index.py

import transformers
from flask import Flask, request, jsonify
from transformers import RobertaTokenizerFast, TFRobertaForSequenceClassification, pipeline
import gradio as gr
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import io
from io import BytesIO  # Import BytesIO for image generation

# Load model and tokenizer
tokenizer = RobertaTokenizerFast.from_pretrained("arpanghoshal/EmoRoBERTa")
model = TFRobertaForSequenceClassification.from_pretrained("arpanghoshal/EmoROBERTa")
emotion = pipeline("sentiment-analysis", model="arpanghoshal/EmoROBERTa")

def analyze_csv(file):
    try:
        # Print file content for debugging
        file_content = file.read()
        print("File content:", file_content)

        # Reset file position to the beginning
        file.seek(0)

        # Read the CSV file into a DataFrame
        df = pd.read_csv(io.BytesIO(file_content))``
        print("DataFrame shape:", df.shape)  # Print DataFrame shape for debugging
        print("DataFrame columns:", df.columns)  # Print DataFrame columns for debugging

        # Check if the DataFrame is empty
        if df.empty:
            return "Empty file. Please upload a CSV file with data.", None

        # Check if the expected column "phrase" is present in the DataFrame
        if "phrase" not in df.columns:
            return "Column 'phrase' not found in the CSV file. Please check the file format.", None

        phrases = df["phrase"]

        # Analyze sentiment for each phrase
        emotion_labels = emotion(phrases)

        # Create summary statistics
        summary_df = pd.DataFrame(emotion_labels).describe()

        # Create a bar chart of emotion distribution
        plt.figure()
        emotion_counts = emotion_labels.get("labels").value_counts()
        emotion_counts.plot(kind="bar")
        plt.title("Emotion Distribution")
        plt.xlabel("Emotion")
        plt.ylabel("Count")

        # Generate PNG image of the chart
        chart_img = BytesIO()
        plt.savefig(chart_img, format="png")
        chart_img.seek(0)

        return summary_df.to_json(), chart_img.read()

    except Exception as e:
        error_message = f"Error processing the CSV file: {str(e)}"
        print(error_message)  # Print the error message for debugging
        return error_message, None


iface = gr.Interface(
    fn=analyze_csv,
    inputs=[gr.File(label="Upload CSV File")],
    outputs=["dataframe", "image"],
    title="Emotion Analyzer with CSV",
    description="Analyzes sentiment and creates charts/tables from a CSV file.",
)

iface.launch()