colinpollock

"""Make initial clusters of categories to bootstrap top-level categories."""

from collections import defaultdict

from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.cluster import KMeans, MiniBatchKMeans

from j_util import get_rows

colinpollock

a = 'a'
b = 'b'
c = 'c'
d = 'd'

def test_empty():
    check([], 0)

def test_all_one():
    check([a, a], 1)

colinpollock

sentence = 'Reggie Miller grew up in Riverside before going to UCLA.'
phrase = 'Riverside'
start_index = sentence.find(phrase)
end_index = start_index + len(phrase)
sentence[start_index-10:end_index+10]
# 'rew up in Riverside before go'

colinpollock

import spacy

nlp = spacy.load('en_vectors_web_lg')

text1 = 'The medical field is moving forward rapidly.'
text2 = 'Medicine is vital to the industry.'
text3 = 'Reggie Miller is a basketball player.'

doc1 = nlp(text1)
doc2 = nlp(text2)

colinpollock

import spacy
from spacy.pipeline import EntityRuler

nlp = spacy.load("en_core_web_sm")

ruler = EntityRuler(nlp, validate=True)
ruler.add_patterns[
    {
        "label": "PHRASE",
        "pattern": [

colinpollock

"""Make initial clusters of categories to bootstrap top-level categories."""

from collections import defaultdict

from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.cluster import KMeans, MiniBatchKMeans

from j_util import get_rows

colinpollock

from collections import defaultdict
import random


restaurants = [
    'za', 'jake', 'spoon', 'hacienda', 'bridgetender', 'river grill', 'rosie', 'sunnyside'
]

people = ['megan', 'jan', 'todd', 'colin', 'michelle']
random.shuffle(people)

colinpollock

IT'S GO TIME

YO YO MA isLessThan20
IT'S IN THE VAULT  @IT'S NOT A LIE IF YOU BELIEVE IT
YO YO MA n
IT'S IN THE VAULT 0
YO YO MA multiple
IT'S IN THE VAULT @IT'S NOT A LIE IF YOU BELIEVE IT

CRY CRY AGAIN isLessThan20

colinpollock

IT'S GO TIME
GIDDYUP "Hello, Newman"
YADA YADA YADA

colinpollock

   public static ClientConfig getClientConfig(String connectionUrl) {
        ClientConfig clientConfig = new ClientConfig();
        LinkedHashSet<String> servers = new LinkedHashSet<String>();
        servers.add(connectionUrl);
        clientConfig.getServerProperties().put(ClientConstants.SERVER_LIST, servers);
        return clientConfig;
    }

    public static JestClient getClient(String connectionUrl) {
        JestClientFactory factory = new JestClientFactory();