def get_name_found_verbose(names, word_lookup):
name_found = {}
ordered_names = sorted(names.items(), key=lambda x: x[1], reverse=True)
for name, count in ordered_names:
if name in name_found:
#print '%s already done' % name
continue
else:
print 'checking %s (%d)' % (name, names[name])
words = name.split()
for i in range(len(words)):
for j in range(len(words)):
new_words = words[:j] + [words[i]] + words[j:]
new_name = ' '.join(new_words)
if new_name in names and new_name not in name_found:
print ' I: %s (%d)' % (new_name, names[new_name])
name_found[new_name] = name
for i in range(len(words)):
new_words = words[:i] + words[i+1:]
new_name = ' '.join(new_words)
if new_name in names and new_name not in name_found:
print ' D: %s (%d)' % (new_name, names[new_name])
name_found[new_name] = name
for i in range(len(words)):
k = loader.sort_word(words[i])
for w, c in word_lookup[k]:
if w != words[i]:
new_words = words[:i] + [w] + words[i+1:]
new_name = ' '.join(new_words)
if new_name in names and new_name not in name_found:
print ' S: %s (%d)' % (new_name, names[new_name])
name_found[new_name] = name
return name_found
def get_name_found(names, word_lookup):
name_found = {}
ordered_names = sorted(names.items(), key=lambda x: x[1], reverse=True)
for name, count in ordered_names:
if name in name_found:
continue
words = name.split()
for i in range(len(words)):
for j in range(len(words)):
new_words = words[:j] + [words[i]] + words[j:]
new_name = ' '.join(new_words)
if new_name in names and new_name not in name_found:
name_found[new_name] = name
for i in range(len(words)):
new_words = words[:i] + words[i+1:]
new_name = ' '.join(new_words)
if new_name in names and new_name not in name_found:
name_found[new_name] = name
for i in range(len(words)):
k = loader.sort_word(words[i])
for w, c in word_lookup[k]:
if w != words[i]:
new_words = words[:i] + [w] + words[i+1:]
new_name = ' '.join(new_words)
if new_name in names and new_name not in name_found:
name_found[new_name] = name
return name_found
import loader
if __name__ == '__main__':
with open('names_short.txt') as infile:
names = [name.strip() for name in infile.readlines()]
ninstances = len(names)
print 'read %d instances' % ninstances
word_lookup = loader.make_word_lookup(names)
print word_lookup
correct = []
for name in names:
words = name.split()
for i in range(len(words)):
words[i] = word_lookup[loader.sort_word(words[i])][0][0]
correct.append(' '.join(words))
with open('corrected.txt', 'w') as outfile:
for i in range(len(names)):
outfile.write('%s\t%s\n' % (names[i], correct[i]))
# for each name in order of descending frequency
# Hypothesis: there is a correct version of the name somewhere in
# the data. For each name in the dataset, especially the ones that
# have only one or two instances, I should be able to remake the
# name by applying the edits. For instance, one of the words is
# probably scrambled, or maybe deleted or doubled. Maybe two of
# the words have been swapped.
