app.py

import streamlit as st
from sklearn.metrics.pairwise import pairwise_distances, cosine_similarity
from scipy.spatial import distance
import pandas as pd
import numpy as np
import torch
from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("cointegrated/rubert-tiny2")
model = AutoModel.from_pretrained("cointegrated/rubert-tiny2")

films = pd.read_csv('Films_finder/movies_2.csv')
films['description'] = films['description'].astype(str)

def embed_bert_cls(text, model, tokenizer):
    t = tokenizer(text, padding=True, truncation=True, return_tensors='pt', max_length=1024)
    with torch.no_grad():
        model_output = model(**{k: v.to(model.device) for k, v in t.items()})
    embeddings = model_output.last_hidden_state[:, 0, :]
    embeddings = torch.nn.functional.normalize(embeddings)
    return embeddings[0].cpu().numpy()
@st.cache_resource
def for_embeded_list(series: pd.Series) -> list:
    return np.array([embed_bert_cls(i.replace('\xa0', ' '), model, tokenizer) for i in series])
embeded_list = for_embeded_list(films['description'])
text = st.text_input('Введите текст')
count_visible = st.number_input("Введите количество отображаемых элементов", 1, 10, step=1)
if text and count_visible:
    embeded_text = embed_bert_cls(text, model, tokenizer).reshape(1,-1)
    cossim = pairwise_distances(embeded_text, embeded_list)[0]
    for i in range(count_visible):
        col1, col2 = st.columns(2)
        with col1:
            st.header(films.iloc[cossim.argsort()].iloc[i][2])
            st.write(films.iloc[cossim.argsort()].iloc[i][3].replace('\xa0', ' '))
            st.write(f'Уверенность состовляет {cossim[i]}')
        with col2:
            st.image(films.iloc[cossim.argsort()].iloc[i][1])
    st.header('Самый не подходящий запрос')
    col3, col4 = st.columns(2)
    with col3:
        st.header(films.iloc[cossim.argsort()].iloc[-1][2])
        st.write(films.iloc[cossim.argsort()].iloc[-1][3].replace('\xa0', ' '))
    with col4:
        st.image(films.iloc[cossim.argsort()].iloc[-1][1])