| _C='rating' | |
| _B='userId' | |
| _A='movieId' | |
| import gradio as gr,numpy as np,pandas as pd | |
| from scipy.sparse import csr_matrix | |
| from sklearn.neighbors import NearestNeighbors | |
| def create_matrix(df):A=df;B=len(A[_B].unique());C=len(A[_A].unique());D=dict(zip(np.unique(A[_B]),list(range(B))));E=dict(zip(np.unique(A[_A]),list(range(C))));F=dict(zip(list(range(B)),np.unique(A[_B])));G=dict(zip(list(range(C)),np.unique(A[_A])));H=[D[A]for A in A[_B]];I=[E[A]for A in A[_A]];J=csr_matrix((A[_C],(I,H)),shape=(C,B));return J,D,E,F,G | |
| def find_similar_movies(movie_id,X,k,metric='cosine',show_distance=False): | |
| A=[];D=movie_mapper[movie_id];B=X[D];k+=1;C=NearestNeighbors(n_neighbors=k,algorithm='brute',metric=metric);C.fit(X);B=B.reshape(1,-1);E=C.kneighbors(B,return_distance=show_distance) | |
| for F in range(0,k):G=E.item(F);A.append(movie_inv_mapper[G]) | |
| A.pop(0);return A | |
| def recommend_movies(movie_name): | |
| A=[A for(A,B)in movie_titles.items()if movie_name.lower()in B.lower()] | |
| if len(A)==0:return'Movie not found. Please check the spelling and try again' | |
| A=A[0];B=find_similar_movies(A,X,k=10);C='\n'.join([movie_titles[A]for A in B]);return C | |
| ratings=pd.read_csv('ratings.csv') | |
| movies=pd.read_csv('movies.csv') | |
| n_ratings=len(ratings) | |
| n_movies=len(ratings[_A].unique()) | |
| n_users=len(ratings[_B].unique()) | |
| user_freq=ratings[[_B,_A]].groupby(_B).count().reset_index() | |
| user_freq.columns=[_B,'n_ratings'] | |
| mean_rating=ratings.groupby(_A)[[_C]].mean() | |
| lowest_rated=mean_rating[_C].idxmin() | |
| highest_rated=mean_rating[_C].idxmax() | |
| movie_stats=ratings.groupby(_A)[[_C]].agg(['count','mean']) | |
| movie_stats.columns=movie_stats.columns.droplevel() | |
| X,user_mapper,movie_mapper,user_inv_mapper,movie_inv_mapper=create_matrix(ratings) | |
| movie_titles=dict(zip(movies[_A],movies['title'])) | |
| movie_name=gr.inputs.Textbox(label='Movie Name') | |
| outputs=gr.outputs.Textbox(label='Recommended Movies',type='text') | |
| iface=gr.Interface(fn=recommend_movies,inputs=movie_name,outputs=outputs,theme=gr.themes.Default(primary_hue='slate')) | |
| iface.launch() |