DEV Community

Discussion on: 30 PyTricks I've Learned By Joining the Real Python Mailing List.

View post
Collapse
 
geraldew profile image
geraldew •

By coincidence I recently encountered the reason why item 7 "Pretty print a dict." is not reliable.
i.e. Answer: Using json.dumps method.

This will fail when the items in the dictionary are not things that can be expressed as JSON. The two I encountered were enumerations and tuples.

Collapse
 
gdledsan profile image
Edmundo Sanchez •

There is a library to solve that, check orjson, its faster, it parses more types, check it out

Collapse
 
wiseai profile image
Mahmoud Harmouch •

Actually, I don't like to use this method, but rather engineer my own thing. For instance, at some point in the past, i created these handy functions:

def indent(string: str, num_spaces: int) -> str:
    indent_str: str = " " * num_spaces
    return "\n".join(indent_str + line for line in string.split("\n")[:-1]) + "\n"

def pretty_dict(dict_: dict[str, str], num_spaces: int) -> str:
  if not isinstance(dict_, dict):
    return repr(dict_)
  rep = []
  for key, val in dict_.items():
    rep.append(f"{key}: {pretty_dict(val, num_spaces)},\n")
  if rep:
    return f'{{\n{indent("".join(rep), num_spaces)}}}'
  else:
    return "{}"
Enter fullscreen mode Exit fullscreen mode

Usage:

>>> dict1 = {'b': 2, 'a': {'b': [1, 2, 3]}, 'c': list(enumerate(range(4)))}
>>> print(pretty_dict(dict1, 2))
{
  b: 2,
  a: {
    b: [1, 2, 3],
  },
  c: [(0, 0), (1, 1), (2, 2), (3, 3)],
}
Enter fullscreen mode Exit fullscreen mode

It seems like, over the years, I developed a hate relationship with the DRY principle. Anyways, I hope you find it useful. Cheers mate!

Collapse
 
tompollard61 profile image
Tom Pollard •

json.dumps() will represent tuples as JSON lists. But there are certainly other Python types that json.dumps() can't handle, like datetimes

The pprint ("pretty-print") module handles all of those. So, instead of json.dumps() you can use pprint.pprint()

Collapse
 
gdledsan profile image
Edmundo Sanchez •

Check orjson, faster and better json handling

Collapse
 
geraldew profile image
geraldew •

Well it simply didn't in the environment that I was using, instead it crashed on the tuple. To be clear, the tuple was being used as a key in the Python dictionary, something that Python allows but I'm guessing has no equivalence in JSON.

Thread Thread
 
tompollard61 profile image
Tom Pollard •

Yes, JSON keys need to be strings.

In any case, pprint was introduced to the stdlib specifically to handle pretty-printing of Python values, but I don't run across many people who seem to know about it.

Thread Thread
 
geraldew profile image
geraldew •

post script - as it happened I was curious enough about JSON versus data structures to try getting it to handle something more than just text keys. So I wrote an experiment about using enumerations - see: