joshnavip/ai-code-detection

1import pandas as pd2 3DATA_PATH = "dataset/processed/dataset_step2_cleaned.csv"4 5def main():6    df = pd.read_csv(DATA_PATH)7 8    # -------------------------------9    # Handle label column safely10    # -------------------------------11    if "Label" not in df.columns:12        if "Label (0- HUMAN, 1-AI)" in df.columns:13            df = df.rename(columns={"Label (0- HUMAN, 1-AI)": "Label"})14        else:15            raise ValueError("Label column not found in dataset.")16 17    # -------------------------------18    # Column validation19    # -------------------------------20    required_cols = ["Label", "Language"]21    missing = [c for c in required_cols if c not in df.columns]22    if missing:23        raise ValueError(f"Missing required columns: {missing}")24 25    total = len(df)26 27    print("\n📊 OVERALL CLASS BALANCE\n")28    class_counts = df["Label"].value_counts().sort_index()29 30    for label, count in class_counts.items():31        percent = (count / total) * 10032        class_name = "Human" if label == 0 else "AI"33        print(f"{class_name} ({label}): {count} samples ({percent:.2f}%)")34 35    print("\nTotal samples:", total)36 37    # -------------------------------38    # Language-wise breakdown39    # -------------------------------40    print("\n🌐 LANGUAGE-WISE CLASS DISTRIBUTION\n")41 42    for lang in df["Language"].unique():43        print(f"Language: {lang}")44        lang_df = df[df["Language"] == lang]45        lang_total = len(lang_df)46 47        lang_counts = lang_df["Label"].value_counts().sort_index()48        for label, count in lang_counts.items():49            percent = (count / lang_total) * 10050            class_name = "Human" if label == 0 else "AI"51            print(f"  {class_name} ({label}): {count} samples ({percent:.2f}%)")52 53        print(f"  Total {lang} samples: {lang_total}\n")54 55 56if __name__ == "__main__":57    main()58