how to ZstdCompressionDict? Some real world examples?

[ewerybody]

Hello!
I'm having a hard time digging through readthedocs ...

First of all: This is not about Python Dictionaries, right? In fact this got me super confused in the beginning. Maybe this could be somehow stated in the docs? :)

So here data is just my data to compress as bytes:

>>> dict_data = zstandard.ZstdCompressionDict(data)

If we want to stick to the python dict idea: This could very well be data = json.dump(py_dict), right?

Now here it gets tricky:

>>> dict_data = zstandard.train_dictionary(size, samples)
>>> dict_size = len(dict_data)  # will not be larger than ``size``

So size is some integer but one line later we get dict_size which is obviously different? Or is it? Why is it just the whole data byte size? Here in the tests there is a function get_optimal_dict_size_heuristically() But why is this optimal? A comment would be good there :)

And samples: apparently this is a list! But how should I select the samples ideally? On the main docs page it says:

Training Zstandard is achieved by providing it with a few samples (one file per sample).

So .. given the json dump from above: Yet we don't have any file. :)

Probably I'm acting really really dumb and all this is rather obvious. Maybe I just need a little hint to pull this off.
Would someone be so kind to help me explaining?
Thanks already!
ëRiC