"""
Competition baseline kit - text classification / regression (CPU only, free).
Usage:
    python baseline_text.py --train train.csv --test test.csv --target label --out submissions/predictions.csv

Train format: CSV with text column 'text' and target column (name passed via --target).
Test format: same CSV WITHOUT the target column. Output: CSV with 'prediction' column.
"""
import argparse, pandas as pd
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import make_pipeline

p = argparse.ArgumentParser()
p.add_argument("--train", required=True)
p.add_argument("--test", required=True)
p.add_argument("--target", default="label")
p.add_argument("--text-col", default="text")
p.add_argument("--out", default="submissions/predictions.csv")
args = p.parse_args()

train = pd.read_csv(args.train)
test = pd.read_csv(args.test)
X = train[args.text_col].fillna("")
y = train[args.target]
X_test = test[args.text_col].fillna("")

clf = make_pipeline(
    TfidfVectorizer(max_features=50_000, ngram_range=(1, 2), sublinear_tf=True),
    LogisticRegression(max_iter=2000, C=1.0),
)
clf.fit(X, y)
pred = clf.predict(X_test)

import os
os.makedirs(os.path.dirname(args.out) or ".", exist_ok=True)
pd.DataFrame({"prediction": pred}).to_csv(args.out, index=False)
print(f"baseline done -> {args.out} | classes: {clf.classes_}")