什么是基于内容的过滤？构建推荐系统指南

import pandas as pd
import ast
import os
import pickle
import requests

# Replace values below with your own if using Redis Cloud instance
REDIS_HOST = os.getenv("REDIS_HOST", "localhost")
REDIS_PORT = os.getenv("REDIS_PORT", "6379")
REDIS_PASSWORD = os.getenv("REDIS_PASSWORD", "")

# If SSL is enabled on the endpoint, use rediss:// as the URL prefix
REDIS_URL = f"redis://:{REDIS_PASSWORD}@{REDIS_HOST}:{REDIS_PORT}"

try:
   df = pd.read_csv("datasets/content_filtering/25k_imdb_movie_dataset.csv")
except:
   # download the file
   url = 'https://redis-ai-resources.s3.us-east-2.amazonaws.com/recommenders/datasets/content-filtering/25k_imdb_movie_dataset.csv'
   r = requests.get(url)

   #save the file as a csv
   if not os.path.exists('./datasets/content_filtering'):
       os.makedirs('./datasets/content_filtering')
   with open('./datasets/content_filtering/25k_imdb_movie_dataset.csv', 'wb') as f:
       f.write(r.content)
   df = pd.read_csv("datasets/content_filtering/25k_imdb_movie_dataset.csv")

roman_numerals = ['(I)','(II)','(III)','(IV)', '(V)', '(VI)', '(VII)', '(VIII)', '(IX)', '(XI)', '(XII)', '(XVI)', '(XIV)', '(XXXIII)', '(XVIII)', '(XIX)', '(XXVII)']

def replace_year(x):
   if x in roman_numerals:
       return 1998 # the average year of the dataset
   else:
       return x

df.drop(columns=['runtime', 'writer', 'path'], inplace=True)

df['year'] = df['year'].apply(replace_year)             # replace roman numerals with average year
df['genres'] = df['genres'].apply(ast.literal_eval)     # convert string representation of list to list
df['keywords'] = df['keywords'].apply(ast.literal_eval) # convert string representation of list to list
df['cast'] = df['cast'].apply(ast.literal_eval)         # convert string representation of list to list
df = df[~df['overview'].isnull()]                       # drop rows with missing overviews
df = df[~df['overview'].isin(['none'])]                 # drop rows with 'none' as the overview

# add a column to the dataframe with all the text we want to embed
df["full_text"] = df["title"] + ". " + df["overview"] + " " + df['keywords'].apply(lambda x: ', '.join(x))

from redisvl.utils.vectorize import HFTextVectorizer

vectorizer = HFTextVectorizer(model = 'sentence-transformers/paraphrase-MiniLM-L6-v2')

df['embedding'] = df['full_text'].apply(lambda x: vectorizer.embed(x, as_buffer=False))
pickle.dump(df['embedding'], open('datasets/content_filtering/text_embeddings.pkl', 'wb'))

try:
    with open('datasets/content_filtering/text_embeddings.pkl', 'rb') as vector_file:
        df['embedding'] = pickle.load(vector_file)
except:
    embeddings_url = 'https://redis-ai-resources.s3.us-east-2.amazonaws.com/recommenders/datasets/content-filtering/text_embeddings.pkl'
    r = requests.get(embeddings_url)
    with open('./datasets/content_filtering/text_embeddings.pkl', 'wb') as f:
        f.write(r.content)
    with open('datasets/content_filtering/text_embeddings.pkl', 'rb') as vector_file:
        df['embedding'] = pickle.load(vector_file)

index:
    name: movies_recommendation
    prefix: movie
    storage_type: json

fields:
    - name: title
      type: text
    - name: rating
      type: numeric
    - name: rating_count
      type: numeric
    - name: genres
      type: tag
    - name: overview
      type: text
    - name: keywords
      type: tag
    - name: cast
      type: tag
    - name: writer
      type: text
    - name: year
      type: numeric
    - name: full_text
      type: text

    - name: embedding
      type: vector
      attrs:
          dims: 384
          distance_metric: cosine
          algorithm: flat
          dtype: float32

movie_schema = IndexSchema.from_yaml("content_filtering_schema.yaml")
index = SearchIndex(movie_schema, redis_client=client)
index.create(overwrite=True, drop=True)

data = df.to_dict(orient='records')
keys = index.load(data)

from redisvl.query import RangeQuery

query_vector = df[df['title'] == '20,000 Leagues Under the Sea']['embedding'].values[0]

query = RangeQuery(vector=query_vector,
                    vector_field_name='embedding',
                    num_results=3,
                    distance_threshold=0.7,
                    return_fields = ['title', 'overview', 'vector_distance'])

results = index.query(query)

[{'id': 'movie:b64fc099d6af440a891e1dd8314e5af7',
 'vector_distance': '0.584870040417',
 'title': 'The Odyssey',
 'overview': 'The aquatic adventure of the highly influential and fearlessly ambitious pioneer, innovator, filmmaker, researcher, and conservationist, Jacques-Yves Cousteau, covers roughly thirty years of an inarguably rich in achievements life.'},

{'id': 'movie:2fbd7803b51a4bf9a8fb1aa79244ad64',
 'vector_distance': '0.63329231739',
 'title': 'The Inventor',
 'overview': 'Inventing flying contraptions, war machines and studying cadavers, Leonardo da Vinci tackles the meaning of life itself with the help of French princess Marguerite de Nevarre.'},

{'id': 'movie:224a785ca7ea4006bbcdac8aad5bf1bc',
 'vector_distance': '0.658123672009',
 'title': 'Ruin',
 'overview': 'The film follows a nameless ex-Nazi captain who navigates the ruins of post-WWII Germany determined to atone for his crimes during the war by hunting down the surviving members of his former SS Death Squad.'}]

from redisvl.query.filter import Tag, Num, Text

def make_filter(genres=None, release_year=None, keywords=None):
    flexible_filter = (
        (Num("year") > release_year) &  # only show movies released after this year
        (Tag("genres") == genres) &     # only show movies that match at least one in list of genres
        (Text("full_text") % keywords)  # only show movies that contain at least one of the keywords
    )
    return flexible_filter

def get_recommendations(movie_vector, num_results=5, distance=0.6, filter=None):
    query = RangeQuery(vector=movie_vector,
                        vector_field_name='embedding',
                        num_results=num_results,
                        distance_threshold=distance,
                        return_fields = ['title', 'overview', 'genres'],
                        filter_expression=filter,
                        )

    recommendations = index.query(query)
    return recommendations