export declare const pythonCode = "\nimport numpy as np\nimport pandas as pd\nfrom typing import Dict, Any\n\ndef summarize_common_metadata(x: pd.DataFrame, trials_expected: int) -> Dict[str, Any]:\n    \"\"\"\n    Extract shared metadata and trial count checks from a grouped DataFrame.\n\n    Args:\n        x (pd.DataFrame): Grouped trial-level DataFrame.\n        trials_expected (int): Expected number of trials.\n\n    Returns:\n        Dict[str, Any]: Summary metadata and quality flags.\n    \"\"\"\n    if \"trial_index\" in x.columns:\n        n_trials = x[\"trial_index\"].nunique()\n    else:\n        # warnings.warn(\"No 'trial_index' column found; assuming single-trial task.\")\n        # For single-trial tasks (like Trailmaking), treat as one trial\n        n_trials = 1\n    return {\n        \"activity_begin_iso8601_timestamp\": x[\"activity_begin_iso8601_timestamp\"].iloc[\n            0\n        ],\n        \"n_trials\": n_trials,\n        \"flag_trials_match_expected\": n_trials == trials_expected,\n        \"flag_trials_lt_expected\": n_trials < trials_expected,\n        \"flag_trials_gt_expected\": n_trials > trials_expected,\n    }\n\ndef score_accuracy(row, legacy=False):\n    try:\n        if legacy:\n            return row[\"user_response\"] == row[\"correct_response\"]\n        else:\n            return row[\"user_response_index\"] == row[\"correct_response_index\"]\n    except Exception as e:\n        print(f\"Error processing row: {e}\")\n        return None\n\ndef summarize(x, trials_expected=20, rt_outlier_low=100, rt_outlier_high=10000):\n   \n    # ABSTRACTION TO APPEAR IN EACH SCORING SCRIPT\n    d = summarize_common_metadata(x, trials_expected)\n\n    \n    # trial counts (for various denominators)\n    d[\"n_trials_total\"] = x[\"trial_index\"].nunique()\n    d[\"n_trials_lure\"] = (x[\"trial_type\"] == \"lure\").sum()\n    d[\"n_trials_normal\"] = (x[\"trial_type\"] == \"normal\").sum()\n\n\n    # tabulate accuracy\n    d[\"n_trials_correct\"] = (\n        x[\"user_response_index\"] == x[\"correct_response_index\"]\n    ).sum()\n\n    d[\"n_trials_incorrect\"] = (\n        x[\"user_response_index\"] != x[\"correct_response_index\"]\n    ).sum()\n\n    # Filter out outliers: RT < 100 ms or RT > 10,000 ms\n    rt_filtered = x.loc[\n        (x[\"response_time_duration_ms\"] >= rt_outlier_low)\n        & (x[\"response_time_duration_ms\"] <= rt_outlier_high),\n        \"response_time_duration_ms\",\n    ]\n    d[\"median_response_time_filtered\"] = rt_filtered.median()\n    \n    # correct AND RT within bounds\n    correct_and_filtered = x.loc[\n        (x[\"user_response_index\"] == x[\"correct_response_index\"])\n        & (x[\"response_time_duration_ms\"] >= rt_outlier_low)\n        & (x[\"response_time_duration_ms\"] <= rt_outlier_high),\n        \"response_time_duration_ms\"\n    ]\n    d[\"median_response_time_correct_filtered\"] = correct_and_filtered.median()\n    d[\"std_response_time_correct_filtered\"] = correct_and_filtered.std()\n\n    # get RTs for correct and incorrect trials\n    d[\"median_response_time_overall\"] = x[\"response_time_duration_ms\"].median()\n    d[\"median_response_time_correct\"] = x.loc[\n        (x[\"user_response_index\"] == x[\"correct_response_index\"]),\n        \"response_time_duration_ms\",\n    ].median()\n    d[\"median_response_time_incorrect\"] = x.loc[\n        (x[\"user_response_index\"] != x[\"correct_response_index\"]),\n        \"response_time_duration_ms\",\n    ].median()\n\n    # return as series\n    indices = list(d.keys())\n    return pd.Series(\n        d,\n        index=indices,\n    )\n";
//# sourceMappingURL=python-code.d.ts.map