Upload app.py

app.py

@@ -1,7 +1,174 @@
+# -*- coding: utf-8 -*-
+"""2_preprocessing_test.ipynb
+
+Automatically generated by Colab.
+
+Original file is located at
+    https://colab.research.google.com/drive/10c3x9G9z70J73l0LJDA8_VDZphQmHEZB
+"""
+
+from google.colab import drive
+drive.mount('/content/drive')
+
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+from sklearn.preprocessing import LabelEncoder
+import os
+from sklearn.model_selection import train_test_split
+import pickle
+import warnings
+warnings.filterwarnings('ignore')
+
+df1 = pd.read_csv("/content/drive/MyDrive/Google Colab/disease-symptom-prediction/data/dataset.csv")
+
+print(df1.shape)
+df1.head()
+
+df1.sort_values(by='Disease', inplace=True)
+df1.head()
+
+df1.drop_duplicates(inplace=True)
+df1.shape
+
+df1['Disease'].value_counts()
+
+df1[df1['Disease']=="Fungal infection"]
+
+df1.fillna("none", inplace=True)
+df1[df1['Disease']=="Fungal infection"]
+
+df1.columns = df1.columns.str.strip().str.lower()
+for col in df1.columns:
+    df1[col] = df1[col].astype(str).str.strip().str.lower()
+
+
+symptom_cols = [col for col in df1.columns if col.startswith('symptom')]
+print(symptom_cols)
+
+all_symptoms = set()
+for col in symptom_cols:
+    for val in df1[col].unique():
+        if val != 'none':
+            all_symptoms.add(val)
+print(f"Unique symptoms: {len(all_symptoms)}")
+
+print(all_symptoms)
+
+df1.head()
+
+df1_num = pd.DataFrame(df1['disease'])
+
+for symptom in all_symptoms:
+    df1_num[symptom] = df1[symptom_cols].apply(lambda row: int(symptom in row.values), axis=1)
+
+df1_num
+
+X = df1_num.drop('disease', axis=1)
+y = df1_num['disease']
+X.shape, y.shape
+
+X.sum(axis=1)
+
+X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, stratify=y)
+
+print(np.unique(y_train, return_counts=True))
+print(np.unique(y_test, return_counts=True))
+
+from sklearn.ensemble import  RandomForestClassifier
+
+model = RandomForestClassifier(n_estimators=100,random_state=42)
+model.fit(X_train, y_train)
+model.fit(X_train, y_train)
+
+import pickle
+
+# Save model
+with open("disease_model.pkl", "wb") as f:
+    pickle.dump(model, f)
+
+# Save symptom list (to use in the app later)
+with open("symptoms.pkl", "wb") as f:
+    pickle.dump(list(all_symptoms), f)
+
+# Original symptoms (keys)
+all_symptoms = sorted(all_symptoms)
+
+# Create display labels by replacing '_' with ' ' and capitalizing each word
+display_symptoms = [symptom.replace('_', ' ').title() for symptom in all_symptoms]
+
+# Create a mapping from display label back to original symptom key
+label_to_symptom = dict(zip(display_symptoms, all_symptoms))
+
+from sklearn.metrics import accuracy_score, f1_score
+
+y_train_pred = model.predict(X_train)
+
+train_accuracy = accuracy_score(y_train, y_train_pred)
+train_f1_score = f1_score(y_train, y_train_pred,average="weighted")
+
+print("Train Accuracy:", train_accuracy)
+print("Train f1 score:", train_f1_score)
+
+y_test_pred = model.predict(X_test)
+test_accuracy = accuracy_score(y_test, y_test_pred)
+test_f1_score = f1_score(y_test, y_test_pred, average="weighted")
+print("Train Accuracy:", test_accuracy)
+print("Train f1 score:", test_f1_score)
+
+import numpy as np
+
+# Example user symptoms
+user_symptoms = ['nausea', 'vomiting', 'abdominal_pain', 'diarrhoea']
+
+# Tip for the user
+if len(user_symptoms) < 4:
+    print("Tip: The model performs better if you enter at least 4 symptoms.\n")
+
+# Convert symptoms to input vector
+input_vector = [1 if symptom in user_symptoms else 0 for symptom in all_symptoms]
+input_vector = np.array([input_vector])
+
+# Make prediction and get probabilities
+probas = model.predict_proba(input_vector)[0]
+max_proba = np.max(probas)
+predicted = model.classes_[np.argmax(probas)]
+
+# Confidence threshold
+threshold = 0.5
+
+# Print predicted disease and confidence
+if max_proba < threshold:
+    print("Warning: The model is not confident about this prediction.")
+    print(f"Predicted disease: {predicted} (Confidence: {max_proba * 100:.1f}%)")
+else:
+    print(f"Predicted disease: {predicted} (Confidence: {max_proba * 100:.1f}%)")
+
+# Function to print top N diseases
+def print_top_diseases(probas, model, top_n=5):
+    classes = model.classes_
+    sorted_indices = np.argsort(probas)[::-1]
+    print(f"\nTop {top_n} possible diseases:")
+    for i in range(min(top_n, len(classes))):
+        disease = classes[sorted_indices[i]]
+        probability = probas[sorted_indices[i]]
+        print(f"{i+1}. {disease}: {probability:.4f}")
+
+# Show top 5 possible diseases
+print_top_diseases(probas, model, top_n=5)
+
+!cp "/content/drive/MyDrive/Google Colab/disease-symptom-prediction/tbc-logo (1).png" "/content/tbc.png"
+
 import gradio as gr
 import pickle
 import numpy as np
 
+# --- 1. Load Disease Prediction Model ---
+with open("disease_model.pkl", "rb") as f:
+    model = pickle.load(f)
+
+with open("symptoms.pkl", "rb") as f:
+    all_symptoms = pickle.load(f)
 
 # Preprocess symptoms
 all_symptoms = sorted(all_symptoms)