import gradio as gr
from transformers import pipeline
import os
import neattext.functions as nfx
import re

model = pipeline("text-classification", model="i0xs0/Emotion_Detection", tokenizer="i0xs0/Emotion_Detection")

def clean_text(text):

    if not isinstance(text, str):
            return text

    text = nfx.remove_userhandles(text)        # Remove user handles (@username)
    text = nfx.remove_punctuations(text)       # Remove punctuation marks (!, ?, .)
    text = nfx.remove_accents(text)            # Remove accents from characters (e.g., é -> e)
    text = nfx.remove_urls(text)               # Remove URLs (e.g., https://example.com)
    text = nfx.remove_emojis(text)             # Remove emojis (e.g., 😊, 🚀)
    text = nfx.remove_emails(text)             # Remove email addresses (e.g., user@example.com)
    text = nfx.remove_phone_numbers(text)      # Remove phone numbers (e.g., +1234567890)
    text = nfx.remove_html_tags(text)          # Remove HTML tags (<div>, <p>)
    text=re.sub(r"[^a-zA-Z0-9\s']", "", text)  # Remove special characters
    text = nfx.remove_multiple_spaces(text)    # Remove multiple spaces and reduce them to a single space
    text = nfx.remove_md5sha(text)             # Remove MD5 or SHA-like hash strings

    return text
    
def predict_emotion(text):

    cleaned_text = clean_text(text)  
    #print("Processed Text:", cleaned_text)  

    results = model(cleaned_text)  
    return {item["label"]: item["score"] for item in results}


#theme = gr.themes.Ocean()
#theme = gr.themes.Glass()
theme = gr.themes.Soft()



demo = gr.Interface(
    fn=predict_emotion,                
    inputs=gr.Textbox(label="Input Text"),  
    outputs=gr.Label(label="Emotion"),
    title="Emotion Classifier",
    description="Enter a text to classify its emotion.",
    allow_flagging="never",  

    theme=theme  
)


demo.launch()