Python中如何合并多词汽车品牌识别结果的拆分项？

问题场景 #

我们有一个包含多词汽车品牌的制造商列表：

['Nissan', 'Honda', 'Rolls Royce', 'Ford', 'Kia', 'Land Rover']

用户输入句子：

I drive a Rolls Royce, a Ford, and a Land Rover.

需要识别句子中的所有汽车制造商，但当前用集合交集实现的代码会将多词品牌拆分为单个单词，返回结果为：

I found 5 automobiles in your sentence
Your automobiles are ['Royce', 'Rolls', 'Ford', 'Land', 'Rover']

期望输出是正确合并多词品牌的结果：

Your automobiles are ['Rolls Royce', 'Ford', 'Land Rover']

当前问题代码 #

import string

result = []
autos = [ 'Mazda',  'Toyota', 'Volkswagon', 'Honda', 'Ford',
    'Chevolet', 'Tesla', 'Kia', 'Hyndai', 'Rolls Royce',
    'Isuzu', 'Jeep', 'Land Rover', 'Mitzubishi', 'Subaru'
]
print("original autos",autos)

### Split each item
new_autos = []
for auto in autos:
    new_autos.extend(auto.split())
print("\nnew_autos", new_autos)


### ask the user for a sentence
sentence = input('Enter a sentence with favorite automobiles of your choice: ')
title_case = sentence.title()

### remove punctuation
new_sentence = title_case.translate(str.maketrans('', '', string.punctuation))

### convert it to a list
sentence_list = list(new_sentence.split(" "))
print(list(sentence_list))


set1 = set(sentence_list)
set2 = set(new_autos)
result = list(set1 & set2)

### tried differnt apporaches to getting the discovered items back in their original order
###result = list(set1.intersection(set2))
###result = list(set1.intersection(set2))
###result = sorted(set1.intersection(set2) ,key=lambda x:set1.index(x))
####result = sorted(set1 & set2, key = set1.index)


if len(result) > 0:
    substitution = sentence.split()
    substitution[sentence_list.index(result[-1])] = 'Brussels Sprouts'
    print(f"\n\nI found {len(result)} automobiles in your sentnece")
    print("Your automobiles are", result)
    print("Your final sentence with some brussels spouts is:", ' '.join(substitution))
else:
    print(f"\n\nI found no automobiles in your sentnece")
    print("The automobiles found is an empty list:", result)
    print("Your final sentence is:", sentence)

问题分析 #

当前代码的核心问题是将所有汽车品牌拆分为单个单词存入集合，丢失了多词品牌的上下文关联。当匹配时，只能识别出单个单词，无法还原成原始的多词品牌名称。

修复方案 #

核心思路是优先匹配多词品牌，避免拆分后丢失上下文。具体步骤：

1. 将汽车品牌列表按单词数量降序排序，确保多词品牌被优先匹配
2. 处理输入句子，去除标点并保留可匹配的格式
3. 遍历排序后的品牌，检查是否存在于处理后的句子中，存在则加入结果，并从句子中移除该品牌（防止重复匹配）

修复后的代码 #

import string

autos = [ 'Mazda',  'Toyota', 'Volkswagon', 'Honda', 'Ford',
    'Chevolet', 'Tesla', 'Kia', 'Hyndai', 'Rolls Royce',
    'Isuzu', 'Jeep', 'Land Rover', 'Mitzubishi', 'Subaru'
]
print("Original autos:", autos)

# 按品牌的单词数量降序排序，优先匹配多词品牌
autos_sorted = sorted(autos, key=lambda x: len(x.split()), reverse=True)

# 获取用户输入并处理
sentence = input('Enter a sentence with your favorite automobiles: ')
# 去除标点，保留原始大小写（或统一为标题格式，确保匹配）
processed_sentence = sentence.translate(str.maketrans('', '', string.punctuation))

found_autos = []
temp_sentence = processed_sentence  # 临时句子，用于移除已匹配的品牌

for auto in autos_sorted:
    if auto in temp_sentence:
        found_autos.append(auto)
        # 从临时句子中移除已匹配的品牌，避免重复匹配
        temp_sentence = temp_sentence.replace(auto, '')

# 保持结果在原句子中的出现顺序
def get_order_in_sentence(item):
    return processed_sentence.index(item)

found_autos_sorted = sorted(found_autos, key=get_order_in_sentence)

# 输出结果
if found_autos_sorted:
    print(f"\nI found {len(found_autos_sorted)} automobiles in your sentence")
    print("Your automobiles are:", found_autos_sorted)
    
    # 替换最后一个品牌为 Brussels Sprouts（保留原功能）
    original_words = sentence.split()
    # 找到最后一个品牌在原句子中的位置
    last_auto = found_autos_sorted[-1]
    # 处理原句子中的标点，找到对应单词位置
    for i, word in enumerate(original_words):
        # 去除单词标点后匹配
        clean_word = word.translate(str.maketrans('', '', string.punctuation))
        if last_auto.startswith(clean_word) and last_auto in processed_sentence:
            original_words[i] = 'Brussels Sprouts'
            # 如果是多词品牌，后续单词也替换为空（可选，这里简化处理）
            if len(last_auto.split()) > 1:
                for j in range(i+1, len(original_words)):
                    clean_j = original_words[j].translate(str.maketrans('', '', string.punctuation))
                    if clean_j in last_auto.split():
                        original_words[j] = ''
                original_words = [w for w in original_words if w]
            break
    print("Your final sentence with some brussels sprouts is:", ' '.join(original_words))
else:
    print("\nI found no automobiles in your sentence")
    print("The automobiles found is an empty list:", found_autos_sorted)
    print("Your final sentence is:", sentence)

最终效果 #

输入句子I drive a Rolls Royce, a Ford, and a Land Rover.后，输出：

Original autos: ['Mazda', 'Toyota', 'Volkswagon', 'Honda', 'Ford', 'Chevolet', 'Tesla', 'Kia', 'Hyndai', 'Rolls Royce', 'Isuzu', 'Jeep', 'Land Rover', 'Mitzubishi', 'Subaru']
Enter a sentence with your favorite automobiles: I drive a Rolls Royce, a Ford, and a Land Rover.

I found 3 automobiles in your sentence
Your automobiles are: ['Rolls Royce', 'Ford', 'Land Rover']
Your final sentence with some brussels sprouts is: I drive a Rolls Royce, a Ford, and a Brussels Sprouts.

内容的提问来源于stack exchange，提问作者Scott Bing