ent
is a basic framework for generating attribute-based data structures from dictionary-like data sources:
> from ent import Ent
> obj = Ent(foo='foo', bar=None)
> obj.foo
u'foo'
> obj.bar
None
ent
can contain arbitrarily-nested hierarchies, as long as every node/leaf is either a primitive type, or an ent:
> obj = Ent({
'foo': 'bar',
'list': [1, 2, 3, 4],
'inner': {
'foo': 'baz',
'bar': 'bang',
},
'ent': Ent(foo='bar'),
})
> obj.foo
u'bar'
> obj.list[2]
3
> obj.inner.foo
u'baz'
ent
can be merged and diffed, and it will even enforce types when merging keys that are shared:
> ent1 = Ent(foo=1, bar=True)
> ent2 = Ent(foo='hi', bar=False, goo='win')
> Ent.merge(ent1, ent2)
<Ent {'foo': 1, 'bar': False}>
> Ent.diff(ent1, ent2)
<Ent {'bar': False}>
ent
even provides a stand-in for the json
and yaml
modules to automatically convert to and from Ent objects and raw JSON/YAML:
> from ent import json, yaml
> json.loads('{"foo": true, "bar": null}')
<Ent {'foo': True, 'bar': None}>
> json.dumps(Ent(foo=True, bar=None))
u'{"foo": true, "bar": null}'
> yaml.safe_load('bar: null\nfoo: true\n')
<Ent {'foo': True, 'bar': None}>
> yaml.dump(Ent(foo=True, bar=None))
u'baz: 1\nfoo: bar\n'
When working with configs, or other content pulled from sources like JSON, it can be really annoying to need to constantly use brackets to access nested data structures.
Let's say we have a small JSON file containing my personal profile:
{
"name": "John Reese",
"urls": {
"blog": "https://noswap.com",
"github": "https://github.com/jreese",
"facebook": "https://www.facebook.com/nucleareclipse"
}
}
Now let's read in that data and print some of it out using the standard json
module:
import json
with open(...) as f:
data = json.load(f)
name = data['name']
url = data['urls']['github']
We can do better with ent
:
from ent import json
with open(...) as f:
data = json.load(f)
name = data.name
url = data.urls.github
ent is compatible with Python 2.7+ and Python 3.3+. You can install it from PyPI with the following command:
$ pip install ent
ent is copyright 2015 John Reese, and is licensed under the MIT license.