Create app.py

app.py

@@ -0,0 +1,90 @@
+import streamlit as st
+import pandas as pd
+import numpy as np
+from sklearn.feature_extraction.text import TfidfVectorizer
+from sklearn.model_selection import train_test_split
+from sklearn.ensemble import RandomForestClassifier
+from sklearn.metrics import accuracy_score
+import joblib
+
+# Title and Description of the App
+st.title("Human vs LLM-Generated Text Differentiator")
+st.write("This app predicts whether a given text is human-written or generated by a language model (LLM).")
+
+# Step 1: Upload Dataset
+st.header("Step 1: Upload the RoFT Dataset")
+uploaded_file = st.file_uploader("Upload your roft.csv file", type="csv")
+
+if uploaded_file is not None:
+    # Load dataset
+    data = pd.read_csv(uploaded_file)
+    st.write("Dataset Loaded Successfully!")
+
+    # Display the first few rows of the dataset
+    st.subheader("Sample of the Dataset:")
+    st.dataframe(data.head())
+
+    # Preprocessing the data
+    st.header("Step 2: Preprocess the Data")
+
+    # Combine prompt_body and gen_body to form the complete text
+    data['text'] = data['prompt_body'].fillna('') + ' ' + data['gen_body'].fillna('')
+    data['label'] = data['true_boundary_index'].apply(lambda x: 1 if x == 9 else 0)  # 1 = Human, 0 = LLM
+
+    st.write("Data Preprocessing Complete!")
+
+    # Show distribution of labels
+    st.subheader("Label Distribution:")
+    st.bar_chart(data['label'].value_counts())
+
+    # Feature Extraction
+    st.header("Step 3: Train the Model")
+    st.write("Extracting features using TF-IDF and training a Random Forest classifier.")
+
+    # TF-IDF Vectorization
+    vectorizer = TfidfVectorizer(max_features=5000)
+    X = vectorizer.fit_transform(data['text']).toarray()
+    y = data['label']
+
+    # Train-Test Split
+    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
+
+    # Train a Random Forest Classifier
+    model = RandomForestClassifier(n_estimators=100, random_state=42)
+    model.fit(X_train, y_train)
+
+    # Evaluate the model
+    y_pred = model.predict(X_test)
+    accuracy = accuracy_score(y_test, y_pred)
+    st.write(f"Model Accuracy: {accuracy * 100:.2f}%")
+
+    # Save the model and vectorizer
+    joblib.dump(model, 'text_classifier.pkl')
+    joblib.dump(vectorizer, 'vectorizer.pkl')
+    st.success("Model Trained and Saved Successfully!")
+
+    # Step 4: User Input for Prediction
+    st.header("Step 4: Predict Human vs LLM-Generated Text")
+
+    # Load the trained model and vectorizer
+    model = joblib.load('text_classifier.pkl')
+    vectorizer = joblib.load('vectorizer.pkl')
+
+    # Input text from the user
+    user_input = st.text_area("Enter the text you want to classify:")
+
+    if st.button("Predict"):
+        if user_input.strip():
+            # Vectorize the input text
+            input_vector = vectorizer.transform([user_input]).toarray()
+
+            # Predict and show the result
+            prediction = model.predict(input_vector)
+            confidence = model.predict_proba(input_vector).max() * 100
+
+            if prediction[0] == 1:
+                st.success(f"The text is likely **Human-Written** with a confidence of {confidence:.2f}%.")
+            else:
+                st.warning(f"The text is likely **LLM-Generated** with a confidence of {confidence:.2f}%.")
+        else:
+            st.error("Please enter some text for prediction.")