High Level API for json file parsing

[marload]

BPO	41284
Nosy	@rhettinger, @ezio-melotti, @methane, @serhiy-storchaka, @remilapeyre, @tirkarthi, @marload
PRs	bpo-41284: High Level API for json file parsing #21453

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2020-07-12.15:52:30.338>
created_at = <Date 2020-07-12.13:20:43.282>
labels = ['library', '3.10']
title = 'High Level API for json file parsing'
updated_at = <Date 2020-07-13.13:01:41.308>
user = 'https://github.com/marload'

bugs.python.org fields:

activity = <Date 2020-07-13.13:01:41.308>
actor = 'methane'
assignee = 'none'
closed = True
closed_date = <Date 2020-07-12.15:52:30.338>
closer = 'serhiy.storchaka'
components = ['Library (Lib)']
creation = <Date 2020-07-12.13:20:43.282>
creator = 'ys19991'
dependencies = []
files = []
hgrepos = []
issue_num = 41284
keywords = ['patch']
message_count = 5.0
messages = ['373552', '373553', '373555', '373559', '373602']
nosy_count = 7.0
nosy_names = ['rhettinger', 'ezio.melotti', 'methane', 'serhiy.storchaka', 'remi.lapeyre', 'xtreak', 'ys19991']
pr_nums = ['21453']
priority = 'normal'
resolution = 'rejected'
stage = 'resolved'
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue41284'
versions = ['Python 3.10']

[marload]

Many Python users use the following snippets to read Json File.

with oepn(filepath, 'r') as f:
    data = json.load(f)

I suggest providing this snippet as a function.

data = json.read(filepath)

Reading Json is very frequent task for python users. I think it is worth providing this with the High Level API.

[remilapeyre]

Hi, using a file object is very common as it makes it possible to use something that is not a file, like an HTTP request or something already in memory. It makes the module serializing / de-serializing the data completely agnostic with regard to the actual physical storage which has some advantages, for example it will not raise FileNotFound.

If you want a oneliner to load data you already can use:

    with open(filepath) as f: data = json.load(f)

[tirkarthi]

There was a previous issue to support filepath for json.load https://bugs.python.org/issue36378 . This just expands the json API that could already be done using one more operation.

[serhiy-storchaka]

json.load() is already a high level API. json.JSONDecoder is more low level API.

Not every two lines of code should be added as a function in the stdlib. Also, such API would be too complex because you would need to combine parameters of open() (8 parameters) and json.load() (7 parameters). If you use these two lines many times in your code you can just add a simple function that supports only options needed for you.

[methane]

(Off topic) There is a very common mistake in this code:

with oepn(filepath, 'r') as f:
    data = json.load(f)

JSON file is encoded in UTF-8. But the default text encoding is locale encoding.

You should do this instead:

with oepn(filepath, 'rb') as f:
    data = json.load(f)

This works for legacy JSON with UTF-16 or UTF-32 (with/without BOM).too.