# Complete code for Hugging Face Spaces deployment
# Your CSV file name: brain_tumor_dataset.csv
import gradio as gr
import pandas as pd
import numpy as np
from sklearn.ensemble import RandomForestClassifier
from sklearn.preprocessing import LabelEncoder
import os
print("Loading Brain Tumor Detection System...")
# Load dataset from CSV file
csv_file = 'brain_tumor_dataset.csv'
if not os.path.exists(csv_file):
print(f"Error: {csv_file} not found!")
print("Please upload brain_tumor_dataset.csv to the current directory")
else:
df = pd.read_csv(csv_file)
print(f"✅ Dataset loaded: {len(df)} patient records")
print(f"📊 Columns in CSV: {list(df.columns)}")
# Check which columns exist and use only available ones
available_columns = df.columns.tolist()
# Define required columns
required_columns = ['Age', 'Gender', 'Headache_Score', 'Memory_Loss', 'Visual_Disturbance',
'Nausea', 'Vomiting', 'Seizures', 'Weakness', 'Coordination_Issues', 'Tumor_Detected']
# Check if all required columns exist
for col in required_columns:
if col not in available_columns:
print(f"⚠️ Warning: '{col}' not found in CSV")
# Use only columns that exist
feature_columns = [col for col in required_columns if col in available_columns]
print(f"✅ Using features: {feature_columns}")
# Train model with available columns
le_gender = LabelEncoder()
le_memory = LabelEncoder()
le_visual = LabelEncoder()
le_nausea = LabelEncoder()
le_vomiting = LabelEncoder()
le_seizures = LabelEncoder()
le_weakness = LabelEncoder()
le_coordination = LabelEncoder()
le_tumor = LabelEncoder()
# Encode categorical columns
df['Gender_E'] = le_gender.fit_transform(df['Gender'])
df['Memory_E'] = le_memory.fit_transform(df['Memory_Loss'])
df['Visual_E'] = le_visual.fit_transform(df['Visual_Disturbance'])
df['Nausea_E'] = le_nausea.fit_transform(df['Nausea'])
df['Vomiting_E'] = le_vomiting.fit_transform(df['Vomiting'])
df['Seizures_E'] = le_seizures.fit_transform(df['Seizures'])
df['Weakness_E'] = le_weakness.fit_transform(df['Weakness'])
df['Coordination_E'] = le_coordination.fit_transform(df['Coordination_Issues'])
df['Tumor_E'] = le_tumor.fit_transform(df['Tumor_Detected'])
# Prepare features
X = df[['Age', 'Gender_E', 'Headache_Score', 'Memory_E', 'Visual_E',
'Nausea_E', 'Vomiting_E', 'Seizures_E', 'Weakness_E', 'Coordination_E']]
y = df['Tumor_E']
# Train model
model = RandomForestClassifier(n_estimators=200, max_depth=15, random_state=42)
model.fit(X, y)
print("✅ Model trained successfully!")
def predict_tumor(age, gender, headache, memory_loss, visual_disturbance,
nausea, vomiting, seizures, weakness, coordination_issues):
gender_e = 0 if gender == 'Male' else 1
memory_e = 1 if memory_loss == 'Yes' else 0
visual_e = 1 if visual_disturbance == 'Yes' else 0
nausea_e = 1 if nausea == 'Yes' else 0
vomiting_e = 1 if vomiting == 'Yes' else 0
seizures_e = 1 if seizures == 'Yes' else 0
weakness_e = 1 if weakness == 'Yes' else 0
coordination_e = 1 if coordination_issues == 'Yes' else 0
input_data = [[age, gender_e, headache, memory_e, visual_e, nausea_e,
vomiting_e, seizures_e, weakness_e, coordination_e]]
prediction = model.predict(input_data)[0]
probability = model.predict_proba(input_data)[0][1]
result = "TUMOR DETECTED" if prediction == 1 else "NO TUMOR"
if prediction == 1:
recommendations = """
🚨 **URGENT RECOMMENDATIONS:**
• Consult a neurologist immediately
• Schedule brain MRI with contrast
• Get a complete neurological examination
• Do not delay treatment
"""
color = "#c8a2c8"
icon = "⚠️"
else:
recommendations = """
✅ **HEALTHY RECOMMENDATIONS:**
• Continue regular health checkups
• Maintain healthy lifestyle
• Monitor any persistent symptoms
• Annual medical examination
"""
color = "#e6e6fa"
icon = "✅"
return result, f"{probability*100:.1f}%", recommendations, color, icon
custom_css = """
"""
with gr.Blocks(css=custom_css, title="Brain Tumor Detection System") as demo:
gr.HTML("""
🧠 BRAIN TUMOR DETECTION SYSTEM
✨ AI-Powered Medical Diagnosis | Instant Results ✨
""")
with gr.Row():
with gr.Column(scale=1):
with gr.Group(elem_classes="card"):
gr.Markdown("### 📋 PATIENT INFORMATION")
age = gr.Slider(
minimum=1, maximum=120, value=45, step=1,
label="🎂 Age (Years)",
info="Patient's current age"
)
gender = gr.Radio(
choices=['Male', 'Female'], label="⚥ Gender",
info="Biological sex", value='Male'
)
headache = gr.Slider(
minimum=1, maximum=60, value=30, step=1,
label="🤕 Headache Severity (1-60)",
info="Pain intensity scale - higher score means more severe headache"
)
with gr.Group(elem_classes="card"):
gr.Markdown("### 🧠 NEUROLOGICAL SYMPTOMS")
memory_loss = gr.Radio(
choices=['No', 'Yes'], label="📝 Memory Loss",
info="Forgetting recent events", value='No'
)
visual_disturbance = gr.Radio(
choices=['No', 'Yes'], label="👁️ Visual Disturbance",
info="Blurred or double vision", value='No'
)
coordination_issues = gr.Radio(
choices=['No', 'Yes'], label="🎯 Coordination Issues",
info="Difficulty walking or maintaining balance", value='No'
)
with gr.Column(scale=1):
with gr.Group(elem_classes="card"):
gr.Markdown("### 🏥 PHYSICAL SYMPTOMS")
nausea = gr.Radio(
choices=['No', 'Yes'], label="🤢 Nausea",
info="Feeling sick to stomach", value='No'
)
vomiting = gr.Radio(
choices=['No', 'Yes'], label="🤮 Vomiting",
info="Throwing up, especially in morning", value='No'
)
seizures = gr.Radio(
choices=['No', 'Yes'], label="⚡ Seizures",
info="Uncontrolled body movements", value='No'
)
weakness = gr.Radio(
choices=['No', 'Yes'], label="💪 Weakness",
info="Unexplained loss of strength", value='No'
)
predict_btn = gr.Button("🔍 PREDICT NOW", variant="primary", size="lg")
with gr.Row():
with gr.Column(scale=1):
with gr.Group(elem_classes="card"):
gr.Markdown("### 📊 DIAGNOSIS RESULT")
result_text = gr.HTML(label="Prediction")
probability = gr.Textbox(
label="Confidence Score", interactive=False,
placeholder="Click predict to see result",
lines=1, show_label=True
)
recommendations = gr.Markdown("### 💡 RECOMMENDATIONS\n\n*Click predict to see recommendations*")
def handle_prediction(age, gender, headache, memory_loss, visual_disturbance,
nausea, vomiting, seizures, weakness, coordination_issues):
result, prob, recs, color, icon = predict_tumor(
age, gender, headache, memory_loss, visual_disturbance,
nausea, vomiting, seizures, weakness, coordination_issues
)
styled_result = f"""
{icon}
{result}
Risk Assessment Complete
"""
return styled_result, prob, recs
predict_btn.click(
handle_prediction,
inputs=[age, gender, headache, memory_loss, visual_disturbance,
nausea, vomiting, seizures, weakness, coordination_issues],
outputs=[result_text, probability, recommendations]
)
gr.HTML("""
""")
demo.launch()