import torch
import gradio as gr
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns


# Use a pipeline as a high-level helper
from transformers import pipeline

# model_path = ("../Models/models--distilbert--distilbert-base-uncased-finetuned-sst-2-english/snapshots/714eb0fa89d2f80546fda750413ed43d93601a13")
# analyzer = pipeline("text-classification", model=model_path)

analyzer = pipeline("text-classification", model="distilbert/distilbert-base-uncased-finetuned-sst-2-english")

# print(analyzer(["This is good", "This product is quite expensive"]))

def sentiment_analyzer(review):
    sentiment = analyzer(review)
    return sentiment[0]['label'], sentiment[0]['score']  # Return sentiment label and confidence score

def create_sentiment_pie_chart(df):
    sentiment_counts = df['Sentiment'].value_counts()

    # Style the pie chart
    colors = ['#e67e22', '#3498db']  # Green for Positive, Red for Negative
    explode = [0.05, 0]  # Slightly separate the positive slice for emphasis

    fig, ax = plt.subplots(figsize=(6, 6))
    sentiment_counts.plot(
        kind='pie',
        colors=colors,
        autopct='%1.1f%%',
        startangle=140,
        explode=explode,
        ax=ax,
        textprops={'fontsize': 10, 'color': 'black'},
        radius=0.7 # Decrease radius to make the pie smaller
    )
    ax.set_title("Sentiment Distribution (Pie Chart)", fontsize=10, fontweight='bold', color='#34495e')
    ax.set_ylabel("")  # Remove default 'Count' label
    plt.tight_layout()

    return fig

# Create a styled bar chart
def create_sentiment_bar_chart(df):
    sentiment_counts = df['Sentiment'].value_counts()

    # Use Seaborn for stylish bar plots
    fig, ax = plt.subplots(figsize=(8, 5))
    # Use a categorical variable for hue to avoid the warning
    sns.barplot(
        x=sentiment_counts.index,
        y=sentiment_counts.values,
        # for hue warning which doesn't effect code
        hue=sentiment_counts.index,  # Use 'Sentiment' as the hue variable
        palette=['#e67e22', '#3498db'],  # Green for Positive, Red for Negative
        ax=ax,
        legend=False,  # We don't need a legend for this plot
        width = 0.5
    )

    # Add value annotations on top of each bar
    for idx, value in enumerate(sentiment_counts.values):
        ax.text(idx, value + 0.5, str(value), ha='center', fontsize=10, color='black', fontweight='bold')

    ax.set_title("Sentiment Distribution (Bar Chart)", fontsize=12, fontweight='bold', color='#34495e')
    ax.set_xlabel("Sentiment", fontsize=10, fontweight='bold', color='#34495e')
    ax.set_ylabel("Count", fontsize=10, fontweight='bold', color='#34495e')
    ax.spines['top'].set_visible(False)  # Clean up borders
    ax.spines['right'].set_visible(False)

    plt.xticks(fontsize=10, fontweight='bold', color='#34495e')
    plt.yticks(fontsize=10, color='#34495e')
    plt.tight_layout()

    return fig


# Define the main function to process the Excel file
def read_reviews_and_analyze_sentiment(file_object):
    # Read the Excel file into a DataFrame
    df = pd.read_excel(file_object)

    # Ensure the column containing reviews exists
    if 'Reviews' not in df.columns:
        raise ValueError("The Excel file must contain a column named 'Reviews'.")

    # Apply the sentiment analyzer and split into two columns
    df[['Sentiment', 'Confidence']] = df['Reviews'].apply(
        lambda x: pd.Series(sentiment_analyzer(x))
    )

    # Generate the pie and bar charts
    pie_chart = create_sentiment_pie_chart(df)
    bar_chart = create_sentiment_bar_chart(df)

    return df, pie_chart, bar_chart



# Example usage:
# df = process_reviews("reviews.xlsx")
# print(df)

# result = read_reviews_and_analyze_sentiment("../Files/Updated_Reviews_Sample.xlsx")
# print(result)

import gradio as gr

# Custom CSS for styling the title
custom_css = """
.title { font-size: 24px; font-weight: bold; color: #34495e; }
.gradio-label, .gradio-description {
    font-size: 18px;  /* Increase font size */
    font-weight: bold;  /* Make text bold */
}
"""

# Interface
interface = gr.Interface(
    fn=read_reviews_and_analyze_sentiment,
    inputs=gr.File(file_types=[".xlsx"], label="Upload Excel File"),  # Accept only Excel files
    outputs=[
        gr.Dataframe(
            headers=["Reviews", "Sentiment", "Confidence"],  # Custom headers for output
            label="Sentiment Analysis Results"
        ),
        gr.Plot(label="Sentiment Distribution (Pie Chart)"),
        gr.Plot(label="Sentiment Distribution (Bar Chart)")
    ],
    title="Sentiment Analyzer with Confidence Scores",
    description=(
        "Upload an Excel file with a 'Reviews' column to analyze sentiments and "
        "visualize the distribution in both pie and bar charts."
    ),
    theme="default",  # Align components vertically
    css=custom_css  # Apply the custom CSS
)

interface.launch()