This commit is contained in:
Ivan Ivanov 2015-11-07 16:47:11 -05:00
parent 1cc6af83c0
commit 617e39a0c0

88
scripts/enrich_dataset.py Normal file
View File

@ -0,0 +1,88 @@
#/usr/bin/env python
import argparse
import os, re, sys
import requests
import urllib
import pandas as pd
IMPORT_API_ENDPOINT = (
'https://api.import.io/store/data/8a8c1017-a2ec-46b2-a20a-a7c711496ee8/_query'
'?input/webpage/url={}'
'&_user=303b474d-8ebb-45f1-aa94-27a7b1719366'
'&_apikey=303b474d8ebb45f1aa9427a7b1719366818dc07c503b3c555038a8d03dae311e551744249d9a429177684fb0ecc415f942aece4a63b938b8d1c796f42a10114778c3182bc5556d245329e6dfcf9f7ad4'
)
MEPS_CODEBOOK = (
'http://meps.ahrq.gov/mepsweb/data_stats/download_data_files_codebook.jsp'
'?PUFId={}&varName={}'
)
def get_labels(PUFId, varName):
"""Loads the mapping of numeric IDs to labels of a categorical field"""
meps_url = MEPS_CODEBOOK.format(PUFId, varName)
import_url = IMPORT_API_ENDPOINT.format(urllib.quote(meps_url, safe=''))
r = requests.get(import_url)
labels = pd.DataFrame(r.json()['results'])
labels = labels[labels['value_value'] != 'TOTAL']
# Extract the int id
labels[varName] = labels.value_value.str.split(' ').apply(lambda x: int(x[0]))
# And the label
labels[varName + '_label'] = labels.value_value.str.split(' ').apply(lambda x: ' '.join(x[1:]))
return labels
def load_dictionary(dictfile):
df = pd.read_csv(dictfile)
return dict(zip(df.NAME, df.DESCRIPTION))
# hardcoded config of which file corresponds to which year.
YEAR_FILES = dict([('h{}e.csv'.format(idx), 2013-i) for i, idx in enumerate([160, 152, 144, 135, 126, 118, 110, 102, 94,
85, 77, 67, 59, 51, 33, 26, 16, 10])]
+ [('h{}f.csv'.format(idx), 2013-i) for i, idx in enumerate([160, 152, 144, 135, 126, 118, 110, 102, 94,
85, 77, 67, 59, 51, 33, 26, 16, 10])])
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--input-file')
parser.add_argument('--output-dir')
parser.add_argument('--column-dictionary')
parser.add_argument('--category')
args = parser.parse_args()
infile = os.path.basename(args.input_file)
puf_id = os.path.splitext(infile)[0].upper()
raw_data = pd.read_csv(args.input_file)
raw_data.columns = [x.upper() for x in raw_data.columns]
for col in raw_data.columns:
try:
labels = get_labels(puf_id, col)
except Exception,e :
continue
joined = raw_data.merge(labels, on=col)
raw_data[col + '_label'] = joined[col + '_label']
if args.column_dictionary:
dictionary = load_dictionary(args.column_dictionary)
srcs = dictionary.keys()
for k in srcs:
dictionary[k + '_label'] = dictionary[k] + '_label'
raw_data.columns = [dictionary.get(col, col) for col in raw_data.columns]
outfile = 'enriched_{}'.format(infile)
if infile in YEAR_FILES:
outfile = 'enriched_{}_{}'.format(YEAR_FILES[infile], infile)
output = '{}/{}'.format(args.output_dir, outfile)
raw_data.to_csv(output)