tabs/sentence_similarity_tab.py

import streamlit as st # type: ignore
from PIL import Image
import os
import ast
import contextlib
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from nltk.corpus import stopwords
from sklearn.manifold import TSNE
from sentence_transformers import SentenceTransformer
from sklearn.metrics.pairwise import cosine_similarity
from translate_app import tr

title = "Sentence Similarity"
sidebar_name = "Sentence Similarity"
dataPath = st.session_state.DataPath


def run():
    st.write("")
    st.title(tr(title))
    sentences = ["This is an example sentence", "Each sentence is converted"]
    sentences[0] = st.text_area(label=tr("Saisir un élément issu de la proposition de valeur (quelque soit la langue):"), value="This is an example sentence")
    sentences[1] = st.text_area(label=tr("Saisir une phrase issue de l'acte de vente (quelque soit la langue):"), value="Each sentence is converted", height=200)
    st.button(label=tr("Validez"), type="primary")

    model = SentenceTransformer('sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2')
    embeddings = model.encode(sentences)
    st.write(tr("Transformation de chaque phrase en vecteur (dimension = 384 ):"))
    st.write(embeddings)
    st.write("")
    # Calculate cosine similarity between the two sentences
    similarity = cosine_similarity([embeddings[0]], [embeddings[1]])
    st.write(tr("**Cosine similarity** comprise entre 0 et 1 :"), similarity[0][0])
    st.write("")
    st.write("")
    st.write("")