from ModelFit import get_best_fit
import matplotlib.pyplot as plt

import os
import numpy as np


def get_all_cell_folders(dir):
    paths = []
    for item in os.listdir(dir):
        cell_path = os.path.join(dir, item)
        if os.path.isdir(cell_path):
            paths.append(cell_path)

    return paths


def main():

    values_models = {}
    values_cells = {}

    for cell in get_all_cell_folders("results/invivo_results/"):
        best_fit = get_best_fit(cell)
        c_behaviour, m_behaviour = best_fit.get_behaviour_values()

        for key in c_behaviour.keys():
            if key not in values_cells.keys():
                values_cells[key] = []
            if key not in values_models.keys():
                values_models[key] = []

            values_cells[key].append(c_behaviour[key])
            values_models[key].append(m_behaviour[key])

    percentage_error_data = []
    labels = []
    for key in values_models.keys():
        errors = []
        for i in range(len(values_models[key])):
            if values_cells[key][i] == 0:
                if values_models[key][i] == 0:
                    errors.append(0)
                else:
                    print("Cannot calc % error if reference is 0")
                continue
            errors.append((values_models[key][i] / values_cells[key][i]) -1)

        percentage_error_data.append(errors)
        labels.append(key)

    plt.boxplot(percentage_error_data)
    plt.xticks(range(1, len(percentage_error_data)+1, 1), labels=labels, rotation=45)
    plt.show()
    plt.close()

    # plt.plot([np.median(x) for x in percentage_error_data], 'o')
    # plt.xticks(range(len(percentage_error_data)), labels=labels, rotation=45)
    # plt.show()


if __name__ == '__main__':
    main()