在尝试从Treebank语料库解析特定句子时，这两个解析器表现得非常奇怪。
这是来自treebank的句子：
['A', 'form', 'of', 'asbestos', 'once', 'used', '*', '*', 'to', 'make', 'Kent', 'cigarette', 'filters', 'has', 'caused', 'a', 'high', 'percentage', 'of', 'cancer', 'deaths', 'among', 'a', 'group', 'of', 'workers', 'exposed', '*', 'to', 'it', 'more', 'than', '30', 'years', 'ago', ',', 'researchers', 'reported', '0', '*T*-1', '.']
EarleyChartParser在使用简单的CFG语法时，会产生一个包含24284棵树的巨大森林来找到正确的那棵树。
然而InsideChartParser似乎在一个队列排序的过程中陷入了无限循环，这是它内部的一个数据结构。
对于我使用的PCFG解析器，我使用这个函数：

def pchart(parser, sentence, gold_tree): #here the parser is a InsideChartParser
    test_trees = list(parser.parse(sentence))
    print("PCHART PARSER - TREES FOUND: ", len(test_trees))
    best_prob = 0.0
    for idx, test_tree in enumerate(test_trees):
        print("TREE: %d" %idx)
        print(test_tree)
        curr_prob = test_tree.prob()
        if curr_prob > best_prob:
            best_prob = curr_prob
            best_tree = test_tree
        if test_tree.productions() == gold_tree.productions(): #check if the tree iterated now is the correct one
            print("CORRECT TREE")
        else:
            print("WRONG TREE")
    return best_tree

而对于早期解析器，我使用这个：

def earley(parser, sentence, gold_tree): #here the parser is a EarleyChartParser
    test_trees = list(parser.parse(sentence)) #creates a forest of trees, with every trees being able to parse that sentence
    print("EARLEY PARSER - TREES FOUND: ", len(test_trees))
    for idx, test_tree in enumerate(test_trees):
        print("TREE: %d" %idx)
        print(test_tree)
        if test_tree.productions() == gold_tree.productions():
            print("CORRECT TREE")
        else:
            print("WRONG TREE")`

在枚举了这么多树之后，Early停止了，而InsideChartParser开始了，但陷入了一个非常长的循环，我认为这是由于此时树或产生式的数量巨大，而且没有很快终止：

<ipython-input-9-641b936c1306> in <module>
     24 
     25 if __name__ == '__main__':
---> 26     main()

<ipython-input-9-641b936c1306> in main()
     18     gold_tree = treebank.parsed_sents()[3]
     19     earley(cfg_earley_parser, sentence, gold_tree) #run earley parser and compare it with gold tree and shows the various trees
---> 20     tree = pchart(pcfg_pchart_parser, sentence, gold_tree) #shows the possible trees and returns the one with maximum probability
     21     print("BEST TREE WITH PROBABILITY: %.12e" %tree.prob())
     22     tree.draw()

<ipython-input-8-5e94e1748d29> in pchart(parser, sentence, gold_tree)
      1 def pchart(parser, sentence, gold_tree):
----> 2     test_trees = list(parser.parse(sentence))
      3     print("PCHART PARSER - TREES FOUND: ", len(test_trees))
      4 
      5     best_prob = 0.0

F:\Development\anaconda3\lib\site-packages\nltk\parse\pchart.py in parse(self, tokens)
    244         while len(queue) > 0:
    245             # Re-sort the queue.
--> 246             self.sort_queue(queue, chart)
    247 
    248             # Prune the queue to the correct size if a beam was defined

F:\Development\anaconda3\lib\site-packages\nltk\parse\pchart.py in sort_queue(self, queue, chart)
    359         :rtype: None
    360         """
--> 361         queue.sort(key=lambda edge: edge.prob())
    362 
    363 

F:\Development\anaconda3\lib\site-packages\nltk\parse\pchart.py in <lambda>(edge)
    359         :rtype: None
    360         """
--> 361         queue.sort(key=lambda edge: edge.prob())
    362 
    363

对于这样一个简单的语法，这是正确的行为吗？
因为这种情况并没有发生在我从treebank语料库中使用的其他句子上。
感谢您的时间！