Skip to main content
 首页 » 编程设计

python之如何解析一个 STIX 模式以生成 OR 逻辑的多模式

2024年11月24日9haluo1

我有一个 STIX 模式如下:

stix_ptn_and_or_case2 = '''[(x:x.x = 'A' OR x:x.x = 'B' ) AND ( x:x.x = 'C' OR x:x.x = 'D' )]''' 

模式的逻辑是(A OR B) AND (C OR D)

该模式可以通过 dendrol(python 库)解析为表达式树,如下所示

(A OR B) AND (C OR D)的逻辑可以拆分成下面4个小的AND only logic

  • A AND C
  • A AND D
  • B AND C
  • B AND D

我想解析表达式树以获得4个AND only逻辑

所以我写了如下的python代码。 python 库 dendrol 可以将 STIX 模式转换为表达式树。

import os, sys, datetime, copy 
from dendrol import Pattern 
 
stix_ptn_and_or_case2 = '''[(x:x.x = 'A' OR x:x.x = 'B' ) AND ( x:x.x = 'C' OR x:x.x = 'D' )]''' 
 
ptn = Pattern(stix_ptn_and_or_case2) 
pdt = ptn.to_dict_tree() 
obj = pdt['pattern'] 
all_ptn_element_list = [] 
 
# the recursion function for expression tree to and-only logic 
def ptnct(obj, ptn_element_list): 
    if('observation' in obj or 'expression' in obj): 
        if('observation' in obj): 
            join = obj['observation']['join'] 
            expressions  = obj['observation']['expressions'] 
        else: 
            join = obj['expression']['join'] 
            expressions  = obj['expression']['expressions'] 
        if(join=='AND' or join==None): 
            for exp in expressions: 
                p = ptnct(exp, ptn_element_list) 
                if(p!=None): 
                    ptn_element_list.append(p) 
            if(len(ptn_element_list)==len(expressions)):# all and  
                all_ptn_element_list.append(ptn_element_list) 
        elif(join=='OR'): 
            for exp in expressions: 
                p = ptnct(exp, ptn_element_list) 
                tmp = copy.deepcopy( ptn_element_list ) 
                tmp.append(p) 
                all_ptn_element_list.append(tmp)                     
 
    elif('comparison' in obj): 
        exp = obj 
        tag = '{0}:{1}.{2}'.format(exp['comparison']['object'], exp['comparison']['path'][0], exp['comparison']['path'][1]) 
        value = exp['comparison']['value'] 
        return value 
 
x=ptnct(obj, []) 
print(all_ptn_element_list) 

代码的输出是

[[u'A'], [u'B'], [u'C'], [u'D']] 

但是想要的输出是

[[u'A', u'C'], [u'A', u'D'], [u'B', u'C'], [u'B', u'D']] 

实际上代码适用于大多数情况,但不适用于此。

对于将 AND-OR 逻辑表达式树解析为 AND-only 表达式(模式)有什么建议吗?

请您参考如下方法:

使用方便的 itertools.product 和 itertools.chain,我们可以将分离表达式树转换为所有可能的连接表达式列表

from itertools import product, chain 
 
from dendrol import Pattern 
 
 
def extract_atomic_chunks(root): 
    container = root.get('observation') or root.get('expression') 
    comparison = root.get('comparison') 
 
    if container: 
        join = container['join'] 
        expressions = container['expressions'] 
 
        if join == 'OR': 
            # Return all possibilities under this node 
            return chain( 
                extract_atomic_chunks(expression) 
                for expression in expressions 
            ) 
 
        elif join == 'AND': 
            # Generate all possible N-tuples, where N=len(expressions)  
            groups = [ 
                extract_atomic_chunks(expression) 
                for expression in expressions 
            ] 
            return product(*groups) 
 
    elif comparison: 
        return comparison['value'] 
 
 
if __name__ == '__main__': 
    stix_expr = "[(x:x.x = 'A' OR x:x.x = 'B') AND (x:x.x = 'C' OR x:x.x = 'D')]" 
 
    pattern = Pattern(stix_expr) 
    tree = pattern.to_dict_tree() 
    root = tree['pattern'] 
 
    print(list(extract_atomic_chunks(root))) 

我想这会让你到达你想要的地方

[('A', 'C'), ('A', 'D'), ('B', 'C'), ('B', 'D')]