协同过滤推荐算法第一版实现实验：课程推荐功能开发

import pandas as pd
import numpy as np
from sklearn.metrics.pairwise import cosine_similarity

读取用户-资源评分数据（学习时长标准化为0-5分）
user_item_matrix = pd.read_csv('user_item_rating.csv', index_col='user_id')

填充缺失值（未学习资源评分为0）
user_item_matrix = user_item_matrix.fillna(0)

计算用户间余弦相似度
user_similarity = cosine_similarity(user_item_matrix)
user_similarity_df = pd.DataFrame(
  user_similarity,
  index=user_item_matrix.index,
  columns=user_item_matrix.index
)

def get_user_recommendations(user_id, top_n=5):
    # 1. 获取目标用户的相似用户（相似度Top 20）
    similar_users = user_similarity_df[user_id].sort_values(ascending=False)[1:21].index
    
    # 2. 获取相似用户已学习但目标用户未学习的资源
    user_learned = user_item_matrix.loc[user_id][user_item_matrix.loc[user_id] > 0].index
    similar_users_items = user_item_matrix.loc[similar_users]
    
    # 3. 计算资源推荐得分（加权平均相似度）
    item_scores = {}    
    for item in similar_users_items.columns:
        if item not in user_learned:
            # 获取评分该资源的相似用户
            users_who_rated = similar_users_items[item][similar_users_items[item] > 0].index
            if len(users_who_rated) > 0:
                # 加权得分 = Σ(用户相似度 * 用户评分) / Σ(用户相似度)
                weighted_score = (user_similarity_df[user_id][users_who_rated] * 
                                 similar_users_items[item][users_who_rated]).sum()
                total_similarity = user_similarity_df[user_id][users_who_rated].sum()
                item_scores[item] = weighted_score / total_similarity
    
    # 4. 返回Top N推荐资源
    sorted_items = sorted(item_scores.items(), key=lambda x: x[1], reverse=True)[:top_n]
    return [item[0] for item in sorted_items]

from fastapi import FastAPI
from pydantic import BaseModel

app = FastAPI(title="学习资源推荐API")

class RecommendRequest(BaseModel):
    user_id: int
    top_n: int = 5

@app.post("/recommend/courses")
def recommend_courses(request: RecommendRequest):
    try:
        recommended_item_ids = get_user_recommendations(
            request.user_id, 
            request.top_n
        )
        
        # 关联资源详情（从数据库获取）
        recommended_items = get_item_details(recommended_item_ids)
        
        return {
            "code": 200,
            "message": "success",
            "data": recommended_items
        }
    except Exception as e:
        return {
            "code": 500,
            "message": f"推荐失败：{str(e)}",
            "data": []
        }