Save a tensorflow.keras model to string

I work with python 3.6 and tensorflow 1.14.

I get a dynamically created tensorflow model and need to store it together with some other information. Therefore I want to serialize the whole trained model into a string. The string gets then later stored together with the other information as file (e.g. json or pickle).

In other words, I want to write the serialize and deserialize function in that example code:

import tensorflow as tf
from tensorflow import keras

def get_trained_model() -> tf.keras.Sequential:
    """
    Dummy code to build a model.
    """
    model = keras.Sequential([
        keras.layers.Dense(8, activation=tf.nn.relu),
        keras.layers.Dense(8, activation=tf.nn.relu),
        keras.layers.Dense(2, activation=tf.nn.softmax)
    ])
    model.compile(optimizer='adam',
                  loss='sparse_categorical_crossentropy',
                  metrics=['accuracy'])

    return model

def serialize(model:tf.keras.Sequential) -> str:
    """
    Serializes the whole tensorflow model (architecture and weights) as string, so that it can be
    saved, shared and later deserialized.
    """
    pass

def deserialize(model_string: str) -> tf.keras.Sequential:
    """
    Deserializes a string into a working tensorflow model object.
    """
    pass

model = get_trained_model()
m = serialize(model)
new_model = deserialize(m)

My current solution is to use model.save(), to write the model into a file, read the string from there and to delete the file again. Since this solution is not very clean I search for a better one.

Thanks a lot for any help or suggestions.