import gradio as gr
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

def plot_likert_scores(file):
    # Load the Excel file into pandas
    df = pd.read_excel(file.name, sheet_name=0, header=None)
    
    # Extract pre and post survey blocks (edit row indices if needed)
    pre_df = df.iloc[1:5, :6].copy()
    post_df = df.iloc[12:15, :6].copy()

    pre_df.columns = ['identifier', 'Degree', 'Confidence', 'Feedback', 'Preparedness', 'Enjoyment']
    post_df.columns = ['identifier', 'Degree', 'Confidence', 'Feedback', 'Preparedness', 'Enjoyment']

    # Clean identifier
    def clean_identifier(x):
        try:
            return int(str(x).strip())
        except:
            return np.nan

    pre_df['identifier'] = pre_df['identifier'].apply(clean_identifier)
    post_df['identifier'] = post_df['identifier'].apply(clean_identifier)
    pre_df = pre_df.dropna(subset=['identifier'])
    post_df = post_df.dropna(subset=['identifier'])
    pre_df['identifier'] = pre_df['identifier'].astype(int)
    post_df['identifier'] = post_df['identifier'].astype(int)

    # Filter for identifier 7
    pre_7 = pre_df[pre_df['identifier'] == 7]
    post_7 = post_df[post_df['identifier'] == 7]

    # If data for identifier 7 is missing
    if pre_7.empty or post_7.empty:
        return "Identifier 7 is missing from pre or post survey data. Please check your file."

    # Map Likert values
    likert_map = {
        'Strongly Disagree': 1,
        'Disagree': 2,
        'Agree': 3,
        'Strongly Agree': 4
    }
    questions = ['Confidence', 'Feedback', 'Enjoyment']
    pretty_labels = ['Confidence', 'Feedback', 'Enjoyment']

    pre_scores = [likert_map.get(pre_7[q].values[0], np.nan) for q in questions]
    post_scores = [likert_map.get(post_7[q].values[0], np.nan) for q in questions]

    # Plot
    x = np.arange(len(questions))
    width = 0.35

    fig, ax = plt.subplots(figsize=(7, 5))
    bars1 = ax.bar(x - width/2, pre_scores, width, label='Pre-survey', color='#3182bd', edgecolor='black')
    bars2 = ax.bar(x + width/2, post_scores, width, label='Post-survey', color='#fdae6b', edgecolor='black')

    # Add value labels
    for bar in bars1 + bars2:
        height = bar.get_height()
        if not np.isnan(height):
            ax.annotate(f'{int(height)}',
                        xy=(bar.get_x() + bar.get_width() / 2, height),
                        xytext=(0, 5),
                        textcoords="offset points",
                        ha='center', va='bottom', fontsize=12, fontweight='bold')

    ax.set_xticks(x)
    ax.set_xticklabels(pretty_labels, fontsize=13, fontweight='bold')
    ax.set_ylim(0.5, 4.5)
    ax.set_yticks([1, 2, 3, 4])
    ax.set_yticklabels(['Strongly Disagree', 'Disagree', 'Agree', 'Strongly Agree'], fontsize=12)
    ax.set_ylabel('Likert Score', fontsize=13, fontweight='bold')
    ax.set_title('Identifier 7: Pre vs Post Likert Scores', fontsize=15, fontweight='bold')
    ax.legend(fontsize=12)
    ax.grid(axis='y', linestyle='--', alpha=0.7)

    plt.tight_layout()

    return fig

# Gradio interface
demo = gr.Interface(
    fn=plot_likert_scores,
    inputs=gr.File(label="Upload your Excel file (.xlsx)"),
    outputs=gr.Plot(label="Likert Plot for Identifier 7"),
    title="Pre vs Post Likert Plot (Identifier 7)",
    description="Upload your survey Excel file. This tool compares pre/post Likert scores for Confidence, Feedback, and Enjoyment for respondent 7.",
    allow_flagging='never'
)

if __name__ == "__main__":
    demo.launch()