Levenshtein.ratio() = 1 - (distance / max(len(s1), len(s2)))

import Levenshtein
 
def levenshtein_demo():
    text1 = "abcd"
    text2 = "aBCD"
    
    distance = Levenshtein.distance(text1, text2)
    ratio = Levenshtein.ratio(text1, text2)
    
    print(f"編輯距離: {distance}")  
    print(f"相似度比率: {ratio:.4f}")  
 
levenshtein_demo()

def jaccard_similarity(s1, s2):
    """
    計(jì)算兩個(gè)字符串的 Jaccard 相似度
    """
    set1 = set(s1.lower())
    set2 = set(s2.lower())
    intersection = len(set1.intersection(set2))
    union = len(set1.union(set2))
    return intersection / union if union != 0 else 0
 
# 示例
text1 = "hello world"
text2 = "hello python"
print(f"Jaccard 相似度: {jaccard_similarity(text1, text2):.4f}")

from collections import Counter
import math
 
def cosine_similarity(s1, s2):
    """
    基于字符頻率的余弦相似度計(jì)算
    """
    # 創(chuàng)建字符頻率向量
    vec1 = Counter(s1.lower())
    vec2 = Counter(s2.lower())
    
    # 計(jì)算點(diǎn)積
    intersection = set(vec1.keys()) & set(vec2.keys())
    dot_product = sum(vec1[x] * vec2[x] for x in intersection)
    
    # 計(jì)算向量模長
    magnitude1 = math.sqrt(sum(v**2 for v in vec1.values()))
    magnitude2 = math.sqrt(sum(v**2 for v in vec2.values()))
    
    if magnitude1 == 0 or magnitude2 == 0:
        return 0
    return dot_product / (magnitude1 * magnitude2)
 
# 示例
text1 = "machine learning"
text2 = "deep learning"
print(f"余弦相似度: {cosine_similarity(text1, text2):.4f}")

def hamming_distance(s1, s2):
    """
    計(jì)算漢明距離
    """
    if len(s1) != len(s2):
        return max(len(s1), len(s2))
    return sum(c1 != c2 for c1, c2 in zip(s1, s2))
 
def hamming_similarity(s1, s2):
    """
    計(jì)算漢明相似度
    """
    if len(s1) != len(s2):
        return 0
    max_len = len(s1)
    distance = hamming_distance(s1, s2)
    return 1 - (distance / max_len)
 
# 示例
binary1 = "1011101"
binary2 = "1001001"
print(f"漢明相似度: {hamming_similarity(binary1, binary2):.4f}")

def get_bigrams(s):
    """
    獲取字符串的二元語法（bigram）
    """
    return set(s[i:i+2] for i in range(len(s)-1))
 
def dice_coefficient(s1, s2):
    """
    計(jì)算 Dice 系數(shù)
    """
    bigrams1 = get_bigrams(s1.lower())
    bigrams2 = get_bigrams(s2.lower())
    
    intersection = len(bigrams1.intersection(bigrams2))
    return 2 * intersection / (len(bigrams1) + len(bigrams2)) if (len(bigrams1) + len(bigrams2)) > 0 else 0
 
# 示例
text1 = "night"
text2 = "nacht"
print(f"Dice 系數(shù): {dice_coefficient(text1, text2):.4f}")

import difflib
 
def sequence_matcher_similarity(s1, s2):
    """
    使用 difflib 計(jì)算相似度
    """
    return difflib.SequenceMatcher(None, s1, s2).ratio()
 
# 示例
text1 = "quick brown fox"
text2 = "quick brown cat"
print(f"difflib 相似度: {sequence_matcher_similarity(text1, text2):.4f}")

pip install fuzzywuzzy

from fuzzywuzzy import fuzz
 
def fuzzy_similarity_demo():
    text1 = "this is a test"
    text2 = "this is a test!"
    
    print(f"簡單比率: {fuzz.ratio(text1, text2)}")
    print(f"部分匹配: {fuzz.partial_ratio(text1, text2)}")
    print(f"詞序不敏感: {fuzz.token_sort_ratio(text1, text2)}")
    print(f"集合比率: {fuzz.token_set_ratio(text1, text2)}")
 
fuzzy_similarity_demo()

import Levenshtein
import difflib
from fuzzywuzzy import fuzz
 
def comprehensive_similarity(text1, text2):
    """
    綜合多種方法計(jì)算相似度
    """
    results = {
        'Levenshtein': Levenshtein.ratio(text1, text2),
        'difflib': difflib.SequenceMatcher(None, text1, text2).ratio(),
        'fuzzy_ratio': fuzz.ratio(text1, text2) / 100,
        'partial_ratio': fuzz.partial_ratio(text1, text2) / 100
    }
    
    print(f"文本1: {text1}")
    print(f"文本2: {text2}")
    print("-" * 30)
    for method, score in results.items():
        print(f"{method:15}: {score:.4f}")
    print()
 
# 測試不同場景
comprehensive_similarity("Hello World", "Hello World!")
comprehensive_similarity("quick brown fox", "fast brown fox")
comprehensive_similarity("machine learning", "deep learning")