如何将生成触觉刺激数组的重复代码改写为循环？

Hey there! I totally get that writing repetitive code to create 10 separate zero arrays feels clunky—let's streamline this with a clean loop. Here are two easy approaches to generate your stim1 through stim10 arrays without redundancy:

Option 1: Store arrays in a list (simple and efficient) #

Using a list keeps things straightforward, and you can access each stimulus array via its index (e.g., stim_list[0] for the first one, equivalent to stim1).

import numpy as np

# First, keep your existing rand_stim_seconds function (or use the optimized version below!)
def rand_stim_seconds(array):
    lst = list(range(0,10))
    random_ints = []
    for i in lst:
        rand_int = np.random.uniform(0.1, 0.6, 1)
        random_ints.append(rand_int)
    values = []
    for j in random_ints:
        d = j * 22050
        e = float(d)
        f = int(np.round(e))
        values.append(f)
    add_lst = [0]
    for k in values:
        add = add_lst[-1] + k
        if add > 44100:
            break
        add_lst.append(add)
    for l in add_lst[1:]:
        array[l] = 1
    return array

# Generate 10 independent stimulus arrays
stim_list = []
for _ in range(10):
    # Create a brand new zero array each time to avoid overwriting data
    new_stim_array = np.zeros(44100, dtype=int)
    stim_list.append(rand_stim_seconds(new_stim_array))

# Access example: stim_list[0] is your first stimulus array, stim_list[9] the tenth

Option 2: Store arrays in a dictionary (for named access) #

If you prefer keeping the stim1 to stim10 naming convention, a dictionary lets you access each array by its name directly.

# Generate 10 named stimulus arrays
stim_dict = {}
for num in range(1, 11):
    new_stim_array = np.zeros(44100, dtype=int)
    stim_dict[f"stim{num}"] = rand_stim_seconds(new_stim_array)

# Access example: stim_dict["stim1"], stim_dict["stim2"], etc.

Bonus: Optimize your rand_stim_seconds function #

Since you're working with numpy, we can simplify your function using vectorized operations (faster and cleaner than loops):

def rand_stim_seconds(array):
    # Generate 10 random intervals in one line (no loop needed)
    random_intervals = np.random.uniform(0.1, 0.6, 10)
    
    # Convert intervals to array indices in a single vectorized step
    values = np.round(random_intervals * 22050).astype(int)
    
    add_lst = [0]
    for k in values:
        next_pos = add_lst[-1] + k
        if next_pos > 44100:
            break
        add_lst.append(next_pos)
    
    # Assign 1s to all target positions at once (no loop)
    array[add_lst[1:]] = 1
    return array

The key thing here is that we create a new zero array inside each loop iteration—this ensures each stimulus array is independent and doesn't get overwritten by previous calls to the function.

内容的提问来源于stack exchange，提问作者sallicap