The process by which I built the openlims list of enrolled bills

extractenrolled.py

#!/usr/bin/env python

import re
import os
import glob
import json
import pymongo
from pymongo import MongoClient
import shutil

limsdb = json.load(open('./out.json','r'))


client = MongoClient('mongodb://user:[email protected]:10049/limsydb')
collection = client.limsydb.measures

for f in glob.glob('./pdfs/B*_Enr*.pdf'):
	bNo = re.match('(B\d+\-\d{1,4})(\_)', os.path.basename(f)).group(1)
	lNo = collection.find_one({"LegislationNo":bNo})['LawNo']
	if (lNo != ''):
		lNo = re.sub(r'L0',r'L',lNo)
		lNo = re.sub(r'-(0){1,4}','-',lNo)
		shutil.copy(f,'./enr/'+lNo+'.pdf')

getPDFs.py

#!/usr/bin/env python

import json
import re
import os
import urllib

f = open('./out.json','r')
measures = json.load(f) 

f_log = open ('./pdfs/log.txt','a')

bname = "http://dcclims1.dccouncil.us"

for m in measures:
	measure_id = m['LegislationNo']
	for doc in m['versions']:
		fname = './pdfs/' + measure_id + "_" + doc['version'] + '.pdf'
		if (not os.path.exists(fname)):
			url = re.sub('\.\.', bname, doc['url'])
			try:
				urllib.urlretrieve(url,fname)
				f_log.write(fname + ' added.\n')
			except:
				f_log.write("ERROR: There was an error here for:" + fname + "url: " + url + '\n')


'''
Naming convention: "B12-0144_[version].pdf"
'''