datasets_page.py

import streamlit as st
import pandas as pd
import numpy as np
import plotly.express as px
import plotly.graph_objects as go

def show():
    st.markdown('<div class="main-header">📁 Dataset: MM-OR</div>', unsafe_allow_html=True)
    
    st.markdown("---")
    
    st.markdown("## 🗂️ MM-OR: A Large-scale Multimodal Operating Room Dataset")
    st.markdown("""
    This project utilizes the **MM-OR** dataset, a comprehensive collection of data recorded in a realistic operating room environment. 
    It is designed to support research in surgical workflow analysis, human activity recognition, and context-aware systems in healthcare.
    """)
    
    # Dataset overview
    st.markdown("### 📊 Dataset High-Level Statistics")
    
    col1, col2, col3, col4 = st.columns(4)
    
    with col1:
        st.metric(
            label="📹 Surgical Procedures",
            value="10",
        )
    
    with col2:
        st.metric(
            label="⏱️ Total Duration",
            value=">100 hours",
        )
    
    with col3:
        st.metric(
            label="🏷️ Modalities",
            value="3 (Video, Audio, Depth)",
        )
    
    with col4:
        st.metric(
            label="📂 Total Size",
            value="~12 TB",
        )
    
    st.markdown("---")
    
    # Dataset categories
    st.markdown("### 🏥 Dataset Details")

    st.info("The MM-OR dataset is the primary source of data for training and evaluating the models in this system.")

    col1, col2 = st.columns(2)

    with col1:
        st.markdown("#### Key Features")
        st.markdown("""
        - **Multimodal Data**: Includes synchronized video, multi-channel audio, and depth information.
        - **Multiple Views**: Video captured from multiple camera perspectives to provide a comprehensive view of the operating room.
        - **Rich Annotations**: Detailed annotations of:
            - Surgical roles (e.g., primary surgeon, assistant, nurse).
            - Atomic actions and complex activities.
            - Interactions between team members.
        - **Realistic Environment**: Data was collected in a high-fidelity simulated operating room.
        """)

    with col2:
        st.markdown("#### Data Modalities")
        st.image("https://www.researchgate.net/publication/359174963/figure/fig1/AS:1143128108556288@1649553881835/An-overview-of-our-data-acquisition-system-in-the-operating-room-OR-We-record.jpg", 
                 caption="Overview of the data acquisition system in the operating room.")

    st.markdown("---")
    st.markdown("### 📈 Data Distribution")

    # Create sample data for visualization
    procedure_data = {
        'Surgical Procedure': [f'Procedure {i+1}' for i in range(10)],
        'Duration (hours)': np.random.uniform(8, 12, 10).round(1),
        'Number of Annotations': np.random.randint(1500, 3000, 10)
    }
    df_procedures = pd.DataFrame(procedure_data)

    fig = px.bar(df_procedures, x='Surgical Procedure', y='Duration (hours)', 
                 title='Duration per Surgical Procedure',
                 labels={'Duration (hours)': 'Duration (hours)'},
                 color='Surgical Procedure')
    st.plotly_chart(fig, use_container_width=True)

    st.markdown("For more information, please refer to the original publication: *MM-OR: A Large-scale Multimodal Operating Room Dataset for Human Activity Recognition*.")
    st.markdown("The dataset is available on GitHub: [MM-OR Dataset](https://github.com/egeozsoy/MM-OR)")