Create a structured, but more compact JSON representation of a native data structure.

compact_structured_json.py

from __future__ import print_function

"""
Create a structured, but more compact JSON representation of a native data structure.

This is inspired by the way native Python data structures are displayed in Jupyter,
e.g.:

In [1]: d = {"count": 7, "limit": 7, "results": [{"path": "wifi", "meaning": "rssi", "points": [{"timestamp": 1458211728242, "value": -67}, {"timestamp": 1458211736170, "value": -74}, {"timestamp": 1458211740399, "value": -62}, {"timestamp": 1458211746265, "value": -64}], "deviceId": "1e31f865-93b9-482a-a7be-858b56c396ae"}, {"path": "/", "meaning": "batteryLevel", "points": [{"timestamp": 1458211728413, "value": 90}, {"timestamp": 1458211736216, "value": 90}, {"timestamp": 1458211740443, "value": 90}], "deviceId": "1e31f865-93b9-482a-a7be-858b56c396ae"}], "offset": 0}

In [2]: d
Out[2]: 
{'count': 7,
 'limit': 7,
 'offset': 0,
 'results': [{'deviceId': '1e31f865-93b9-482a-a7be-858b56c396ae',
   'meaning': 'rssi',
   'path': 'wifi',
   'points': [{'timestamp': 1458211728242, 'value': -67},
    {'timestamp': 1458211736170, 'value': -74},
    {'timestamp': 1458211740399, 'value': -62},
    {'timestamp': 1458211746265, 'value': -64}]},
  {'deviceId': '1e31f865-93b9-482a-a7be-858b56c396ae',
   'meaning': 'batteryLevel',
   'path': '/',
   'points': [{'timestamp': 1458211728413, 'value': 90},
    {'timestamp': 1458211736216, 'value': 90},
    {'timestamp': 1458211740443, 'value': 90}]}]}

This code is based on using pretty-printing and minimizing indentation
(and is not yet finished...).

To-do:
- treat/test Booleans
- test single and double quotes
"""

import re
import sys
import pprint
import collections
if sys.version_info.major < 3:
    from cStringIO import StringIO
else:
    from io import StringIO


def convert_keys(data):
    "Convert Unicode strings into normal string ones."

    if sys.version_info.major >= 3:
       return data
    else:
        # http://stackoverflow.com/questions/1254454/fastest-way-to-convert-a-dicts-keys-values-from-unicode-to-str
        if isinstance(data, basestring):
            return str(data)
        elif isinstance(data, collections.Mapping):
            return dict(map(convert_keys, data.iteritems()))
        elif isinstance(data, collections.Iterable):
            return type(data)(map(convert_keys, data))
        else:
            return data


def pretty_struc(obj):
    "Return a pretty-printed structured string of a native data structure."

    obj1 = convert_keys(obj)
    out = StringIO()
    pp = pprint.PrettyPrinter(indent=1, stream=out)
    pp.pprint(obj1)
    out.seek(0)
    pretty = out.read()
    out.close()
    return pretty


def compact_json(obj):
    "Return a more compact, but still structured valid JSON representation of 'obj'."

    pp = pretty_struc(obj)

    # detect all indentation levels
    indents = list(set(re.findall('\n\s+[^\s]', pp)))
    counts = sorted([i.count(' ') for i in indents])
    if counts:
        counts2 = [counts[0]] + list(range(counts[0]+1, len(counts)+1))

    def repl(m):
        "Minimize indentation."
        l = m.end() - m.start() - 2
        c = counts2[counts.index(l)]
        return '\n' + ' ' * c + m.string[m.end()-1]

    # minimize indentation
    if indents:
        pp_min = re.sub('\n\s+[^\s]', repl, pp)
    else:
        pp_min = pp

    # convert single quotes into double quotes if not preceeded by backslash
    res = re.sub(r"(?!\\)'", '"', pp_min)

    # handle the rest, mainly Booleans
    # ...

    obj_json = res

    return obj_json


if __name__ == '__main__':
    d = {u"count": 7, "limit": 7, "results": [{"path": "wifi", "meaning": "rssi", "points": [{"timestamp": 1458211728242, "value": -67}, {"timestamp": 1458211736170, "value": -74}, {"timestamp": 1458211740399, "value": -62}, {"timestamp": 1458211746265, "value": -64}], "deviceId": "1e31f865-93b9-482a-a7be-858b56c396ae"}, {"path": "/", "meaning": "batteryLevel", "points": [{"timestamp": 1458211728413, "value": 90}, {"timestamp": 1458211736216, "value": 90}, {"timestamp": 1458211740443, "value": 90}], "deviceId": "1e31f865-93b9-482a-a7be-858b56c396ae"}], "offset": 0}

    print(d)
    print(convert_keys(d))
    print(pretty_struc(d))
    print(compact_json(d))