app.py

import gradio as gr
from collections import Counter
import string
import csv
import io

def process_text(text, sorting_option):
    # Remove punctuation from the input text
    translator = str.maketrans('', '', string.punctuation)
    text = text.translate(translator)
    
    words = text.split()
    
    # Count word frequencies
    word_counts = Counter(words)
    
    # Sort words based on the selected option
    if sorting_option == 'alphabetically':
        sorted_words = sorted(word_counts.items(), key=lambda x: x[0])
    elif sorting_option == 'by_frequency':
        sorted_words = sorted(word_counts.items(), key=lambda x: x[1], reverse=True)
    else:
        sorted_words = word_counts.items()

    # Get the top 5 words with their frequencies
    top_5_words = sorted_words
    
    # Format the top 5 words and frequencies as an HTML table
    table_html = "<table style='width:100%'><tr><th>Index</th><th>Word</th><th>Frequency</th></tr>"
    for index, (word, freq) in enumerate(top_5_words, start=1):
        table_html += f"<tr><td>{index}</td><td>{word}</td><td>{freq}</td></tr>"
    table_html += "</table>"
    
    # Create a CSV file
    csv_data = [["Index", "Word", "Frequency"]]
    for index, (word, freq) in enumerate(top_5_words, start=1):
        csv_data.append([index, word, freq])
    
    # Write CSV data to a string buffer
    csv_buffer = io.StringIO()
    csv_writer = csv.writer(csv_buffer)
    csv_writer.writerows(csv_data)
    csv_buffer.seek(0)
    
    # Create a download link for the CSV file
    csv_download_link = f"<a href='data:application/csv;charset=utf-8,{csv_buffer.getvalue()}' download='word_frequencies.csv'>Download CSV</a>"
    
    # Wrap the table in a div with a fixed height and scrolling
    div_with_scroll = f"<div style='height: 200px; overflow-y: scroll;'>{table_html}</div>"
    
    return div_with_scroll, csv_download_link

iface = gr.Interface(
    fn=process_text,
    inputs=[gr.Textbox("text", label="Paste Text Here"), 
            gr.Radio(["alphabetically", "by_frequency", "none"], label="Select Sorting Option")],
    outputs=[gr.HTML(label="Top 5 Words with Frequencies"), gr.HTML(label="Download CSV")]
)

iface.launch(share=True)