b', 2, 'c', 'b', 'c']
"""
assert isinstance(in_string, str)
out = []
for item, group in groupby(in_string):
cnt = sum(1 for x in group)
if cnt == 1:
out.append(item)
else:
out.extend((item, item, cnt))
assert isinstance(out, list)
assert all(isinstance(x, (str, int)) for x in out)
return out
def run_length_decoder(in_list: List[Union[str, int]]) -> str:
"""
>>> run_length_decoder(['a', 'a', 5, 'b', 'b', 2, 'c', 'b', 'c'])
"aaaaabbcbc"
"""
assert isinstance(in_list, list)
assert all(isinstance(x, (str, int)) for x in in_list)
out: str = ""
for item in in_list:
if isinstance(item, int):
out += out[-1] * (item - 2)
else:
out += item
# alternative
# for n, item in enumerate(in_list):
# if isinstance(item, int):
# char = in_list[n - 1]
# assert isinstance(char, str)
# out += char * (item - 2)
# else:
# out += item
assert isinstance(out, str)
return out
I can choose the form of the test: encode(decode(in_list)) or decode(encode(in_string)).
@given(
in_string = st.text()
)
def test_roundtrip_run_length_encoder_decoder(in_string):
in_string = in_string
encoded_list = run_length_encoder(in_string)
assert isinstance(encoded_list, list)
assert all(isinstance(x, (str, int)) for x in encoded_list)
decoded_string = run_length_decoder(encoded_list)
assert isinstance(decoded_string, str)
assert in_string == decoded_string, (in_string, decoded_string)
test_roundtrip_run_length_encoder_decoder()
This was easy but in_string doesn’t have enough repetitions. They ask me to do something better (suggestion: use one_of).
So i should add a random numbers of ripetition with a random lenght (ee vs eee) in random positions. How can I do that with Hypothesis? Maybe they are asking me something more simple
I think to generate the list is more difficoult: I should generate a list without repetitions and add some sequence like [..., 'k', 'k', #, ... where k is a character (string) and # is an integer (int). Ofcourse before 'k' I need a different character.
2
def text_with_repeats(alphabet=st.characters()):
# It's fine if this encoder wouldn't round-trip; we just
# want to make a string with more repeats and it does that.
return st.lists(
st.tuples(st.integers(1, 5), alphabet)
).map(lambda ls: "".join(n * char for n, char in ls))
@given(
in_string = st.text() | text_with_repeats()
)