详解如何使用Python从零开始构建文本统计模型

with open("文档1.md", "r", encoding="utf-8") as f:
    lines = f.readlines()

# 单字符统计
vocab_1 = Counter()
for line in lines:
    vocab_1.update(Counter(list(line)))

# 多字符片段统计（2-5字符）
def multi_char_counter(n):
    vocab = Counter()
    for line in lines:
        for j in range(0, len(line), n):
            segment = line[j:j + n]
            if len(segment) == n:
                vocab.update(Counter([segment]))
    return vocab

vocab_2 = multi_char_counter(2)
vocab_3 = multi_char_counter(3)
vocab_4 = multi_char_counter(4)
vocab_5 = multi_char_counter(5)

# 转换为DataFrame并排序
def create_vocab_df(counter, top_n=None):
    df = pd.DataFrame(counter.most_common(top_n), columns=["word", "count"])
    return df.sort_values(by="count", ascending=False)

vocab_dfs = {
    f"vocab_{i}_df": create_vocab_df(globals()[f"vocab_{i}"], 50)
    for i in range(2, 6)
}

# 调整单字符计数
for df, n in [(vocab_5_df, 5), (vocab_4_df, 4), (vocab_3_df, 3), (vocab_2_df, 2)]:
    for word, count in df[["word", "count"]].values:
        if count > 1:
            for i in range(n):
                char = word[i:i + 1]
                if char in vocab_1_df['word'].values:
                    vocab_1_df.loc[vocab_1_df['word'] == char, 'count'] -= count