print "By '%s' did you mean:"%orig

for i, (name, uri) in enumerate(possibles):

print "%d)\t%s"%((i+1),name)

print "%d)\tNone of the above."%(len(possibles)+1)

valid = False

n = 0

while not valid:

valid = True

try:

n = int(raw_input())

except Exception:

valid = False

if not 1<=n<=len(possibles)+1:

valid = False

if n-1==len(possibles):

return None

orig_name = name

name = ''.join(re.findall('[A-Z0-9]+',name.upper()))

best_authors = []

with open('sample_data/author_names.json', 'r') as f:

j = json.load(f)

for b in j['results']['bindings']:

author_orig = b['name']['value']

uri = b['author']['value']

author = b['name']['value'].upper()

subnames = author_orig.split()

author = ''.join(re.findall('[A-Z0-9]+',author))

dist = jaccard_ngram_dist(name,author,3)

best_authors.append(((author_orig,uri),dist))

if len(subnames)>=2:

for sname in [subnames[0], subnames[-1]]:

sname = ''.join(re.findall('[A-Z0-9]+',sname))

dist = jaccard_ngram_dist(name,sname,3)

best_authors.append(((author_orig,uri),dist))

if len(best_authors)>20:

best_authors.sort(key=lambda x:x[1])

best_authors = best_authors[:5]

best_authors.sort(key=lambda x:x[1])

best_authors = best_authors[:5]

best_dist = best_authors[0][1]

possibles = [best_authors[0][0]]

for author, dist in best_authors[1:]:

percent_diff = (dist-best_dist)*2/float(dist+best_dist)

if percent_diff < __CUTOFF__:

possibles.append(author)

if len(possibles)>1:

identified = pp(orig_name, possibles)

else:

identified = possibles[0]

test_values = [

"Paolini", "JK Rowling",

"Cristopher Paolini",

"Robert Heinline"

]

for test in test_values:

print "\nIdentifying '%s'..."%test

print ident_author(test)

while True:

print "\nTry inputing a name. (or 'q' to exit)"

name = raw_input()

if name=='q':

break