{"cells": [{"cell_type": "markdown", "metadata": {}, "source": "# CAFA6 Submit Conservative\n\nSubmits the output from `cafa6-pure-goa-conservative` notebook.\n\n**Strategy**: Pure GOA + higher threshold + term limit = better precision."}, {"cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": "import os\nimport shutil\n\nINPUT_PATH = '/kaggle/input/cafa6-pure-goa-conservative/submission.tsv'\nOUTPUT_PATH = 'submission.tsv'\n\nif os.path.exists(INPUT_PATH):\n    shutil.copy(INPUT_PATH, OUTPUT_PATH)\n    size = os.path.getsize(OUTPUT_PATH)\n    print(f\"SUCCESS: Copied {size:,} bytes\")\nelse:\n    print(f\"ERROR: {INPUT_PATH} not found!\")\n    for root, dirs, files in os.walk('/kaggle/input/'):\n        for f in files:\n            print(os.path.join(root, f))"}, {"cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": "if os.path.exists(OUTPUT_PATH):\n    import pandas as pd\n    df = pd.read_csv(OUTPUT_PATH, sep='\\t', header=None, names=['protein', 'term', 'score'])\n    print(f\"Predictions: {len(df):,}\")\n    print(f\"Proteins: {df['protein'].nunique():,}\")\n    print(f\"Avg terms/protein: {len(df)/df['protein'].nunique():.1f}\")"}], "metadata": {"kernelspec": {"display_name": "Python 3", "language": "python", "name": "python3"}}, "nbformat": 4, "nbformat_minor": 5}