1用python将csv文件中乱码的用户列重新赋值。最后从1开始用数字代替现有的用户id，不删除重复，用相同数字代替相同id 2.统一两种算法的用户列id排列方式，然后对比两种算法对同一首歌的推荐结果。最好能用图表展示出来。 3.将scipy中的svd计算方式封装成一个单独的模块，放到自己的一个py文件中。可以被调用。

1.

import csv

def fix_user_column(csv_file):
    user_dict = {}
    new_user_id = 1
    with open(csv_file, 'r', encoding='utf-8') as file:
        reader = csv.reader(file)
        rows = [row for row in reader]
        for row in rows:
            user = row[0]
            if user not in user_dict:
                user_dict[user] = new_user_id
                new_user_id += 1
            row[0] = user_dict[user]
    with open(csv_file, 'w', encoding='utf-8', newline='') as file:
        writer = csv.writer(file)
        writer.writerows(rows)

fix_user_column('data.csv')

2.

import pandas as pd
import matplotlib.pyplot as plt

def compare_recommendations(algorithm1_file, algorithm2_file, song_id):
    df1 = pd.read_csv(algorithm1_file)
    df2 = pd.read_csv(algorithm2_file)
    df1 = df1[df1['song_id'] == song_id]
    df2 = df2[df2['song_id'] == song_id]
    merged_df = pd.merge(df1, df2, on='user_id')
    merged_df = merged_df[['user_id', 'algorithm1_score', 'algorithm2_score']]
    merged_df.plot(kind='bar', x='user_id')
    plt.show()

compare_recommendations('algorithm1.csv', 'algorithm2.csv', '1234')

3.

from scipy.sparse.linalg import svds

def svd_factorization(matrix, k):
    u, s, vt = svds(matrix, k=k)
    return u, s, vt

该函数可以被调用，前提是导入了 `scipy` 包。