app.py

import gradio as gr
import numpy as np
import pandas as pd
import plotly.express as px
from sklearn.decomposition import PCA
from sentence_transformers import SentenceTransformer

# Load pre-trained sentence transformer model
model = SentenceTransformer('all-MiniLM-L6-v2')

# Function to calculate embeddings and PCA
def compute_pca(texts):
    # Generate embeddings
    embeddings = model.encode(texts)
    
    # Compute PCA
    pca = PCA(n_components=2)
    pca_result = pca.fit_transform(embeddings)
    
    # Create DataFrame for visualization
    df = pd.DataFrame({
        'Text': texts,
        'PC1': pca_result[:, 0],
        'PC2': pca_result[:, 1]
    })
    
    # Plot the PCA result
    fig = px.scatter(df, x='PC1', y='PC2', text='Text', title='PCA of Text Embeddings')
    return fig

# Define Gradio app layout and interactions
def text_editor_app():
    with gr.Blocks() as demo:
        # Text box to input texts
        text_input = gr.Textbox(lines=10, placeholder="Enter or paste your texts here, one per line...", label="Text Inputs")
        
        # Display the list of texts
        texts = gr.Dataframe(headers=["Texts"], label="Text List", interactive=True)
        
        # Button to process texts
        submit_button = gr.Button("Compute Embeddings and PCA")

        # Output plot
        output_plot = gr.Plot(label="PCA Visualization")
        
        # Define button click interaction
        def process_texts(text_input):
            # Split input texts by newline
            text_list = text_input.strip().split('\n')
            return gr.DataFrame.update(value=[[t] for t in text_list], row_count=len(text_list))
        
        submit_button.click(fn=lambda x: compute_pca([t[0] for t in x]), inputs=texts, outputs=output_plot)
        text_input.change(fn=process_texts, inputs=text_input, outputs=texts)

    return demo

# Launch the app
text_editor_app().launch()