model_setup.py

import pandas as pd
import re
import joblib
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.preprocessing import MultiLabelBinarizer
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import MultinomialNB
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import LinearSVC
from sklearn.multiclass import OneVsRestClassifier
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score

# Load and clean the dataset
df = pd.read_csv("all_tweets.csv")

def clean_text(text):
    text = re.sub(r"http\S+", "", text)
    text = re.sub(r"@\w+", "", text)
    text = re.sub(r"#\w+", "", text)
    text = re.sub(r"[^a-zA-Z\s]", "", text.lower())
    return text.strip()

df["clean_text"] = df["text"].apply(clean_text)
df["hashtag_list"] = df["hashtags"].apply(lambda x: x.lower().split())

# Convert features and labels
vectorizer = TfidfVectorizer(max_features=3000)
X = vectorizer.fit_transform(df["clean_text"])

mlb = MultiLabelBinarizer()
Y = mlb.fit_transform(df["hashtag_list"])

# Save the vectorizer and label binarizer
joblib.dump(vectorizer, "vectorizer.pkl")
joblib.dump(mlb, "label_binarizer.pkl")

# Split the dataset
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2, random_state=42)

# Define models to compare
models = {
    "Logistic Regression": LogisticRegression(max_iter=1000),
    "Random Forest": RandomForestClassifier(n_estimators=100),
    "Multinomial NB": MultinomialNB(),
    "Linear SVC": LinearSVC()
}

# Train and evaluate each model
results = []

for name, base_model in models.items():
    print(f"🧪 Training {name}...")
    model = OneVsRestClassifier(base_model)
    model.fit(X_train, Y_train)
    Y_pred = model.predict(X_test)

    # Save the trained model
    model_filename = f"{name.replace(' ', '_').lower()}_model.pkl"
    joblib.dump(model, model_filename)

    accuracy = accuracy_score(Y_test, Y_pred)
    precision = precision_score(Y_test, Y_pred, average='micro', zero_division=0)
    recall = recall_score(Y_test, Y_pred, average='micro', zero_division=0)
    f1 = f1_score(Y_test, Y_pred, average='micro', zero_division=0)

    results.append({
        "Model": name,
        "Accuracy": accuracy,
        "Precision": precision,
        "Recall": recall,
        "F1 Score": f1
    })

# Display comparison
results_df = pd.DataFrame(results)
print("\n📊 Model Performance Comparison:")
print(results_df.sort_values(by="F1 Score", ascending=False).reset_index(drop=True))