Reexpress MCP Server

# Copyright Reexpress AI, Inc. All rights reserved. import numpy as np MARGINAL_ACC_KEY = "m" MARGINAL_ADMITTED_KEY = "m_prop" CLASS_CONDITIONAL_ACC_KEY = "c" CLASS_CONDITIONAL_ADMITTED_KEY = "c_prop" PREDICTION_CONDITIONAL_ACC_KEY = "p" PREDICTION_CONDITIONAL_ADMITTED_KEY = "p_prop" def get_acc_prop_tuple(list_to_process, total=None): if total is not None and total > 0: prop = len(list_to_process)/total else: prop = 0.0 acc = np.mean(list_to_process) if len(list_to_process) > 0 else 0 return acc, prop def get_decorated_string(float_proportion, display_string, decorate=False, decorate_alpha=0.95, is_fully_rejected=False): if not decorate: return display_string if is_fully_rejected: return rf"\colorbox{{correctPredictionColor}}{{{display_string}}}" if float_proportion >= decorate_alpha: return rf"\colorbox{{correctPredictionColor}}{{{display_string}}}" else: return rf"\colorbox{{wrongPredictionColor}}{{{display_string}}}" def get_float_as_display_significant_digits_string(float_proportion, decorate=False, decorate_alpha=0.95, is_fully_rejected=False) -> str: if is_fully_rejected: return get_decorated_string(float_proportion, r"\allRejected", decorate=decorate, decorate_alpha=decorate_alpha, is_fully_rejected=is_fully_rejected) if float_proportion == 0.0: return get_decorated_string(float_proportion, "0.", decorate=decorate, decorate_alpha=decorate_alpha) elif float_proportion == 1.0: return get_decorated_string(float_proportion, "1.", decorate=decorate, decorate_alpha=decorate_alpha) if float_proportion < 0.005 and float_proportion != 0.0: return get_decorated_string(float_proportion, "<0.01", decorate=decorate, decorate_alpha=decorate_alpha) else: return get_decorated_string(float_proportion, f"{float_proportion:.2f}", decorate=decorate, decorate_alpha=decorate_alpha) def get_latex_row(dataset_name, model_name, alpha, latex_rows_dict, estimator_label, numberOfClasses): running_latex_rows = [] for class_label in range(numberOfClasses): conditional_acc = \ get_float_as_display_significant_digits_string( latex_rows_dict[f"{CLASS_CONDITIONAL_ACC_KEY}{class_label}"], decorate=True, decorate_alpha=alpha, is_fully_rejected=latex_rows_dict[f"{CLASS_CONDITIONAL_ADMITTED_KEY}{class_label}"] == 0.0) running_latex_rows.append(conditional_acc) admitted_proportion = \ get_float_as_display_significant_digits_string( latex_rows_dict[f"{CLASS_CONDITIONAL_ADMITTED_KEY}{class_label}"], decorate=False, decorate_alpha=alpha, is_fully_rejected=False) running_latex_rows.append(admitted_proportion) for class_label in range(numberOfClasses): conditional_acc = \ get_float_as_display_significant_digits_string( latex_rows_dict[f"{PREDICTION_CONDITIONAL_ACC_KEY}{class_label}"], decorate=True, decorate_alpha=alpha, is_fully_rejected=latex_rows_dict[f"{PREDICTION_CONDITIONAL_ADMITTED_KEY}{class_label}"] == 0.0) running_latex_rows.append(conditional_acc) admitted_proportion = \ get_float_as_display_significant_digits_string( latex_rows_dict[f"{PREDICTION_CONDITIONAL_ADMITTED_KEY}{class_label}"], decorate=False, decorate_alpha=alpha, is_fully_rejected=False) running_latex_rows.append(admitted_proportion) marginal_acc = get_float_as_display_significant_digits_string( latex_rows_dict[MARGINAL_ACC_KEY], decorate=True, decorate_alpha=alpha, is_fully_rejected=latex_rows_dict[MARGINAL_ADMITTED_KEY] == 0.0) running_latex_rows.append(marginal_acc) marginal_admitted_proportion = get_float_as_display_significant_digits_string( latex_rows_dict[MARGINAL_ADMITTED_KEY], decorate=False, decorate_alpha=alpha, is_fully_rejected=False) running_latex_rows.append(marginal_admitted_proportion) return " & ".join([dataset_name, model_name, estimator_label]) + " & " + " & ".join(running_latex_rows) + r"\\" def init_latex_rows_dict(numberOfClasses): latex_rows_dict = {} latex_rows_dict[MARGINAL_ACC_KEY] = 0.0 # marginal accuracy latex_rows_dict[MARGINAL_ADMITTED_KEY] = 0.0 # marginal |Admitted| / |N| for class_label in range(numberOfClasses): latex_rows_dict[f"{CLASS_CONDITIONAL_ACC_KEY}{class_label}"] = 0.0 # class-conditional accuracy latex_rows_dict[f"{CLASS_CONDITIONAL_ADMITTED_KEY}{class_label}"] = 0.0 # class-conditional |Admitted| / |N| latex_rows_dict[f"{PREDICTION_CONDITIONAL_ACC_KEY}{class_label}"] = 0.0 # prediction-conditional accuracy latex_rows_dict[f"{PREDICTION_CONDITIONAL_ADMITTED_KEY}{class_label}"] = 0.0 # prediction-conditional |Admitted| / |N| return latex_rows_dict def print_latex_row(options, model, alpha_prime, latex_rows_dict_no_reject, latex_rows_dict_softmax_f, latex_rows_dict_softmax_df, latex_rows_dict_sdm, latex_rows_dict_sdm_hr): if options.construct_results_latex_table_rows: latex_meta_data_list = options.additional_latex_meta_data.strip().split(",") if len(latex_meta_data_list) == 2: dataset_name = rf'$\{latex_meta_data_list[0]}$' model_name = rf'$\{latex_meta_data_list[1]}$' else: dataset_name = rf'DATASET-NAME-HERE' model_name = rf'MODEL-NAME-HERE' print(f"Latex-formatted results table rows (alpha={alpha_prime})") print( get_latex_row(dataset_name, model_name, alpha_prime, latex_rows_dict_no_reject, r'$\estimatorNoReject$', model.numberOfClasses) ) print( get_latex_row(dataset_name, model_name, alpha_prime, latex_rows_dict_softmax_f, r'$\estimatorSoftmax$', model.numberOfClasses) ) print( get_latex_row(dataset_name, model_name, alpha_prime, latex_rows_dict_softmax_df, r'$\estimatorSoftmaxOverDistanceMagnitude$', model.numberOfClasses) ) print( get_latex_row(dataset_name, model_name, alpha_prime, latex_rows_dict_sdm, r'$\sdm$', model.numberOfClasses) ) print( get_latex_row(dataset_name, model_name, alpha_prime, latex_rows_dict_sdm_hr, r'$\sdmHR$', model.numberOfClasses) )