app.py

import gradio as gr
import pandas as pd
import faiss
import numpy as np
import os
from FlagEmbedding import BGEM3FlagModel
from io import BytesIO

# Load the pre-trained embedding model
model = BGEM3FlagModel('BAAI/bge-m3', use_fp16=True)

# Load the JSON data into a DataFrame
df = pd.read_json('White-Stride-Red-68.json')
df['embeding_context'] = df['embeding_context'].astype(str).fillna('')

# Filter out any rows where 'embeding_context' might be empty or invalid
df = df[df['embeding_context'] != '']

# Load the FAISS index
index = faiss.read_index('vector_store_bge_m3.index')

# Function to perform search and return all columns
def search_query(query_text):
    num_records = 50

    # Encode the input query text
    embeddings_query = model.encode([query_text], batch_size=12, max_length=1024)['dense_vecs']
    embeddings_query_np = np.array(embeddings_query).astype('float32')

    # Search in FAISS index for nearest neighbors
    distances, indices = index.search(embeddings_query_np, num_records)

    # Get the top results based on FAISS indices
    result_df = df.iloc[indices[0]].drop(columns=['embeding_context']).drop_duplicates().reset_index(drop=True)

    return result_df

# Gradio interface function
def gradio_interface(query_text):
    search_results = search_query(query_text)
    
    # Save search_results to an Excel file in memory
    output = BytesIO()
    with pd.ExcelWriter(output, engine='xlsxwriter') as writer:
        search_results.to_excel(writer, index=False)
    excel_data = output.getvalue()
    
    # Return the DataFrame and update the download button
    return search_results, gr.update(value=excel_data)

with gr.Blocks() as app:
    gr.Markdown("<h1>White Stride Red Search (BEG-M3)</h1>")

    # Input text box for the search query
    search_input = gr.Textbox(label="Search Query", placeholder="Enter search text", interactive=True)

    # Search button below the text box
    search_button = gr.Button("Search")

    # Output table for displaying results
    search_output = gr.DataFrame(label="Search Results")

    # Download button for Excel file
    download_button = gr.DownloadButton(label="Download Excel", file_name="search_results.xlsx")

    # Link button click to action
    search_button.click(fn=gradio_interface, inputs=search_input, outputs=[search_output, download_button])

# Launch the Gradio app
app.launch()