create datapackage for unihan-tabular

create_datapackage.py

#!/usr/bin/env python
# -*- coding: utf8 -*-
"""
Based off http://frictionlessdata.io/guides/creating-tabular-data-packages-in-python/

For use on https://github.com/cihai/unihan-tabular
License: MIT
"""

from __future__ import (absolute_import, division, print_function,
                        unicode_literals, with_statement)

import io
import csv
import os
from jsontableschema import infer

import datapackage

# this and name/title descriptors can be replaced with your package name/title
about = {}
about_file = os.path.join(
    os.path.dirname(__file__), 'unihan_tabular', '__about__.py')
with open(about_file) as fp:
    exec(fp.read(), about)


dp = datapackage.DataPackage()
dp.descriptor['name'] = about['__title__']
dp.descriptor['title'] = about['__title__']

filepath = './data/unihan.csv'

# On Python 2, this would crash due to poor Unicode support (was using this
# on a unicode-rich CSV. Python 3 is slow, so had to chop off the top
with io.open(filepath) as stream:
    headers = stream.readline().rstrip('\n').split(',')
    values = csv.reader(stream)
    schema = infer(headers, values)
    dp.descriptor['resources'] = [
        {
            'name': 'data',
            'path': filepath,
            'schema': schema
        }
    ]

# datapackage Version: 0.8.8 would wrongly attribute a date-like type
# to many fields, so had to find-replace them with "string"
with open('datapackage.json', 'w') as f:
    f.write(dp.to_json())