{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.14","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"none","dataSources":[{"sourceId":84493,"databundleVersionId":9871156,"sourceType":"competition"},{"sourceId":203900450,"sourceType":"kernelVersion"},{"sourceId":211920287,"sourceType":"kernelVersion"}],"dockerImageVersionId":30786,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Inference\n\nIn this notebook I've experimented an approach to do online retraining of a simple lightgbm model.\n\n- EDA: https://www.kaggle.com/code/simonedegasperis/starter-eda\n- train: https://www.kaggle.com/code/simonedegasperis/lgbm-model-training\n\nKudos to https://www.kaggle.com/code/motono0223/js24-preprocessing-create-lags for historical data with added lags feature that I've used to train the initial version of the model.\n\nFor retraining I've sampled from each date of historical data a random 1% of the data and I've added online data. I tried to retrain the model each N days by gradually increasing the cache which was defined as global variable and I've then averaged the solution of the initial model and the new retrained model.\n\nThe purpose of choosing a small fraction of data for retraining was to try to stay within 10 minutes limit between 2 consecutive batches.\nI hope you will find the solution helpfull to build a better model.\n\nIn the following I present a schematic view of the idea\n\n![image.png](attachment:fa7b2a6d-90d0-4d8e-be68-47f9d313a67c.png)\n\nNote: While this current implentation does not improve the score of the single model, it may be a good starting point to implement some solution based on online retrain of solutions that want ot leverage somehow the data online.","metadata":{},"attachments":{"fa7b2a6d-90d0-4d8e-be68-47f9d313a67c.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"# Changelog\n\n**v18**: Remove nan labels from training\n\n**v17**: Fix notebook timeout\n\n**v16**: Reduce cache size\n\n**v15**: Fix simulator so lags and test dataframe match the same date_id\n\n**v14**: Increase cache size and retrain start day\n\n**v13**: Improve retraining strategy. Use all lags from previous day as ground truth and not only those correspoding to the the last time id of the previous day.\n         Improve debugging by simulating the submission api\n\n**v12**: Empty cache list after cache update\n\n**v11**: fix reset day count after each retrain, store only last 2 lags and update cache only when retraining to reduce processing time","metadata":{}},{"cell_type":"markdown","source":"# Libraries","metadata":{}},{"cell_type":"code","source":"# imports\nimport os\nimport glob\nimport numpy as np\nimport pandas as pd\nimport polars as pl\nimport lightgbm as lgb\nimport xgboost as xgb\nimport pickle\nimport kaggle_evaluation.jane_street_inference_server","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:15:59.484187Z","iopub.execute_input":"2025-01-09T22:15:59.484661Z","iopub.status.idle":"2025-01-09T22:16:04.756039Z","shell.execute_reply.started":"2025-01-09T22:15:59.484582Z","shell.execute_reply":"2025-01-09T22:16:04.754755Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import torch\nimport torch.nn as nn\nimport torch.nn.functional as F\nimport tqdm\nfrom pytorch_lightning import (LightningDataModule, LightningModule, Trainer)\nfrom pytorch_lightning.callbacks import EarlyStopping, ModelCheckpoint, Timer","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:16:04.759594Z","iopub.execute_input":"2025-01-09T22:16:04.761028Z","iopub.status.idle":"2025-01-09T22:16:12.628757Z","shell.execute_reply.started":"2025-01-09T22:16:04.760968Z","shell.execute_reply":"2025-01-09T22:16:12.627598Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Configs","metadata":{}},{"cell_type":"code","source":"class CONFIG:\n    seed = 42\n    target_col = \"responder_6\"\n    all_cols = [\"date_id\", \"symbol_id\", \"time_id\", \"weight\"] + [f\"feature_{idx:02d}\" for idx in range(79)]+ [f\"responder_{idx}_lag_1\" for idx in range(9)] + [target_col]\n    test_cols = [\"row_id\", \"date_id\", \"symbol_id\", \"time_id\"] + [f\"feature_{idx:02d}\" for idx in range(79)]+ [f\"responder_{idx}_lag_1\" for idx in range(9)] + [target_col]\n    feature_cols = [\"symbol_id\", \"time_id\"] + [f\"feature_{idx:02d}\" for idx in range(79)]+ [f\"responder_{idx}_lag_1\" for idx in range(9)]\n    lag_cols_rename = { f\"responder_{idx}\" : f\"responder_{idx}_lag_1\" for idx in range(9)}\n    only_features = [\"row_id\", \"date_id\", \"symbol_id\", \"time_id\"] + [f\"feature_{idx:02d}\" for idx in range(79)]\n    only_lags = [\"row_id\", \"date_id\", \"symbol_id\", \"time_id\"] + [f\"responder_{idx}_lag_1\" for idx in range(9)] \n    data_paths = [\"/kaggle/input/lgbm-model-training/lgbm_model_0.json\",\"/kaggle/input/js24-preprocessing-create-lags/validation.parquet/\"]\n    retrain = True\n    EVAL = False","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:16:12.634470Z","iopub.execute_input":"2025-01-09T22:16:12.634843Z","iopub.status.idle":"2025-01-09T22:16:12.643621Z","shell.execute_reply.started":"2025-01-09T22:16:12.634808Z","shell.execute_reply":"2025-01-09T22:16:12.642457Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Load model","metadata":{}},{"cell_type":"code","source":"# load model\nlgbm_model = lgb.Booster(model_file=CONFIG.data_paths[0])","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:16:12.645689Z","iopub.execute_input":"2025-01-09T22:16:12.646131Z","iopub.status.idle":"2025-01-09T22:16:12.737257Z","shell.execute_reply.started":"2025-01-09T22:16:12.646085Z","shell.execute_reply":"2025-01-09T22:16:12.735960Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Params used to retrain\ninput_params = {\"num_leaves\": 31, \"feature_fraction\": 0.9, \"n_estimators\": 100, \"learning_rate\": 0.1}\n\n# Define Parameters\nparams = {\n    'objective': 'regression',\n    'metric': 'rmse',                                      # Root Mean Squared Error\n    'boosting_type': 'gbdt',                               # Gradient Boosted Decision Trees\n    'num_leaves': input_params['num_leaves'],\n    'learning_rate': input_params['learning_rate'],\n    'feature_fraction': input_params['feature_fraction'],\n    'n_estimators': input_params['n_estimators']      \n}","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:16:12.738956Z","iopub.execute_input":"2025-01-09T22:16:12.739353Z","iopub.status.idle":"2025-01-09T22:16:12.745923Z","shell.execute_reply.started":"2025-01-09T22:16:12.739315Z","shell.execute_reply":"2025-01-09T22:16:12.744723Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Inference","metadata":{}},{"cell_type":"code","source":"# Initialize global vars\ncache = None\ncache_list = []\n# tot nb of days counter\nday_count = 0\n# training counter to be reset after each train\ntrain_counter = 0\nnew_lgbm_model = None\nlags_ : pl.DataFrame | None = None\nlabels : pl.DataFrame | None = None\n# Each batch of predictions (except the very first) must be returned within 1 minute of the batch features being provided.\ndef predict(test: pl.DataFrame, lags: pl.DataFrame | None) -> pl.DataFrame | pd.DataFrame:\n    \"\"\"Make a prediction.\"\"\"\n    # All the responders from the previous day are passed in at time_id == 0. We save them in a global variable for access at every time_id.\n    # Use them as extra features, if you like.\n    global cache          # Declare the global cache\n    global day_count\n    global new_lgbm_model\n    global lags_\n    global cache_list\n    global labels\n    global train_counter\n    global gt\n\n    # Store lags, since they are provided at the start of each day but no for other tiòe_id of the same day\n    if lags is not None:\n        # print(f\"Day count: {day_count}\")\n        lags_ = lags\n        day_count += 1\n        train_counter += 1\n        # store ground truth from previous day\n        update_labels = lags_[\"date_id\", \"symbol_id\", \"time_id\",\"responder_6_lag_1\"]\n        lag_cols_rename = {\"responder_6_lag_1\": \"responder_6\"}\n        update_labels = update_labels.rename(lag_cols_rename)\n        if labels is not None:\n            labels = pl.concat([labels, update_labels], rechunk=True)\n        else:\n            labels = update_labels\n\n    # Init prediction\n    predictions = test.select(\n        'row_id',\n        pl.lit(0.0).alias('responder_6'),\n    )\n\n    if not lags_ is None:\n        lags = lags_.group_by([\"date_id\", \"symbol_id\"], maintain_order=True).last() # pick up last record of previous date\n        lags = lags.drop([\"time_id\"])\n        test = test.join(lags, on=[\"date_id\", \"symbol_id\"],  how=\"left\")\n    else:\n        test = test.with_columns(\n            ( pl.lit(0.0).alias(f'responder_{idx}_lag_1') for idx in range(9) )\n        )\n\n    if CONFIG.retrain:\n        # store data for each batch\n        cache_list.append(test)\n\n    # initialize preds\n    preds = np.zeros((test.shape[0],))\n\n    # lightgbm model\n    X = test[CONFIG.feature_cols].to_numpy()\n  \n    # re-train a model on the fly every N days\n    if CONFIG.retrain and train_counter % 4 == 0 and day_count>=60:\n        # print(\"Start retraining\")\n        if cache is not None:\n            cache_update = pl.concat(cache_list, rechunk=True)\n            cache = pl.concat([cache, cache_update], rechunk=True)\n        else:\n            cache = pl.concat(cache_list, rechunk=True)\n\n        # filter labels\n        # move data back to the previous day (we receive the lags at the same day but they are the ground truth of the previous day)\n        df = labels.with_columns(\n            (pl.col(\"date_id\") -1).alias(\"date_id\")\n        )\n        df = df.filter(pl.col(\"date_id\") >= np.min(cache[\"date_id\"].to_numpy()))\n        \n        # prepare data for training\n        train = cache.join(df, on=[\"date_id\", \"symbol_id\", \"time_id\"],  how=\"left\")\n\n        # print(f\"Shape x_train {train[CONFIG.feature_cols].shape}\")\n        # print(f\"Shape y_train {train[CONFIG.target_col].shape}\")\n\n        # drop columns where labels are none (normally last day)\n        train_cleaned = train.filter(pl.col(CONFIG.target_col).is_not_nan())\n        \n        X_train = train_cleaned[CONFIG.feature_cols].to_numpy()\n        y_train = train_cleaned[CONFIG.target_col].to_numpy().flatten()\n\n        train_data = lgb.Dataset(X_train, label=y_train)\n\n        # Re-train the model\n        new_lgbm_model = lgb.train(\n            params,\n            train_data,\n            num_boost_round=40\n        )\n        # reset counter otherwise we will retrain for each time_id of the same day\n        train_counter = 1\n        # empty cache list\n        cache_list = []\n\n        # store only last 50 days\n        days = np.unique(cache[\"date_id\"].to_numpy())\n        days = days[-40:]\n        min_day = np.min(days)\n        cache = cache.filter(pl.col(\"date_id\") >= min_day)\n\n    # average original model with new retrained model\n    if new_lgbm_model:\n        # lightgbm models\n        y_pred1 = new_lgbm_model.predict(X, num_iteration=lgbm_model.best_iteration)\n        y_pred2 = lgbm_model.predict(X, num_iteration=lgbm_model.best_iteration)\n        # simple average\n        # weight more the prediction from the new model retrained on the new data\n        y_pred = (0.6*y_pred1+0.4*y_pred2)\n    else:\n        # lightgbm model\n        y_pred = lgbm_model.predict(X, num_iteration=lgbm_model.best_iteration)\n\n    preds = y_pred\n    # print(f\"predict> preds.shape =\", preds.shape)\n    \n    predictions = \\\n    test.select('row_id').\\\n    with_columns(\n        pl.Series(\n            name   = 'responder_6', \n            values = np.clip(preds, a_min = -5, a_max = 5),\n            dtype  = pl.Float64,\n        )\n    )\n\n    if isinstance(predictions, pl.DataFrame):\n        assert predictions.columns == ['row_id', 'responder_6']\n    elif isinstance(predictions, pd.DataFrame):\n        assert (predictions.columns == ['row_id', 'responder_6']).all()\n    else:\n        raise TypeError('The predict function must return a DataFrame')\n    # Confirm has as many rows as the test data.\n    assert len(predictions) == len(test)\n\n    return predictions","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:16:12.749366Z","iopub.execute_input":"2025-01-09T22:16:12.749812Z","iopub.status.idle":"2025-01-09T22:16:12.772755Z","shell.execute_reply.started":"2025-01-09T22:16:12.749773Z","shell.execute_reply":"2025-01-09T22:16:12.771322Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Debug","metadata":{}},{"cell_type":"markdown","source":"Kudos to https://www.kaggle.com/code/chumajin/janestreet-easy-to-understand-new-time-series-api for proposing a way to test the api\n\nThis section is used to test the model. To run it set DEBUG flag to True.","metadata":{}},{"cell_type":"code","source":"DEBUG = False","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:16:12.775092Z","iopub.execute_input":"2025-01-09T22:16:12.775662Z","iopub.status.idle":"2025-01-09T22:16:12.794808Z","shell.execute_reply.started":"2025-01-09T22:16:12.775586Z","shell.execute_reply":"2025-01-09T22:16:12.793038Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def makelag(date_id):\n    \"\"\"\n    Making lag at the previout day\n\n    Args:\n    date_id (int): date_id at the previous day\n    \n    Returns:\n    pl.dataframe\n    \"\"\"\n    responder_cols = [s for s in train.columns if \"responder\" in s]\n    lag = alltraindata.filter(pl.col(\"date_id\")==date_id).select([\"date_id\",\"time_id\",\"symbol_id\"] + responder_cols).collect()\n    # move forward one day to match the test date_id\n    lag = lag.with_columns(\n        (pl.col(\"date_id\") + 1).alias(\"date_id\")\n    )\n    lag.columns = lag_sample.columns\n    \n    return lag","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:16:12.796609Z","iopub.execute_input":"2025-01-09T22:16:12.797183Z","iopub.status.idle":"2025-01-09T22:16:12.810715Z","shell.execute_reply.started":"2025-01-09T22:16:12.797129Z","shell.execute_reply":"2025-01-09T22:16:12.809339Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def weighted_zero_mean_r2(y_true, y_pred, weights):\n    \"\"\"\n    Calculate the sample weighted zero-mean R-squared score.\n\n    Parameters:\n    y_true (numpy.ndarray): Ground-truth values for responder_6.\n    y_pred (numpy.ndarray): Predicted values for responder_6.\n    weights (numpy.ndarray): Sample weight vector.\n\n    Returns:\n    float: The weighted zero-mean R-squared score.\n    \"\"\"\n    numerator = np.sum(weights * (y_true - y_pred)**2)\n    denominator = np.sum(weights * y_true**2)\n    \n    r2_score = 1 - numerator / denominator\n    return r2_score","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:16:12.812698Z","iopub.execute_input":"2025-01-09T22:16:12.813258Z","iopub.status.idle":"2025-01-09T22:16:12.834008Z","shell.execute_reply.started":"2025-01-09T22:16:12.813206Z","shell.execute_reply":"2025-01-09T22:16:12.832398Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"if DEBUG:\n    lag_sample = pl.read_parquet(\"/kaggle/input/jane-street-real-time-market-data-forecasting/lags.parquet/date_id=0/part-0.parquet\")\n    alltraindata = pl.scan_parquet(\"/kaggle/input/jane-street-real-time-market-data-forecasting/train.parquet\")\n    # pick last 200 days\n    nb_days = 200\n    train = alltraindata.filter(pl.col(\"date_id\")>1698-nb_days).collect()\n    train = train.with_columns(pl.Series(range(len(train))).alias(\"row_id\"))","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:16:12.836101Z","iopub.execute_input":"2025-01-09T22:16:12.836708Z","iopub.status.idle":"2025-01-09T22:16:42.089117Z","shell.execute_reply.started":"2025-01-09T22:16:12.836645Z","shell.execute_reply":"2025-01-09T22:16:42.087525Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"## Step 1 The data is split by day using group_by.\nif DEBUG:\n    all_submission_dataframe = []\n    # Initialize global vars\n    cache = None\n    # tot nb of days counter\n    day_count = 0\n    # training counter to be reset after each train\n    train_counter = 0\n    new_lgbm_model = None\n    lags_ : pl.DataFrame | None = None\n    labels : pl.DataFrame | None = None\n    for num_days, df_per_day in tqdm.tqdm(train.group_by(\"date_id\",maintain_order=True)):\n        \n        ## Step 2 The data is split by time_id using group_by, and the lag is generated (for time_id == 0).\n        \n        for time_id, test in df_per_day.group_by(\"time_id\",maintain_order=True):\n            \n            ## when time_id == 0, makelags\n            \n            if time_id[0] == 0:\n                lag = makelag(num_days[0] - 1)\n            else:\n                lag = None\n            \n            submission_dataframe = predict(test, lag)\n            \n            all_submission_dataframe.append(submission_dataframe)\n            \n    all_submission_dataframe = pl.concat(all_submission_dataframe)\n    all_submission_dataframe\n\n    print(weighted_zero_mean_r2(train.select(\"responder_6\").to_numpy().reshape(-1), all_submission_dataframe.select(\"responder_6\").to_numpy().reshape(-1), train.select(\"weight\").to_numpy().reshape(-1)))","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-09T22:16:42.090917Z","iopub.execute_input":"2025-01-09T22:16:42.091430Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Evaluation with API","metadata":{}},{"cell_type":"code","source":"if CONFIG.EVAL:\n    test_dir = '/kaggle/input/janestreet-updated-simulator-for-time-series-api/debug/test.parquet'\n    lags_dir = '/kaggle/input/janestreet-updated-simulator-for-time-series-api/debug/lags.parquet'\nelse:\n    test_dir = '/kaggle/input/jane-street-real-time-market-data-forecasting/test.parquet'\n    lags_dir = '/kaggle/input/jane-street-real-time-market-data-forecasting/lags.parquet'","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"inference_server = kaggle_evaluation.jane_street_inference_server.JSInferenceServer(predict)\n\nif os.getenv('KAGGLE_IS_COMPETITION_RERUN'):\n    inference_server.serve()\nelse:\n    inference_server.run_local_gateway(\n        (\n            '/kaggle/input/jane-street-real-time-market-data-forecasting/test.parquet',\n            '/kaggle/input/jane-street-real-time-market-data-forecasting/lags.parquet',\n        )\n    )","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}