import streamlit as st
import pandas as pd
import numpy as np
from datetime import datetime
import plotly.graph_objects as go
import plotly.express as px
from streamlit_autorefresh import st_autorefresh
import pickle
import os
from io import BytesIO
import pydicom
import cv2
import random
from ultralytics import YOLO
from plotly.subplots import make_subplots
# Load languages
def load_languages():
return {
'English': {
'welcome': 'Welcome',
'dashboard': 'Dashboard',
'profile': 'Profile',
'settings': 'Settings',
'emergency': 'Emergency Contact',
'about': 'About Us',
'prediction': 'Patient Prediction',
'analytics': 'Analytics'
},
'Spanish': {
'welcome': 'Bienvenido',
'dashboard': 'Tablero',
'profile': 'Perfil',
'settings': 'Ajustes',
'emergency': 'Contacto de Emergencia',
'about': 'Sobre Nosotros',
'prediction': 'Predicción de Pacientes',
'analytics': 'Análisis'
},
'French': {
'welcome': 'Bienvenue',
'dashboard': 'Tableau de Bord',
'profile': 'Profil',
'settings': 'Paramètres',
'emergency': 'Contact d\'urgence',
'about': 'À Propos',
'prediction': 'Prédiction de Patients',
'analytics': 'Analytique'
}
}
# Theme configurations
dark_theme = {
'primary_color': '#4da6ff',
'background_color': '#0e1117',
'secondary_bg': '#262730',
'text_color': '#fafafa',
'font': 'sans-serif',
'card_bg': '#1a1a1a',
'success_color': '#2fd36e',
'warning_color': '#ffd534',
'danger_color': '#ff4961'
}
# 3D Scatter plot for visual appeal
def create_3d_scatter():
x = np.random.randn(100)
y = np.random.randn(100)
z = np.random.randn(100)
fig = go.Figure(data=[go.Scatter3d(
x=x, y=y, z=z,
mode='markers',
marker=dict(
size=5,
color=z,
colorscale='Viridis',
opacity=0.8
)
)])
fig.update_layout(
scene=dict(
xaxis_title='X',
yaxis_title='Y',
zaxis_title='Z',
bgcolor='rgba(0,0,0,0)'
),
margin=dict(r=0, b=0, l=0, t=0),
paper_bgcolor='rgba(0,0,0,0)',
plot_bgcolor='rgba(0,0,0,0)'
)
return fig
def create_dynamic_dashboard():
st.title("Hospital Dashboard")
st_autorefresh(interval=10000, key="dashboard_refresh")
current_time = datetime.now()
times = pd.date_range(end=current_time, periods=20, freq='1min')
# 3D Scatter plot for visual appeal
st.subheader("Hospital Activity Visualization")
fig_3d = create_3d_scatter()
st.plotly_chart(fig_3d, use_container_width=True)
# Interactive metrics with hover effect
col1, col2, col3 = st.columns(3)
with col1:
current_patients = np.random.randint(80, 120)
st.metric("Current Patients", current_patients, delta=np.random.randint(-5, 5))
with col2:
bed_capacity = f"{np.random.randint(60, 90)}%"
st.metric("Bed Capacity", bed_capacity, delta=f"{np.random.randint(-3, 3)}%")
with col3:
staff_on_duty = np.random.randint(40, 60)
st.metric("Staff on Duty", staff_on_duty, delta=np.random.randint(-2, 2))
# Interactive charts
fig = make_subplots(rows=1, cols=2, subplot_titles=("Patient Flow (Last 20 minutes)", "Department Load (%)"))
# Patient Flow
fig.add_trace(
go.Scatter(
x=times,
y=np.random.randint(50, 100, size=20),
name="Admissions",
mode='lines+markers',
line=dict(color='#4da6ff', width=3),
marker=dict(size=8, symbol='circle', line=dict(color='#ffffff', width=2))
),
row=1, col=1
)
fig.add_trace(
go.Scatter(
x=times,
y=np.random.randint(40, 90, size=20),
name="Discharges",
mode='lines+markers',
line=dict(color='#ff4961', width=3),
marker=dict(size=8, symbol='circle', line=dict(color='#ffffff', width=2))
),
row=1, col=1
)
# Department Load
departments = ['ER', 'ICU', 'Surgery', 'Pediatrics', 'General']
values = np.random.randint(40, 100, size=len(departments))
fig.add_trace(
go.Bar(
x=departments,
y=values,
marker_color='#4da6ff',
hoverinfo='y',
textposition='auto',
textfont=dict(color='white'),
hoverlabel=dict(bgcolor='#1a1a1a', font_size=14)
),
row=1, col=2
)
fig.update_layout(
height=500,
showlegend=False,
paper_bgcolor='rgba(0,0,0,0)',
plot_bgcolor='rgba(0,0,0,0)',
font=dict(color='#fafafa'),
margin=dict(l=20, r=20, t=60, b=20),
)
fig.update_xaxes(showgrid=False, zeroline=False)
fig.update_yaxes(showgrid=True, gridcolor='rgba(255,255,255,0.1)', zeroline=False)
st.plotly_chart(fig, use_container_width=True)
emergency_data = pd.DataFrame({
'Time': times[-5:],
'Type': np.random.choice(['Critical', 'Moderate', 'Minor'], size=5),
'Department': np.random.choice(['ER', 'ICU', 'Surgery'], size=5),
'Status': np.random.choice(['In Progress', 'Waiting', 'Completed'], size=5)
})
st.subheader("Recent Emergency Cases")
st.dataframe(
emergency_data.style
.applymap(lambda x: f'color: {dark_theme["text_color"]}; background-color: {dark_theme["card_bg"]}')
.set_properties(**{'background-color': dark_theme['card_bg'], 'color': dark_theme['text_color'], 'border-color': dark_theme['primary_color']})
.highlight_max(axis=0, props='color: #ff4961; font-weight: bold;')
.highlight_min(axis=0, props='color: #2fd36e; font-weight: bold;'),
use_container_width=True
)
def user_profile_section():
st.title("User Profile")
if 'user_profile' not in st.session_state:
st.session_state.user_profile = {
'name': '',
'email': '',
'phone': '',
'department': '',
'role': '',
'profile_picture': None
}
col1, col2 = st.columns(2)
with col1:
st.session_state.user_profile['name'] = st.text_input("Full Name", st.session_state.user_profile['name'])
st.session_state.user_profile['email'] = st.text_input("Email", st.session_state.user_profile['email'])
st.session_state.user_profile['phone'] = st.text_input("Phone", st.session_state.user_profile['phone'])
with col2:
st.session_state.user_profile['department'] = st.selectbox(
"Department",
["Cardiology", "Emergency", "Pediatrics", "Surgery", "Other"],
index=0 if not st.session_state.user_profile['department'] else None
)
st.session_state.user_profile['role'] = st.selectbox(
"Role",
["Doctor", "Nurse", "Administrator", "Other"],
index=0 if not st.session_state.user_profile['role'] else None
)
uploaded_file = st.file_uploader("Upload Profile Picture", type=['jpg', 'png'])
if uploaded_file is not None:
st.session_state.user_profile['profile_picture'] = uploaded_file
st.image(uploaded_file, width=150)
if st.button("Save Profile"):
st.success("Profile updated successfully!")
def emergency_contact_section():
st.title("Emergency Contact")
st.header("Emergency Department")
col1, col2, col3 = st.columns(3)
with col1:
st.info("📞 Emergency Hotline\n\n1-800-HOSPITAL")
with col2:
st.info("🚑 Ambulance Service\n\n911")
with col3:
st.info("👨⚕️ On-call Doctor\n\n+1-555-0123")
st.header("Contact Form")
emergency_type = st.selectbox(
"Emergency Type",
["Medical Emergency", "Fire Emergency", "Security Issue", "Other"]
)
description = st.text_area("Description of Emergency")
location = st.text_input("Location in Hospital")
if st.button("Submit Emergency Alert"):
if description and location:
st.success("Emergency alert submitted successfully!")
st.info("Emergency response team has been notified.")
else:
st.error("Please fill in all required fields.")
def about_us_section():
st.title("About Us")
st.markdown("""
## Our Mission
To provide exceptional healthcare services with compassion and innovation,
ensuring the best possible outcomes for our patients and communities.
## Our Vision
To be the leading healthcare provider known for excellence in patient care,
medical research, and healthcare technology innovation.
## Our Values
- **Excellence** in all aspects of healthcare
- **Compassion** towards all patients
- **Innovation** in medical practices
- **Integrity** in our actions
- **Teamwork** in our approach
## Hospital Statistics
""")
col1, col2, col3 = st.columns(3)
with col1:
st.metric("Years of Service", "50+")
with col2:
st.metric("Healthcare Professionals", "1000+")
with col3:
st.metric("Patients Served Annually", "50,000+")
st.header("Our Departments")
departments = [
("🫀 Cardiology", "Specialized heart care and treatment"),
("🧠 Neurology", "Expert neurological care"),
("👶 Pediatrics", "Comprehensive children's healthcare"),
("🏥 Emergency Care", "24/7 emergency services"),
("🔬 Research", "Cutting-edge medical research")
]
for dept, desc in departments:
st.subheader(dept)
st.write(desc)
def prediction_page():
st.title("Patient Readmission Prediction")
col1, col2 = st.columns(2)
with col1:
gender = st.selectbox('Gender:', ["Female", "Male", "Other"])
admission_type = st.selectbox('Admission Type:', ['Emergency', 'Urgent', 'Elective'])
diagnosis = st.selectbox('Diagnosis:', ['Heart Disease', 'Diabetes', 'Injury', 'Infection'])
lab_procedures = st.number_input('Number of Lab Procedures:', 1, 100, 1)
medications = st.number_input('Number of Medications:', 1, 36, 1)
with col2:
outpatient_visits = st.number_input('Number of Outpatient Visits:', 0, 5, 0)
inpatient_visits = st.number_input('Number of Inpatient Visits:', 0, 5, 0)
emergency_visits = st.number_input('Number of Emergency Visits:', 0, 5, 0)
num_diagnoses = st.number_input('Number of Diagnoses:', 1, 10, 1)
a1c_result = st.selectbox('A1C Result:', ['Normal', 'Abnormal'])
if st.button("Predict Readmission", key="predict_button"):
gender_code = {"Female": 0, "Male": 1, "Other": 2}[gender]
admission_code = {"Emergency": 1, "Urgent": 2, "Elective": 0}[admission_type]
diagnosis_code = {"Heart Disease": 1, "Diabetes": 0, "Injury": 3, "Infection": 2}[diagnosis]
a1c_code = {"Normal": 1, "Abnormal": 0}[a1c_result]
input_data = np.array([[
gender_code, admission_code, diagnosis_code, lab_procedures,
medications, outpatient_visits, inpatient_visits,
emergency_visits, num_diagnoses, a1c_code
]])
try:
model_path = os.path.join(os.path.dirname(__file__), "Readmission_Model.pkl")
with open(model_path, "rb") as m:
model = pickle.load(m)
result = model.predict(input_data)[0]
if result == 1:
st.error("⚠️ High Risk: Readmission is Required")
st.markdown("""
### Recommended Actions:
1. Schedule follow-up appointment within 7 days
2. Review medication compliance
3. Coordinate with care management team
""")
else:
st.success("✅ Low Risk: Readmission is Not Required")
st.markdown("""
### Recommended Actions:
1. Schedule routine follow-up within 30 days
2. Provide standard discharge instructions
3. Document any concerns for future reference
""")
except FileNotFoundError:
st.error("Model file not found. Please ensure 'Readmission_Model.pkl' is in the correct directory.")
except Exception as e:
st.error(f"An error occurred during prediction: {str(e)}")
def analytics_page():
st.title("Hospital Analytics Dashboard")
time_period = st.selectbox(
"Select Time Period",
["Last 24 Hours", "Last Week", "Last Month", "Last Year"]
)
if time_period == "Last 24 Hours":
dates = pd.date_range(end=datetime.now(), periods=24, freq='H')
elif time_period == "Last Week":
dates = pd.date_range(end=datetime.now(), periods=7, freq='D')
elif time_period == "Last Month":
dates = pd.date_range(end=datetime.now(), periods=30, freq='D')
else:
dates = pd.date_range(end=datetime.now(), periods=12, freq='M')
analytics_data = pd.DataFrame({
'Date': dates,
'Admissions': np.random.randint(50, 150, size=len(dates)),
'Discharges': np.random.randint(40, 140, size=len(dates)),
'Readmissions': np.random.randint(5, 30, size=len(dates)),
'Average_Stay': np.random.uniform(2, 7, size=len(dates))
})
col1, col2, col3, col4 = st.columns(4)
with col1:
total_admissions = analytics_data['Admissions'].sum()
st.metric("Total Admissions", f"{total_admissions:,}")
with col2:
avg_stay = analytics_data['Average_Stay'].mean()
st.metric("Average Stay (days)", f"{avg_stay:.1f}")
with col3:
readmission_rate = (analytics_data['Readmissions'].sum() / total_admissions) * 100
st.metric("Readmission Rate", f"{readmission_rate:.1f}%")
with col4:
bed_turnover = total_admissions / len(dates)
st.metric("Daily Bed Turnover", f"{bed_turnover:.1f}")
fig = make_subplots(rows=2, cols=2, subplot_titles=("Admissions vs Discharges Trend", "Readmission Trend", "Average Length of Stay Trend", "Department-wise Statistics"))
# Admissions vs Discharges Trend
fig.add_trace(go.Scatter(
x=analytics_data['Date'],
y=analytics_data['Admissions'],
name="Admissions",
line=dict(color='#4da6ff', width=3)
), row=1, col=1)
fig.add_trace(go.Scatter(
x=analytics_data['Date'],
y=analytics_data['Discharges'],
name="Discharges",
line=dict(color='#ff4961', width=3)
), row=1, col=1)
# Readmission Trend
fig.add_trace(go.Scatter(
x=analytics_data['Date'],
y=analytics_data['Readmissions'],
name="Readmissions",
line=dict(color='#ffd534', width=3)
), row=1, col=2)
# Average Length of Stay Trend
fig.add_trace(go.Scatter(
x=analytics_data['Date'],
y=analytics_data['Average_Stay'],
name="Average Stay",
line=dict(color='#2fd36e', width=3)
), row=2, col=1)
# Department-wise Statistics
departments = ['Emergency', 'Surgery', 'Cardiology', 'Pediatrics', 'Neurology']
occupancy_rates = np.random.uniform(60, 95, len(departments))
fig.add_trace(go.Bar(
x=departments,
y=occupancy_rates,
name="Occupancy Rate",
marker_color='#4da6ff'
), row=2, col=2)
fig.update_layout(
height=800,
showlegend=True,
paper_bgcolor='rgba(0,0,0,0)',
plot_bgcolor='rgba(0,0,0,0)',
font=dict(color='#fafafa'),
legend=dict(
bgcolor='rgba(0,0,0,0)',
bordercolor='rgba(0,0,0,0)'
)
)
fig.update_xaxes(showgrid=False, zeroline=False)
fig.update_yaxes(showgrid=True, gridcolor='rgba(255,255,255,0.1)', zeroline=False)
st.plotly_chart(fig, use_container_width=True)
st.subheader("Department-wise Statistics")
dept_data = pd.DataFrame({
'Department': departments,
'Occupancy_Rate': occupancy_rates,
'Avg_Stay': np.random.uniform(2, 8, len(departments)),
'Patient_Satisfaction': np.random.uniform(75, 95, len(departments))
})
st.dataframe(
dept_data.style.format({
'Occupancy_Rate': '{:.2f}%',
'Avg_Stay': '{:.2f}',
'Patient_Satisfaction': '{:.2f}%'
})
.applymap(lambda x: f'color: {dark_theme["text_color"]}; background-color: {dark_theme["card_bg"]}')
.set_properties(**{'background-color': dark_theme['card_bg'], 'color': dark_theme['text_color'], 'border-color': dark_theme['primary_color']})
.highlight_max(axis=0, props='color: #2fd36e; font-weight: bold;')
.highlight_min(axis=0, props='color: #ff4961; font-weight: bold;'),
use_container_width=True
)
def medical_image_analysis_page():
st.title("Medical Image Analysis")
# Disease details dictionary
DISEASE_DETAILS = {
'Aortic enlargement': {
'description': 'Abnormal enlargement of the aorta',
'precautions': [
'Immediate cardiovascular consultation',
'Regular blood pressure monitoring',
'Avoid heavy lifting and strenuous activities'
],
'admission': 'Immediate hospitalization if risk of rupture'
},
'Cardiomegaly': {
'description': 'Abnormal enlargement of the heart',
'precautions': [
'Restrict physical activities',
'Follow strict medication regimen',
'Regular cardiac monitoring'
],
'admission': 'Urgent cardiac care if symptoms are severe'
},
'Pneumothorax': {
'description': 'Collapsed or partially collapsed lung',
'precautions': [
'Oxygen therapy',
'Chest tube insertion may be required',
'Complete bed rest'
],
'admission': 'Immediate hospitalization'
},
'No finding': {
'description': 'No significant medical conditions detected',
'precautions': ['Regular health check-ups'],
'admission': 'Not required'
}
}
# Ensure uploads directory exists
os.makedirs('uploads', exist_ok=True)
# Load YOLO model
MODEL_PATH = 'yolov8n.pt'
model = YOLO(MODEL_PATH)
# Disease classes
CLASSES = [
'Aortic enlargement', 'Atelectasis', 'Calcification', 'Cardiomegaly',
'Consolidation', 'ILD', 'Infiltration', 'Lung Opacity', 'Nodule/Mass',
'Other lesion', 'Pleural effusion', 'Pleural thickening', 'Pneumothorax',
'Pulmonary fibrosis', 'No finding'
]
st.markdown("### Upload Medical Image")
uploaded_file = st.file_uploader(
"Choose a medical image (DICOM, JPG, PNG)",
type=['dcm', 'dicom', 'jpg', 'png']
)
if uploaded_file is not None:
# Save the uploaded file
with open(os.path.join("uploads", uploaded_file.name), "wb") as f:
f.write(uploaded_file.getbuffer())
file_path = os.path.join("uploads", uploaded_file.name)
try:
# Read the image
if uploaded_file.name.endswith('.dcm') or uploaded_file.name.endswith('.dicom'):
dicom = pydicom.dcmread(file_path)
img = dicom.pixel_array
else:
img = cv2.imread(file_path)
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
# Normalize and convert to RGB
img = (img / img.max() * 255).astype(np.uint8)
img = cv2.cvtColor(img, cv2.COLOR_GRAY2RGB)
# Perform inference
results = model(img)
# Process results
predictions = []
for r in results:
boxes = r.boxes.xyxy.cpu().numpy()
confidences = r.boxes.conf.cpu().numpy()
class_ids = r.boxes.cls.cpu().numpy().astype(int)
for box, confidence, class_id in zip(boxes, confidences, class_ids):
x_min, y_min, x_max, y_max = box
predictions.append({
'class_name': CLASSES[class_id],
'confidence': float(confidence),
'box': {
'x_min': int(x_min),
'y_min': int(y_min),
'x_max': int(x_max),
'y_max': int(y_max)
}
})
# If no findings, add a default prediction
if not predictions:
predictions.append({
'class_name': 'No finding',
'confidence': 1.0,
'box': {'x_min': 0, 'y_min': 0, 'x_max': 1, 'y_max': 1}
})
# Visualize results
visualized_img = img.copy()
for pred in predictions:
color = [random.randint(0, 255) for _ in range(3)]
visualized_img = draw_bbox(visualized_img,
[pred['box']['x_min'], pred['box']['y_min'],
pred['box']['x_max'], pred['box']['y_max']],
pred['class_name'],
pred['confidence'],
color)
st.subheader("Analysis Results")
# Create responsive layout
col1, col2 = st.columns([2, 1])
with col1:
# Display image responsively
st.image(visualized_img, use_column_width=True, caption="Analyzed Medical Image")
with col2:
# Detailed analysis report
st.markdown("### Detected Conditions")
# Check if predictions exist
if not predictions:
st.info("No significant findings detected.")
else:
for pred in predictions:
disease = pred['class_name']
confidence = pred['confidence']
st.markdown(f"#### {disease}")
st.markdown(f"**Confidence:** {confidence:.2%}")
# Retrieve disease details
if disease in DISEASE_DETAILS:
details = DISEASE_DETAILS[disease]
st.markdown(f"**Description:** {details['description']}")
st.markdown("**Precautions:**")
for precaution in details['precautions']:
st.markdown(f"- {precaution}")
st.markdown(f"**Hospital Admission:** {details['admission']}")
st.markdown("---")
# Clean up temporary file
os.remove(file_path)
except Exception as e:
st.error(f"An error occurred during image analysis: {str(e)}")
def draw_bbox(image, box, label, confidence, color):
alpha = 0.1
alpha_box = 0.4
overlay_bbox = image.copy()
overlay_text = image.copy()
output = image.copy()
text_width, text_height = cv2.getTextSize(label.upper(), cv2.FONT_HERSHEY_SIMPLEX, 0.6, 1)[0]
cv2.rectangle(overlay_bbox, (int(box[0]), int(box[1])), (int(box[2]), int(box[3])), color, -1)
cv2.addWeighted(overlay_bbox, alpha, output, 1 - alpha, 0, output)
cv2.rectangle(overlay_text, (int(box[0]), int(box[1])-7-text_height), (int(box[0])+text_width+2, int(box[1])), (0, 0, 0), -1)
cv2.addWeighted(overlay_text, alpha_box, output, 1 - alpha_box, 0, output)
cv2.rectangle(output, (int(box[0]), int(box[1])), (int(box[2]), int(box[3])), color, 2)
cv2.putText(output, label.upper(), (int(box[0]), int(box[1])-5), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 255, 255), 1, cv2.LINE_AA)
cv2.putText(output, f"{confidence:.2f}", (int(box[0]), int(box[3])+20), cv2.FONT_HERSHEY_SIMPLEX, 0.5, color, 1, cv2.LINE_AA)
return output
import google.generativeai as genai
from typing import List
from datetime import datetime
from fpdf import FPDF
import io
import PIL.Image
from gtts import gTTS
import base64
def chatbot_page():
st.title("🏥 Hospital AI Assistant")
st.write("Chat with our AI assistant for hospital-related queries and image analysis.")
# Language settings in sidebar
st.sidebar.title("Language Settings")
language = st.sidebar.radio("Choose Response Language:", ["English", "Hindi", "Marathi"])
# Configure Gemini API
genai.configure(api_key=st.secrets["google"]["api_key"])
# Initialize models
text_model = genai.GenerativeModel('gemini-pro')
vision_model = genai.GenerativeModel('gemini-pro-vision')
if "chat_history" not in st.session_state:
st.session_state.chat_history = []
st.session_state.chat = text_model.start_chat(history=[])
# Image upload section
uploaded_image = st.file_uploader("📤 Upload an image for analysis", type=['jpg', 'jpeg', 'png'])
if uploaded_image:
# Save and display uploaded image
image = PIL.Image.open(uploaded_image)
st.image(image, caption="Uploaded Image", use_column_width=True)
# Add image analysis button
# if st.button("🔍 Analyze Image"):
# with st.spinner("Analyzing image..."):
# try:
# prompt = f"Analyze this medical image and describe what you see in {language.lower()}. If it's not a medical image, please provide a general description."
# response = vision_model.generate_content([prompt, image])
# if response and hasattr(response, "text"):
# analysis_text = response.text
# st.session_state.chat_history.append({
# "role": "assistant",
# "content": analysis_text,
# "type": "image_analysis"
# })
# # Add text-to-speech option
# st.write("#### 📢 Listen to the analysis:")
# if st.button("🔊 Play Analysis"):
# lang_code = 'hi' if language == "Hindi" else 'mr' if language == "Marathi" else 'en'
# tts = gTTS(text=analysis_text, lang=lang_code)
# audio = BytesIO()
# tts.write_to_fp(audio)
# audio.seek(0)
# audio_base64 = base64.b64encode(audio.read()).decode()
# audio_tag = f'<audio controls autoplay><source src="data:audio/mp3;base64,{audio_base64}" type="audio/mp3"></audio>'
# st.markdown(audio_tag, unsafe_allow_html=True)
# except Exception as e:
# st.error(f"Error during image analysis: {str(e)}")
# Display chat history
for message in st.session_state.chat_history:
with st.chat_message(message["role"]):
st.write(message["content"])
if message.get("type") == "image_analysis":
# Add audio playback option for image analysis
if st.button(f"🔊 Play Response", key=f"play_{len(st.session_state.chat_history)}"):
lang_code = 'hi' if language == "Hindi" else 'mr' if language == "Marathi" else 'en'
tts = gTTS(text=message["content"], lang=lang_code)
audio = BytesIO()
tts.write_to_fp(audio)
audio.seek(0)
audio_base64 = base64.b64encode(audio.read()).decode()
audio_tag = f'<audio controls autoplay><source src="data:audio/mp3;base64,{audio_base64}" type="audio/mp3"></audio>'
st.markdown(audio_tag, unsafe_allow_html=True)
# Chat input
if prompt := st.chat_input("Ask me anything about the hospital..."):
st.session_state.chat_history.append({"role": "user", "content": prompt})
with st.chat_message("user"):
st.write(prompt)
try:
# Handle regular text queries
response = text_model.generate_content([
f"Respond in {language.lower()} to the following query: {prompt}"
])
bot_response = response.text
st.session_state.chat_history.append({
"role": "assistant",
"content": bot_response,
"type": "text"
})
with st.chat_message("assistant"):
st.write(bot_response)
# Add audio option for text responses
# if st.button("🔊 Play Response", key=f"play_text_{len(st.session_state.chat_history)}"):
# lang_code = 'hi' if language == "Hindi" else 'mr' if language == "Marathi" else 'en'
# tts = gTTS(text=bot_response, lang=lang_code)
# audio = BytesIO()
# tts.write_to_fp(audio)
# audio.seek(0)
# audio_base64 = base64.b64encode(audio.read()).decode()
# audio_tag = f'<audio controls autoplay><source src="data:audio/mp3;base64,{audio_base64}" type="audio/mp3"></audio>'
# st.markdown(audio_tag, unsafe_allow_html=True)
except Exception as e:
st.error(f"Error: {str(e)}")
# Export options
# st.sidebar.title("Chat Options")
# if st.sidebar.button("🗑️ Clear Chat"):
# st.session_state.chat_history = []
# st.session_state.chat = text_model.start_chat(history=[])
# st.experimental_rerun()
if st.sidebar.button("📥 Export Chat"):
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
txt_content = ""
for msg in st.session_state.chat_history:
txt_content += f"{msg['role'].title()}: {msg['content']}\n\n"
st.sidebar.download_button(
"📄 Download Chat",
data=txt_content,
file_name=f"chat_export_{timestamp}.txt",
mime="text/plain"
)
# Footer
st.markdown("---")
st.markdown("Made with ❤️ for Tatva-AI")
# Add chatbot to main navigation
import tensorflow as tf
from PIL import Image
# Add this function after the medical_image_analysis_page() function
def brain_tumor_detection_page():
st.title("Brain Tumor Detection")
def load_model_with_custom_objects(model_path):
def custom_depthwise_conv2d(*args, **kwargs):
if 'groups' in kwargs:
del kwargs['groups']
return tf.keras.layers.DepthwiseConv2D(*args, **kwargs)
custom_objects = {
'DepthwiseConv2D': custom_depthwise_conv2d,
'tf': tf
}
try:
model = tf.keras.models.load_model(
model_path,
custom_objects=custom_objects,
compile=False
)
return model
except Exception as e:
st.error(f"Model loading failed: {e}")
return None
def preprocess_image(image):
image = image.convert("RGB")
image = image.resize((224, 224))
image_array = np.array(image) / 255.0
image_array = np.expand_dims(image_array, axis=0)
return image_array
model_path = 'models/keras_model.h5'
model = load_model_with_custom_objects(model_path)
if model is None:
st.error("Could not load the model.")
return
uploaded_file = st.file_uploader("Upload brain scan image", type=['jpg', 'png', 'jpeg'])
if uploaded_file is not None:
image = Image.open(uploaded_file)
st.image(image, caption="Uploaded Scan", use_column_width=True)
try:
processed_image = preprocess_image(image)
predictions = model.predict(processed_image)
CLASS_LABELS = ["Pituitary", "No Tumor", "Meningioma", "Glioma"]
st.subheader("Prediction Results:")
for label, prob in zip(CLASS_LABELS, predictions[0]):
st.progress(float(prob))
st.write(f"{label}: {prob*100:.2f}%")
predicted_class = CLASS_LABELS[np.argmax(predictions)]
st.success(f"Predicted Condition: {predicted_class}")
except Exception as e:
st.error(f"Analysis failed: {e}")
def main():
st.set_page_config(
page_title="Hospital Management System",
layout="wide",
initial_sidebar_state="expanded"
)
if 'language' not in st.session_state:
st.session_state.language = 'English'
languages = load_languages()
# Sidebar Navigation
with st.sidebar:
st.markdown("<h2 style=' color: #4da6ff;'>Tatva AI</h2>", unsafe_allow_html=True)
st.image("logo.jpeg", width=100)
if 'user_profile' in st.session_state and st.session_state.user_profile['name']:
st.write(f"Welcome, {st.session_state.user_profile['name']}")
if st.session_state.user_profile['profile_picture']:
st.image(st.session_state.user_profile['profile_picture'], width=100)
# Navigation menu
menu_options = [
"Dashboard",
"Patient Prediction",
"Analytics",
"Brain Tumor Detection",
"Medical Image Analysis",
"User Profile",
"Emergency Contact",
"About Us",
"Settings",
"Chatbot"
]
selected_page = st.selectbox(
languages[st.session_state.language]['welcome'],
menu_options
)
if selected_page == "Chatbot":
chatbot_page()
# Page rendering based on selection
if selected_page == "Dashboard":
create_dynamic_dashboard()
elif selected_page == "Patient Prediction":
prediction_page()
elif selected_page == "Brain Tumor Detection":
brain_tumor_detection_page()
elif selected_page == "Analytics":
analytics_page()
elif selected_page == "Medical Image Analysis":
medical_image_analysis_page()
elif selected_page == "User Profile":
user_profile_section()
elif selected_page == "Emergency Contact":
emergency_contact_section()
elif selected_page == "About Us":
about_us_section()
elif selected_page == "Settings":
st.title("Settings")
st.subheader("Language Settings")
new_language = st.selectbox(
"Select Language",
list(languages.keys()),
index=list(languages.keys()).index(st.session_state.language)
)
if new_language != st.session_state.language:
st.session_state.language = new_language
st.experimental_rerun()
st.subheader("Theme Settings")
theme = st.selectbox(
"Choose Theme",
["Dark"]
)
current_theme = dark_theme
st.markdown(f"""
<style>
:root {{
--primary-color: {current_theme['primary_color']};
--background-color: {current_theme['background_color']};
--secondary-bg: {current_theme['secondary_bg']};
--text-color: {current_theme['text_color']};
--font: {current_theme['font']};
--card-bg: {current_theme['card_bg']};
--success-color: {current_theme['success_color']};
--warning-color: {current_theme['warning_color']};
--danger-color: {current_theme['danger_color']};
}}
</style>
""", unsafe_allow_html=True)
# Add 3D effect and continuous movement
st.markdown("""
<script>
const cursor = document.createElement('div');
cursor.className = 'custom-cursor';
document.body.appendChild(cursor);
document.addEventListener('mousemove', (e) => {
cursor.style.left = e.clientX + 'px';
cursor.style.top = e.clientY + 'px';
const elements = document.querySelectorAll('.stPlotlyChart, .stDataFrame, .stMetric');
elements.forEach(el => {
const rect = el.getBoundingClientRect();
const x = (e.clientX - rect.left) / rect.width - 0.5;
const y = (e.clientY - rect.top) / rect.height - 0.5;
el.style.transform = `perspective(1000px) rotateY(${x * 5}deg) rotateX(${-y * 5}deg)`;
});
});
</script>
<style>
.custom-cursor {
width: 20px;
height: 20px;
border: 2px solid #ffffff;
border-radius: 50%;
position: fixed;
pointer-events: none;
z-index: 9999;
mix-blend-mode: difference;
}
.stPlotlyChart, .stDataFrame, .stMetric {
transition: transform 0.1s ease;
}
.stMetric {
background: linear-gradient(145deg, rgba(26,26,26,0.6) 0%, rgba(26,26,26,0.8) 100%);
border-radius: 10px;
padding: 10px;
box-shadow: 0 4px 6px rgba(0,0,0,0.1);
}
.stMetric:hover {
background: linear-gradient(145deg, rgba(26,26,26,0.8) 0%, rgba(26,26,26,1) 100%);
}
</style>
""", unsafe_allow_html=True)
if __name__ == "__main__":
main()
Datasets
Models
