根据来自另一个列表/字典的顺序对列表中的文件名进行排序

Question

I have a sorted list of filenames given like this:

files = ['root/base/val1/apples/pkernel', 
         'root/base/val1/oranges/pkernel',  
         'root/base/val1/eng_scope_lattice/p2_kernel',
         'root/base/val2/grapes/pkernel',
         'root/base/val2/exact_scope_lattice/p2_kernel',
         'root/base/val2/peaches/pkernel',
         'root/base/val2/pineapple/pkernel']

and I have a dictionary value_dict :

value_dict = {'val1':[oranges,apples], 'val2':[peaches, grapes, pineapples]}

I have sorted the list files but I also want to sort the files that end with "pkernel" inside each value ( val1 and val2 in this case) based on the order in the value_dict . So "oranges" will come before "apples" for val1 and similarly we will use the order specified in value_dict . I also have other files with extensions different than p2kernel whose order does not need to be changed.

So my final_list will be

final_list = ['root/base/val1/oranges/pkernel', 
              'root/base/val1/apples/pkernel',  
              'root/base/val1/eng_scope_lattice/p2_kernel',
              'root/base/val2/peaches/pkernel',
              'root/base/val2/exact_scope_lattice/p2_kernel',
              'root/base/val2/grapes/pkernel',
              'root/base/val2/pineapple/pkernel']

I was trying to use the sorted(s, key = operator.itemgetter(1, 2)) methods but I don't know how to apply the results of the dictionary in subsets of sorting. So I end up doing this process manually using sublime text operations. Is there a way to automate this?

For file extensions other than p1kernel the order in the original list should stay intact.

Answer 1

It is very unclear to me exactly what rules you intend to apply to the sorting. Also this way of doing things is pretty hack-y and makes a lot of assumptions about what your inputs are allowed to look like. That being said this almost does what you ask except in your example you put items from the lists before other items for the val 1 folder but not the val2 folder. Anyway I think you aught to be able to make things work based on this code. Edited: fixed copy paste error in the code.

def getKey(val): 
    for k,v in value_dict.items: 
        if val.find(k) != -1: 
            for i in range(len(v)): 
                val = val.replace(v[i],str(i)) 
    return val

                 
sorted(file, key=getKey)

Answer 2

Yes, there is more than a way to automate this. I will explain you one very simple algorithm, maybe not the fastest but it's better than using sublime text operations.

1. Transform your array of files to an array of arrays, I mean use split("/") on each string in order to convert them on a list.
2. You can use sorted(new_list, key=lambda path_file_list: customKeyFunction(path_file_list))

Where

def customKeyFunction(path_file):
  val_path = path_file[2]
  try:
    key = value_dict[val_path].index(path_file[3])
  except ValueError:
    key = -1
  return key


sorted(new_list, key=lambda path_file_list: customKeyFunction(path_file_list))

Disclaimer: this will change the order of the p2kernel files. But with this you have something to start from.

Answer 3

Your question is ill-defined with "order will not be changed", so I will assume that keys not present are sorted to the beginning or the end.

Here is one option:

>>> sorted(files, key=sort_order)
['root/base/val1/eng_scope_lattice/p2_kernel',
 'root/base/val1/oranges/pkernel',
 'root/base/val1/apples/pkernel',
 'root/base/val2/exact_scope_lattice/p2_kernel',
 'root/base/val2/pineapple/pkernel',
 'root/base/val2/peaches/pkernel',
 'root/base/val2/grapes/pkernel']

where we define sort_order as follows:

import math

files = [
    'root/base/val1/apples/pkernel',
    'root/base/val1/oranges/pkernel',  
    'root/base/val1/eng_scope_lattice/p2_kernel',  
    'root/base/val2/grapes/pkernel', 
    'root/base/val2/exact_scope_lattice/p2_kernel',
    'root/base/val2/peaches/pkernel',
    'root/base/val2/pineapple/pkernel'
]
_orders = {
    'val1': ['oranges', 'apples'],
    'val2': ['peaches', 'grapes', 'pineapples']
}
orders = {k: {val: ind for ind, val in enumerate(v)} for k, v in _orders.items()}
digits = {k: int(math.ceil(math.log(len(v), 10))) for k, v in orders.items()}

BASE = ['root', 'base']

def sort_order(file):
    fragments = file.split('/')
    if fragments[:2] == BASE:
        if len(fragments) > 3:
            folder, subfolder = fragments[2:4]
            if folder in orders:
                index = orders[folder].get(subfolder, '')  # Put unknown first
                str_index = index and f'{index:0{digits[folder]}d}'
            fragments[3] = f'{str_index}/{subfolder}'
    return fragments

We have to do silly things because python3 doesn't allow sticking eg (1, 'foo') in the middle of a list of strings and comparing them, in the form

[['root', 'base', 'val1', '1/apples', 'pkernel'],
 ['root', 'base', 'val1', '0/oranges', 'pkernel'],
 ['root', 'base', 'val1', '/eng_scope_lattice', 'p2_kernel'],
 ['root', 'base', 'val2', '1/grapes', 'pkernel'],
 ['root', 'base', 'val2', '/exact_scope_lattice', 'p2_kernel'],
 ['root', 'base', 'val2', '0/peaches', 'pkernel'],
 ['root', 'base', 'val2', '/pineapple', 'pkernel']]