从文件Python查找2d文件中分离的块数

Question

Sorry in advance if the title was a bit confusing, not sure how to word it better

In python, if you were to have a seperate 'file.txt' containing the following text:

.....%%%%%..% %%%...%%...%% %.....%%%..%% ...%%.....%%% ....%%.....%% .....%%%..%%% %%%%....%%%.. Representing splatters of paint, I'm trying to find a way to output the number of individual chunks, in this case this would return '4', I'm having trouble visualising how I would go about doing this, any help would greatly be appreciated. Thanks!

Answer 1

I'm still not sure what is a "chunk" but this assumes that it's a cluster of % elements:

import numpy as np
from scipy import ndimage

pattern = [];
with open('patterns.txt', 'r') as f:
    for line in f:
        line = line.replace('%','1').replace('.','0').replace('',' ')
        pattern.append((line.strip(' \n ')).split())

image = np.array(pattern, dtype = int)

pstruct = [[1,1,1],[1,1,1],[1,1,1]]
labeled_array, num_features = ndimage.measurements.label(image,structure = pstruct)

This program reads your input file of . and % into a nested list. It turns all % to 1 and . to 0 and generates a numpy array of integers with 1 and 0 representing your original pattern. It then uses a scipy function to look for clusters of 1s.

pstruct defines what is considered a cluster, and when the default is used it produces 5 chunks, for details look here https://docs.scipy.org/doc/scipy-.14.0/reference/generated/scipy.ndimage.measurements.label.html .

labeled_array are chunks numbered by their cluster number and num_features is the number of clusters.

If this does not properly define your chunk, clarify it a bit more or even better - upload an image where you explicitly mark the chunks.