cluster list elements into sub-lists if an element is greater than threshold

I have a list which has str , int and float elements.
I want to create sub-lists (cluster elements) in such a way that if any numeric element is greater than threshold (here threshold is 3), a cluster will be created containing the previous elements. Code and expected output is given below.

L = ["this is", "my", 1, "first line", 4, "however this", 3.5 , "is my last line", 4] 
out = []
temp = []
for x in L:
  if isinstance(x, str) or x <3:
    temp.append(x)
  else:
    out.append(temp)
    temp = []
  
print(out)
[['this is', 'my', 1, 'first line'], ['however this'], ['is my last line']]

I want to check how can I can cluster in more advanced way such as itertools.groupby or filter or any other advanced methods ?

>Solution :

You can use itertools.groupby using key as the predicate:

import itertools

L = ["this is", "my", 1, "first line", 4, "however this", 3.5 , "is my last line", 4] 

print([
    list(values)  # Convert the groups as a list
    for key, values in itertools.groupby(L, key=lambda x: isinstance(x, str) or x < 3)
    if key        # If the predicate is true
])

It outputs

[['this is', 'my', 1, 'first line'], ['however this'], ['is my last line']]

and how can I remove int/float elements with same groupby method

Use a nested list comprehension:

import itertools

L = ["this is", "my", 1, "first line", 4, "however this", 3.5 , "is my last line", 4] 

print([
    [value for value in values if not isinstance(value, (int, float))]
    for key, values in itertools.groupby(L, key=lambda x: isinstance(x, str) or x < 3)
    if key
])

It outputs

[['this is', 'my', 'first line'], ['however this'], ['is my last line']]