added load_checkpoint function

import gym

import json

import datetime

import glob

import os

np.random.seed(43)

        f.savefig(filename)

def load_checkpoint(parent_directory="params/", battery_name="battery1", checkpoint=""):

    """

    @brief: returns a dictionary of the latest battery checkpoint

    @input:

        parent_directory: the parent directory with the battery sub directories

        battery_name: name of the battery (subdirectory) to load the file

        checkpoint: the checkpoint to load. defaults to latest checkpoint

    @output:

        a dictionary of battery parameters

    """

    path = f"{parent_directory}/{battery_name}"

    checkpoint = max(glob.iglob(path+"/*.json"), key=os.path.getctime).replace('//','\\')

    #checkpoint = path + "/" + newest

    print(f"[INFO] loading checkpoint: {checkpoint}")

    with open(checkpoint) as f:

        checkpoint_file = json.load(f)

    return checkpoint_file

# ----------------------------------------------------------------------------------- cycle_test

def cycle_test(cell, random_load=True, verbose=0, show_plot=False,

               dt=1.0, save_plot=False, reset=True, file_name="",

        amp = np.random.uniform(1,2)

        pwr = np.random.uniform(.8, 1.4)

        fct = np.random.uniform(3,5)

        ofs = np.random.uniform(3,5)

        ofs = np.random.uniform(int(c)-3,int(c)+3)

        y = np.sin(amp*-x**pwr)/fct+ofs

    done = False

    i = 0

            self.Ir = _i

            self.z = self.z - _z

            self.h = _h

            self.ocv = np.polyval(self.z_coef, self.z * 100.0) - self.R * self.Ir - self.R0 * current + self.h + self.M0 * np.sign(current)

            self.ocv = (np.polyval(self.z_coef, self.z * 100.0) - self.R * self.Ir - self.R0 * current + self.h + self.M0 * np.sign(current))

            if(type(self.ocv) is not np.float64):

                    self.ocv = self.ocv[0]

            # noisy observations

            noisy_z = np.clip(np.random.normal(self.z, .01), 0.0, 100.0)

            else:  # charging, assume some amount of degradation occurs

                self.age += self.avg_load / self.cycle_time * self.charging_age_factor

            self.age = np.clip(self.age, 0.0, self.eol)

            if(type(self.age) is not np.float64):

                    self.age = self.age[0]

        # living reward

        reward = -.01