import pandas as pd
from pandas import Series, DataFrame

from itertools import islice

import datetime

from itertools import islice
import codecs
import re
import csv

import os

if os.getcwd() == '/home/picloud/notebook':
    ON_PICLOUD = True
    DATA_DIR = '/home/picloud/working-open-data/data'
    PYDATA_DIR = '/home/picloud/pydata-book/'
else:
    ON_PICLOUD = False
    DATA_DIR = os.path.join(os.pardir, "data")
    PYDATA_DIR = os.path.join(os.pardir, "pydata-book")
    
dataset_fname = os.path.join (DATA_DIR, "census/DataSet.txt")
datadict_fname = os.path.join (DATA_DIR, "census/DataDict.txt")
fips_fname = os.path.join (DATA_DIR, "census/FIPS_CountyName.txt")

assert os.path.exists(DATA_DIR)
assert os.path.exists(PYDATA_DIR)
assert os.path.exists(dataset_fname)
assert os.path.exists(datadict_fname)
assert os.path.exists(fips_fname)


# read in fips code
fips_file = codecs.open(fips_fname, encoding='iso-8859-1')

fips = dict()
for row in islice(fips_file, None):
    fips[row[:5]] = row[6:-1]
    

# read in data set
ds_file = codecs.open(dataset_fname, encoding='iso-8859-1')
reader = csv.DictReader(ds_file)
dataset = dict([(row["fips"], row) for row in islice(reader, None)])
    
states_fips = sorted([k for k in fips.keys() if k[-3:] == '000' and k != '00000'])


!head $fips_fname

dataset["00000"]["POP010210"]

for f in states_fips[:5]:
    print f, fips[f]

list(islice(sorted(dataset.keys()),5))

fips['06000']

# number of counties in CA

from collections import Counter
print Counter([k[0:2] for k in dataset.keys() if k[2:5] != '000'])['06']



























































import string
alphabet = string.lowercase

alphabet

print "alphabet:", alphabet
print "alphabet[0]:", alphabet[0]
print "alphabet[-1], alphabet[0:5], alphabet[-2:]:", alphabet[-1], alphabet[0:5], alphabet[-2:]

alphabet[5]





alphabet[0:3]





alphabet[1:4:2]





alphabet[-6:]





alphabet[-1:-3:-1]






a = array([0,1,2,3])

a + 5

sum(2*a)















# round off atomic weight

elements = DataFrame([{'number': 1, 'name': 'hydrogen', 'weight':1}, 
                      {'number': 2, 'name': 'helium', 'weight':4},
                      {'number': 3, 'name': 'lithium', 'weight':7},
                      {'number': 4, 'name': 'beryllium', 'weight':9},
                      {'number': 5, 'name': 'boron', 'weight':11},
                      {'number': 6, 'name': 'carbon', 'weight':12},
                     ], index= ['H', 'He', 'Li', 'Be', 'B', 'C'])

# add group information

elements['group'] = Series([1, 18, 1, 2, 13, 14], index = ['H', 'He', 'Li', 'Be', 'B', 'C'])

elements








len(elements.index)




elements[elements.number > 4]["weight"].sum()










set(elements[elements['group'] == 1].name)










elements.sort_index(by='weight')['number'][::-1][:2].sum()













comments = Series(['first and most common element', 'the C in organic'], index=['H', 'C'])

elements['comments'] = comments
elements

elements.comments.dropna().count()






"".join(elements.comments.dropna().apply(lambda x: x[0]).values)











"".join(['a','b'])

elements.number.apply(lambda x: 2*x)

"".join(elements.number.apply(lambda x: str(2*x)).values)

import datetime

(datetime.datetime.now() + datetime.timedelta(days=20)).month