Alamgirapi's picture
Update app.py
8d810b6 verified
raw
history blame
26.4 kB
import streamlit as st
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
import seaborn as sns
from sklearn.feature_extraction.text import TfidfVectorizer, CountVectorizer
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
from sklearn.linear_model import LogisticRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import LinearSVC, SVC
from sklearn.naive_bayes import MultinomialNB, GaussianNB
from sklearn.metrics import accuracy_score, classification_report, confusion_matrix
import os
import pickle
import re
import string
from collections import Counter
import plotly.express as px
import plotly.graph_objects as go
# Configure Streamlit page
st.set_page_config(
page_title="Text Classification App",
page_icon="📝",
layout="wide"
)
# Text preprocessing class
class TextCleaner:
def __init__(self):
self.stop_words = set(['the', 'a', 'an', 'and', 'or', 'but', 'in', 'on', 'at', 'to', 'for', 'of', 'with', 'by'])
def clean_text(self, text):
"""Clean and preprocess text"""
if pd.isna(text):
return ""
text = str(text).lower()
text = re.sub(r'http\S+', '', text) # Remove URLs
text = re.sub(r'[^a-zA-Z\s]', '', text) # Remove non-alphabetic characters
text = re.sub(r'\s+', ' ', text) # Remove extra whitespace
text = text.strip()
# Remove stop words (optional)
words = text.split()
words = [word for word in words if word not in self.stop_words]
return ' '.join(words)
# Data analysis functions
def get_data_insights(df, text_col, target_col):
"""Get basic insights from the dataset"""
insights = {
'shape': df.shape,
'missing_values': df.isnull().sum().to_dict(),
'class_distribution': df[target_col].value_counts().to_dict(),
'text_length_stats': {
'mean': df[text_col].str.len().mean(),
'median': df[text_col].str.len().median(),
'min': df[text_col].str.len().min(),
'max': df[text_col].str.len().max()
}
}
return insights
# Model training functions
def train_model(model_name, X_train, X_test, y_train, y_test):
"""Train and evaluate a model"""
models = {
'Logistic Regression': LogisticRegression(random_state=42, max_iter=1000),
'Decision Tree': DecisionTreeClassifier(random_state=42),
'Random Forest': RandomForestClassifier(random_state=42, n_estimators=100),
'Linear SVC': LinearSVC(random_state=42, max_iter=1000),
'SVC': SVC(random_state=42, probability=True),
'Multinomial Naive Bayes': MultinomialNB(),
'Gaussian Naive Bayes': GaussianNB()
}
model = models[model_name]
# For Gaussian NB, convert sparse matrix to dense
if model_name == 'Gaussian Naive Bayes':
X_train = X_train.toarray()
X_test = X_test.toarray()
# Train model
model.fit(X_train, y_train)
# Make predictions
y_pred = model.predict(X_test)
# Calculate metrics
accuracy = accuracy_score(y_test, y_pred)
# Save model
os.makedirs("models", exist_ok=True)
model_filename = f"{model_name.replace(' ', '_').lower()}.pkl"
with open(os.path.join("models", model_filename), 'wb') as f:
pickle.dump(model, f)
return model, accuracy, y_pred, model_filename
# Utility functions
def save_artifacts(obj, folder_name, file_name):
"""Save artifacts like encoders and vectorizers"""
os.makedirs(folder_name, exist_ok=True)
with open(os.path.join(folder_name, file_name), 'wb') as f:
pickle.dump(obj, f)
def load_artifacts(folder_name, file_name):
"""Load saved artifacts"""
try:
with open(os.path.join(folder_name, file_name), 'rb') as f:
return pickle.load(f)
except FileNotFoundError:
st.error(f"File {file_name} not found in {folder_name} folder")
return None
def predict_text(model_filename, text, vectorizer_type="tfidf"):
"""Make prediction on new text"""
try:
# Load model
with open(os.path.join('models', model_filename), 'rb') as f:
model = pickle.load(f)
# Load vectorizer
vectorizer_file = f"{vectorizer_type}_vectorizer.pkl"
vectorizer = load_artifacts("artifacts", vectorizer_file)
if vectorizer is None:
return None, None
# Load label encoder
encoder = load_artifacts("artifacts", "encoder.pkl")
if encoder is None:
return None, None
# Clean and vectorize text
text_cleaner = TextCleaner()
clean_text = text_cleaner.clean_text(text)
# Transform text
text_vector = vectorizer.transform([clean_text])
# For Gaussian NB, convert to dense
if 'gaussian' in model_filename:
text_vector = text_vector.toarray()
# Make prediction
prediction = model.predict(text_vector)
prediction_proba = None
# Get prediction probabilities if available
if hasattr(model, 'predict_proba'):
try:
prediction_proba = model.predict_proba(text_vector)[0]
except:
pass
# Decode prediction
predicted_label = encoder.inverse_transform(prediction)[0]
return predicted_label, prediction_proba
except Exception as e:
st.error(f"Error during prediction: {str(e)}")
return None, None
# Streamlit App
st.title('📝 No Code Text Classification App')
st.markdown('---')
st.write('Analyze your text data and train machine learning models without coding!')
# Sidebar
st.sidebar.title("Navigation")
section = st.sidebar.radio("Choose Section", ["📊 Data Analysis", "🤖 Train Model", "🔮 Predictions"])
# Upload Data
st.sidebar.markdown("---")
st.sidebar.subheader("📁 Upload Your Dataset")
train_data = st.sidebar.file_uploader("Upload training data", type=["csv"])
test_data = st.sidebar.file_uploader("Upload test data (optional)", type=["csv"])
# Global variables to store data and settings
if 'vectorizer_type' not in st.session_state:
st.session_state.vectorizer_type = "tfidf"
if train_data is not None:
try:
# Try different encodings
encodings = ['utf-8', 'latin1', 'cp1252', 'iso-8859-1']
train_df = None
for encoding in encodings:
try:
train_df = pd.read_csv(train_data, encoding=encoding)
break
except UnicodeDecodeError:
continue
if train_df is None:
st.error("Unable to read the CSV file. Please check the file encoding.")
else:
if test_data is not None:
for encoding in encodings:
try:
test_df = pd.read_csv(test_data, encoding=encoding)
break
except UnicodeDecodeError:
continue
else:
test_df = None
# Show data preview
with st.sidebar.expander("📋 Data Preview", expanded=True):
st.write("Shape:", train_df.shape)
st.write(train_df.head(2))
columns = train_df.columns.tolist()
text_data = st.sidebar.selectbox("📝 Choose the text column:", columns)
target = st.sidebar.selectbox("🎯 Choose the target column:", columns)
# Process data
if text_data and target:
# Clean text
text_cleaner = TextCleaner()
train_df['clean_text'] = train_df[text_data].apply(text_cleaner.clean_text)
train_df['text_length'] = train_df[text_data].str.len()
# Handle label encoding
label_encoder = LabelEncoder()
train_df['target_encoded'] = label_encoder.fit_transform(train_df[target])
# Save label encoder
save_artifacts(label_encoder, "artifacts", "encoder.pkl")
except Exception as e:
st.error(f"Error loading data: {str(e)}")
train_df = None
# Data Analysis Section
if section == "📊 Data Analysis":
if train_data is not None and 'train_df' in locals() and train_df is not None:
st.header("📊 Data Analysis")
# Get insights
insights = get_data_insights(train_df, text_data, target)
# Display insights in columns
col1, col2, col3, col4 = st.columns(4)
with col1:
st.metric("Total Samples", insights['shape'][0])
with col2:
st.metric("Features", insights['shape'][1])
with col3:
st.metric("Classes", len(insights['class_distribution']))
with col4:
st.metric("Avg Text Length", f"{insights['text_length_stats']['mean']:.1f}")
st.markdown("---")
# Data quality section
col1, col2 = st.columns(2)
with col1:
st.subheader("📋 Dataset Overview")
st.write("**Shape:**", insights['shape'])
st.write("**Missing Values:**")
missing_df = pd.DataFrame.from_dict(insights['missing_values'], orient='index', columns=['Count'])
st.dataframe(missing_df[missing_df['Count'] > 0])
st.write("**Sample Data:**")
st.dataframe(train_df[[text_data, target, 'text_length']].head())
with col2:
st.subheader("📊 Class Distribution")
class_dist = pd.DataFrame.from_dict(insights['class_distribution'], orient='index', columns=['Count'])
st.dataframe(class_dist)
# Plot class distribution
fig = px.bar(
x=class_dist.index,
y=class_dist['Count'],
title="Class Distribution",
labels={'x': 'Class', 'y': 'Count'}
)
st.plotly_chart(fig, use_container_width=True)
st.markdown("---")
# Text analysis section
st.subheader("📝 Text Analysis")
col1, col2 = st.columns(2)
with col1:
# Text length distribution
fig = px.histogram(
train_df,
x='text_length',
title="Text Length Distribution",
nbins=30
)
st.plotly_chart(fig, use_container_width=True)
with col2:
# Text length by class
fig = px.box(
train_df,
x=target,
y='text_length',
title="Text Length by Class"
)
st.plotly_chart(fig, use_container_width=True)
# Word frequency analysis
st.subheader("🔤 Most Common Words")
all_text = ' '.join(train_df['clean_text'].astype(str))
word_freq = Counter(all_text.split())
top_words = word_freq.most_common(20)
if top_words:
words_df = pd.DataFrame(top_words, columns=['Word', 'Frequency'])
fig = px.bar(
words_df,
x='Frequency',
y='Word',
orientation='h',
title="Top 20 Most Common Words"
)
fig.update_layout(yaxis={'categoryorder': 'total ascending'})
st.plotly_chart(fig, use_container_width=True)
else:
st.warning("📁 Please upload training data to perform analysis")
# Train Model Section
elif section == "🤖 Train Model":
if train_data is not None and 'train_df' in locals() and train_df is not None:
st.header("🤖 Train Machine Learning Model")
col1, col2 = st.columns(2)
with col1:
st.subheader("⚙️ Model Configuration")
model_name = st.selectbox("Choose Model", [
"Logistic Regression", "Decision Tree",
"Random Forest", "Linear SVC", "SVC",
"Multinomial Naive Bayes", "Gaussian Naive Bayes"
])
with col2:
st.subheader("📊 Vectorization Method")
vectorizer_choice = st.selectbox("Choose Vectorizer", ["TF-IDF", "Count Vectorizer"])
# Model parameters
st.subheader("🔧 Parameters")
col1, col2 = st.columns(2)
with col1:
max_features = st.slider("Max Features", 1000, 20000, 10000, step=1000)
test_size = st.slider("Test Size", 0.1, 0.4, 0.2, step=0.05)
with col2:
random_state = st.number_input("Random State", 0, 1000, 42)
min_df = st.slider("Min Document Frequency", 1, 10, 1)
# Initialize vectorizer
if vectorizer_choice == "TF-IDF":
vectorizer = TfidfVectorizer(
max_features=max_features,
min_df=min_df,
stop_words='english'
)
st.session_state.vectorizer_type = "tfidf"
else:
vectorizer = CountVectorizer(
max_features=max_features,
min_df=min_df,
stop_words='english'
)
st.session_state.vectorizer_type = "count"
# Show data info
st.subheader("📋 Training Data Info")
col1, col2, col3 = st.columns(3)
with col1:
st.metric("Total Samples", len(train_df))
with col2:
st.metric("Unique Classes", train_df[target].nunique())
with col3:
st.metric("Avg Text Length", f"{train_df['text_length'].mean():.1f}")
if st.button("🚀 Start Training", type="primary"):
with st.spinner("Training model... This may take a few minutes."):
try:
# Vectorize text data
X = vectorizer.fit_transform(train_df['clean_text'])
y = train_df['target_encoded']
# Split data
X_train, X_test, y_train, y_test = train_test_split(
X, y,
test_size=test_size,
random_state=random_state,
stratify=y
)
st.success(f"✅ Data split - Train: {X_train.shape}, Test: {X_test.shape}")
# Save vectorizer
vectorizer_filename = f"{st.session_state.vectorizer_type}_vectorizer.pkl"
save_artifacts(vectorizer, "artifacts", vectorizer_filename)
# Train model
model, accuracy, y_pred, model_filename = train_model(
model_name, X_train, X_test, y_train, y_test
)
st.success("🎉 Model training completed!")
# Display results
col1, col2 = st.columns(2)
with col1:
st.metric("🎯 Test Accuracy", f"{accuracy:.4f}")
# Classification report
st.subheader("📊 Classification Report")
report = classification_report(
y_test, y_pred,
target_names=label_encoder.classes_,
output_dict=True
)
report_df = pd.DataFrame(report).transpose()
st.dataframe(report_df.round(4))
with col2:
# Confusion matrix
st.subheader("🔄 Confusion Matrix")
cm = confusion_matrix(y_test, y_pred)
fig = px.imshow(
cm,
text_auto=True,
aspect="auto",
title="Confusion Matrix",
labels=dict(x="Predicted", y="Actual"),
x=label_encoder.classes_,
y=label_encoder.classes_
)
st.plotly_chart(fig, use_container_width=True)
st.info(f"✅ Model saved as: {model_filename}")
st.info("🔮 You can now use the 'Predictions' section to classify new text!")
except Exception as e:
st.error(f"❌ Error during training: {str(e)}")
else:
st.warning("📁 Please upload training data to train a model")
# Predictions Section
elif section == "🔮 Predictions":
st.header("🔮 Text Classification Predictions")
# Check if models exist
if os.path.exists("models") and os.listdir("models"):
available_models = [f for f in os.listdir("models") if f.endswith('.pkl')]
if available_models:
# Single prediction
st.subheader("📝 Single Text Classification")
col1, col2 = st.columns([2, 1])
with col1:
text_input = st.text_area("Enter text to classify:", height=150)
with col2:
selected_model = st.selectbox("Choose model:", available_models)
predict_button = st.button("🔮 Predict", type="primary")
if predict_button and text_input.strip():
with st.spinner("Making prediction..."):
predicted_label, prediction_proba = predict_text(
selected_model,
text_input,
st.session_state.get('vectorizer_type', 'tfidf')
)
if predicted_label is not None:
st.success("✅ Prediction completed!")
# Display results
col1, col2 = st.columns(2)
with col1:
st.markdown("### 🎯 Results")
st.markdown(f"**Input Text:** {text_input[:200]}{'...' if len(text_input) > 200 else ''}")
st.markdown(f"**Predicted Class:** `{predicted_label}`")
with col2:
# Display probabilities if available
if prediction_proba is not None:
st.markdown("### 📊 Class Probabilities")
encoder = load_artifacts("artifacts", "encoder.pkl")
if encoder is not None:
prob_df = pd.DataFrame({
'Class': encoder.classes_,
'Probability': prediction_proba
}).sort_values('Probability', ascending=False)
fig = px.bar(
prob_df,
x='Probability',
y='Class',
orientation='h',
title="Prediction Confidence"
)
fig.update_layout(yaxis={'categoryorder': 'total ascending'})
st.plotly_chart(fig, use_container_width=True)
elif predict_button:
st.warning("⚠️ Please enter some text to classify")
# Batch predictions
st.markdown("---")
st.subheader("📊 Batch Predictions")
uploaded_file = st.file_uploader("Upload CSV file with texts to classify", type=['csv'])
if uploaded_file is not None:
try:
# Try different encodings for batch file
encodings = ['utf-8', 'latin1', 'cp1252', 'iso-8859-1']
batch_df = None
for encoding in encodings:
try:
batch_df = pd.read_csv(uploaded_file, encoding=encoding)
break
except UnicodeDecodeError:
continue
if batch_df is not None:
st.write("📋 Uploaded data preview:")
st.dataframe(batch_df.head())
col1, col2 = st.columns(2)
with col1:
text_column = st.selectbox("Select text column:", batch_df.columns.tolist())
with col2:
batch_model = st.selectbox("Choose model:", available_models, key="batch_model")
if st.button("🚀 Run Batch Predictions", type="primary"):
with st.spinner("Processing batch predictions..."):
predictions = []
confidences = []
progress_bar = st.progress(0)
total_texts = len(batch_df)
for i, text in enumerate(batch_df[text_column]):
pred, proba = predict_text(
batch_model,
str(text),
st.session_state.get('vectorizer_type', 'tfidf')
)
predictions.append(pred if pred is not None else "Error")
# Get confidence (max probability)
if proba is not None:
confidences.append(max(proba))
else:
confidences.append(0.0)
progress_bar.progress((i + 1) / total_texts)
batch_df['Predicted_Class'] = predictions
batch_df['Confidence'] = confidences
st.success("✅ Batch predictions completed!")
# Show results
st.subheader("📊 Results")
result_df = batch_df[[text_column, 'Predicted_Class', 'Confidence']]
st.dataframe(result_df)
# Summary statistics
st.subheader("📈 Summary")
col1, col2, col3 = st.columns(3)
with col1:
st.metric("Total Predictions", len(predictions))
with col2:
successful_preds = sum(1 for p in predictions if p != "Error")
st.metric("Successful", successful_preds)
with col3:
avg_confidence = sum(confidences) / len(confidences) if confidences else 0
st.metric("Avg Confidence", f"{avg_confidence:.3f}")
# Class distribution of predictions
pred_counts = pd.Series(predictions).value_counts()
if len(pred_counts) > 0:
fig = px.pie(
values=pred_counts.values,
names=pred_counts.index,
title="Distribution of Predictions"
)
st.plotly_chart(fig, use_container_width=True)
# Download results
csv = batch_df.to_csv(index=False)
st.download_button(
label="📥 Download Results as CSV",
data=csv,
file_name="batch_predictions.csv",
mime="text/csv"
)
else:
st.error("❌ Unable to read the CSV file. Please check the file encoding.")
except Exception as e:
st.error(f"❌ Error in batch prediction: {str(e)}")
else:
st.warning("⚠️ No trained models found. Please train a model first.")
else:
st.warning("⚠️ No models directory found. Please go to 'Train Model' section to train a model first.")
# Footer
st.markdown("---")
st.markdown("🚀 Built with Streamlit | 📊 No-Code Text Classification")