# train.py

import pandas as pd
import joblib
import numpy as np
from sklearn.linear_model import LogisticRegression
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.multiclass import OneVsRestClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import f1_score

LABELS = ['admiration','anger','disgust','fear','hope','joy','love','pride','sadness']

def to_binary(label_string):
    present = [e.strip() for e in str(label_string).split(',')]
    return [1 if label in present else 0 for label in LABELS]

# ── load ──────────────────────────────────────────────────────────
print("Loading dataset...")
df = pd.read_excel("Multi-Labeled_Emotions_Modified.xlsx")
df = df[['Tweets (text)', 'Emotions (Multi-labeled)']].dropna()
print(f"Total rows: {len(df)}")

X = df['Tweets (text)'].tolist()
y = [to_binary(row) for row in df['Emotions (Multi-labeled)']]

# ── split ─────────────────────────────────────────────────────────
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)
print(f"Train: {len(X_train)} rows")
print(f"Test:  {len(X_test)} rows")

# ── save test set as hidden test data ─────────────────────────────
test_emotions = [
    ', '.join([LABELS[i] for i, val in enumerate(row) if val == 1])
    for row in y_test
]
test_df = pd.DataFrame({
    'Tweets (text)': X_test,
    'Emotions (Multi-labeled)': test_emotions
})
test_df.to_excel("test_set.xlsx", index=False)
print("Saved test_set.xlsx")

# ── train ─────────────────────────────────────────────────────────
print("Training...")
vectorizer = TfidfVectorizer(max_features=10000, ngram_range=(1,2))
X_train_tfidf = vectorizer.fit_transform(X_train)
X_test_tfidf  = vectorizer.transform(X_test)

classifier = OneVsRestClassifier(
    LogisticRegression(max_iter=1000, C=1.0)
)
classifier.fit(X_train_tfidf, np.array(y_train))
print("Training done.")

# ── quick check ───────────────────────────────────────────────────
y_pred = classifier.predict(X_test_tfidf)
f1 = f1_score(np.array(y_test), y_pred, average='macro', zero_division=0)
print(f"F1 score: {f1:.4f}")

# ── save ──────────────────────────────────────────────────────────
model_bundle = {
    "vectorizer": vectorizer,
    "classifier": classifier,
    "labels": LABELS
}
joblib.dump(model_bundle, "model.pkl")
print("Saved model.pkl")