haginile / pyjstat Goto Github PK

LAST RELEASE 1.0: update to 1.0 version is highly recommended, since it supports JSON-stat 2.0 and brings other improvements.

pyjstat is a python library for JSON-stat formatted data manipulation which allows reading and writing JSON-stat [1] format with python,using the DataFrame structures provided by the widely accepted pandas library [2]. The JSON-stat format is a simple lightweight JSON format for data dissemination, currently in its 2.0 version. Pyjstat is inspired in rjstat [3], a library to read and write JSON-stat with R, by ajschumacher. Note that, like in the rjstat project, not all features are supported (i.e. not all metadata are converted). pyjstat is provided under the Apache License 2.0.

This library was first developed to work with Python 2.7. With some fixes (thanks to @andrekittredge), now it works with Python 3.4 too.

Support for JSON-stat 1.3 and 2.0 is provided. JSON-stat 1.3 methods are deprecated now and shouldn't be used in the future, but backwards compatibility has been preserved.

Pyjstat 1.0 is aimed for simplicity. JSON-stat classes have been replicated (Dataset, Collection and Dimension) and provided with simple read() and write() methods. Funcionality covers common use cases as having a URL or dataframe as data sources.

Methods for retrieving the value of a particular cube cell are taken from the JSON-stat Javascript sample code. Thanks to @badosa for this.

Also, version 1.0 makes use of the requests package internally, which should make downloading of datasets easier.

Test coverage is 88% and Travis CI is used.

Finally, note that the new classes and methods are inspired by JSON-stat 2.0, and hence, won't work with previous versions of JSON-stat. However, older methods are still available incorporating bug fixes and performance improvements.

pyjstat requires pandas package. For installation:

pip install pyjstat

Typical usage often looks like this:

from pyjstat import pyjstat

EXAMPLE_URL = 'http://json-stat.org/samples/galicia.json'

# read from json-stat
dataset = pyjstat.Dataset.read(EXAMPLE_URL)

# write to dataframe
df = dataset.write('dataframe')
print(df)

# read from dataframe
dataset_from_df = pyjstat.Dataset.read(df)

# write to json-stat
print(dataset_from_df.write())

This operation mimics the Javascript example in the JSON-stat web page:

from pyjstat import pyjstat

EXAMPLE_URL = 'http://json-stat.org/samples/oecd.json'
query = [{'concept': 'UNR'}, {'area': 'US'}, {'year': '2010'}]

dataset = pyjstat.Dataset.read(EXAMPLE_URL)
print(dataset.get_value(query))

A collection can be parsed into a list of dataframes:

from pyjstat import pyjstat

EXAMPLE_URL = 'http://json-stat.org/samples/collection.json'

collection = pyjstat.Collection.read(EXAMPLE_URL)
df_list = collection.write('dataframe_list')
print(df_list)

In the following example, apikey parameter must be replaced by a real api key from ONS. This dataset corresponds to residence type by sex by age in London:

EXAMPLE_URL = 'http://web.ons.gov.uk/ons/api/data/dataset/DC1104EW.json?'\
              'context=Census&jsontype=json-stat&apikey=yourapikey&'\
              'geog=2011HTWARDH&diff=&totals=false&'\
              'dm/2011HTWARDH=E12000007'
dataset = pyjstat.Dataset.read(EXAMPLE_URL)
df = dataset.write('dataframe')
print(df)

More examples can be found in the examples directory, both for versions 1.3 and 2.0.

This syntax is deprecated and therefore not recommended anymore.

Typical usage often looks like this:

from pyjstat import pyjstat
import requests
from collections import OrderedDict

EXAMPLE_URL = 'http://json-stat.org/samples/us-labor.json'

data = requests.get(EXAMPLE_URL)
results = pyjstat.from_json_stat(data.json(object_pairs_hook=OrderedDict))
print (results)

The same data can be converted into JSON-stat, with some unavoidable metadata loss:

from pyjstat import pyjstat
import requests
from collections import OrderedDict
import json

EXAMPLE_URL = 'http://json-stat.org/samples/us-labor.json'

data = requests.get(EXAMPLE_URL)
results = pyjstat.from_json_stat(data.json(object_pairs_hook=OrderedDict))
print (results)
print (json.dumps(json.loads(pyjstat.to_json_stat(results))))

For a changes, fixes, improvements and new features reference, see CHANGES.txt.