Cfrtjryk

So I've picked up PyYAML over the past few days but I feel like some of the documentation is vague or I'm misinterpreting it. Here is a function for writing data to a .yaml which can write the values to each element I have in the list:

def writeSave(filename, **kwargs):

    with open(filename, 'r') as yamlFile:

        yamlDict = yaml.load(yamlFile)

        if not yamlDict: yamlDict = {}



    with open(filename, 'w') as write:

        z = yamlDict.copy()

        z.update(kwargs)

        yaml.dump(z, write)

Which will then overwrite the values below when called:

saveData:

        {mission_state: 7, playername: 'steve'}

Using this function seems simple:

writeSave('../path/savedata.yaml', mission_state=2)

Which will then change the mission_state in the list.

I don't know if I've done this smoothly or if there is a better practice to performing these kind of functions. If there are, it'd be nice to have some kind of pointers.

Is there a way to make this more readable? Maybe more readable vars or more commenting?

• First, Python has an official style-guide, PEP8, which you should follow. It recommends using lower_case for functions and variables.




• I would also rename your function, since write_save does not say what this function is doing. I think overwrite_setting or update_value or something similar would be better.




• You might want to add a check if any keyword arguments are passed to the function. Either raise an exception (a simple AssertionError might be enough), or at least don't waste time reading the file and writing the same content back.




• While open opens a file in read-mode by default, here I think it is actually fine to open it once explicitly with 'r' and then with 'w'.




• Your z = yamDict.copy() should be completely unnecessary. Mutating the original dictionary does not do any harm. The updating does not need to be within the with block either.




• Finally, you could use or for the check of an empty dictionary.

With all of the above changes, your code would become:

import yaml



def update_value(filename, **kwargs):

    assert kwargs

    with open(filename, 'r') as f:

        yaml_dict = yaml.load(f) or {}

    yaml_dict.update(kwargs)

    with open(filename, 'w') as f:

        yaml.dump(yaml_dict, f)

If you have only one thread, you could skip the reading of the config everytime you want to change a value and instead keep the dictionary in memory (after having it read once at the beginning). You will still want the write to save state, though. For this you might want to have a Config class, derived from dict:

import os



class Config(dict):

    def __init__(self, filename):

        self.filename = filename

        if os.path.isfile(filename):

            with open(filename) as f:

                # use super here to avoid unnecessary write

                super(Config, self).update(yaml.load(f) or {})



    def __setitem__(self, key, value):

        super(Config, self).__setitem__(key, value)

        with open(self.filename, "w") as f:

            yaml.dump(self, f)



    def __delitem__(self, key):

        super(Config, self).__delitem__(key)

        with open(self.filename, "w") as f:

            yaml.dump(self, f)



    def update(self, kwargs):

        super(Config, self).update(kwargs)

        with open(self.filename, "w") as f:

            yaml.dump(self, f)

You might want to define a decorator to add the dumping part to the methods:

import functools



def dumps(func):

    @functools.wraps(func)

    def wrapper(self, *args, **kwargs):

        ret = func(self, *args, **kwargs)

        with open(self.filename, "w") as f:

            yaml.dump(self, f)

        return ret

    return wrapper



class Config(dict):

    def __init__(self, filename):

        self.filename = filename

        if os.path.isfile(filename):

            with open(filename) as f:

                # use super here to avoid unnecessary write

                super(Config, self).update(yaml.load(f) or {})



    __setitem__ = dumps(dict.__setitem__)

    __delitem__ = dumps(dict.__delitem__)

    update = dumps(dict.update)

You can use this class like a dict, but every change to the dictionary is also dumped to the yaml file. The file is only read once (when the config is initialized), so changes to the file will only be visible when the script is re-started.

cfg = Config("test.yaml")

print(cfg)

cfg['a'] = 3

print(cfg)

cfg.update({"b": 4})

cfg.update(c=5)

del cfg['a']

print(cfg)