def cascade(self, instance):
"""Cascade the index from this instance to others that depend on it.
This causes index_instance() to be called on each instance that depends
on the instance supplied.
"""
for descriptor in self.cascades:
cascade_inst = None
# find the instance we're being told to cascade the reindex onto
try:
if callable(descriptor):
cascade_inst = descriptor(instance)
elif isinstance(descriptor, str):
cascade_inst = getattr(instance, descriptor)
except:
cascade_inst = None
# if we found one, check if it's searchable, check if it
# wants to accept the cascade, and if so, reindex it
if cascade_inst:
# If it's not an iterable already, make it into one
if not hasattr(cascade_inst, '__iter__'):
cascade_insts = [cascade_inst]
else:
cascade_insts = cascade_inst
for cascade_inst in cascade_insts:
indexer = get_indexer(cascade_inst)
if indexer and indexer.reindex_on_cascade(instance, cascade_inst):
indexer.index_instance(cascade_inst, with_cascade=False)
def main():
train_data = get_train_data_from_csv('data/train_15_dns.csv')
#shuffle(train_data)
train_data = train_data[0:50000]
dev_data = get_dev_data_from_csv('data/dev_15_dns.csv')
#shuffle(dev_data)
dev_data = dev_data[0:10000]
print('len of training data:', len(train_data))
print('len of dev data:', len(dev_data))
vocab_size = get_vocab_size(train_data)
print('calculated vocab size:', vocab_size)
indexer = get_indexer('indexer_15_dups.csv')
model = train_rnn_classifier(train_data, vocab_size, indexer)
print_evaluation(dev_data, model)
from keras.preprocessing.sequence import pad_sequences
from keras.models import Sequential
from keras.layers import Dense, Embedding, LSTM, Dropout
from sklearn.model_selection import train_test_split
from keras.utils.np_utils import to_categorical
import re
import json
import pandas as pd
from utils import get_train_data_from_csv, get_dev_data_from_csv, get_test_data_from_csv, Indexer, get_indexer
from nltk.tokenize import TweetTokenizer
from sklearn.metrics import classification_report
include_test = True
tknr = TweetTokenizer()
indexer = get_indexer('indexer_15_dups.csv')
word_indexer = Indexer()
word_indexer.add_and_get_index("UNK")
train_data = get_train_data_from_csv('data/train_15_ds.csv')[0:1000]
dev_data = get_dev_data_from_csv('data/dev_15_ds.csv')[:200]
test_data = get_test_data_from_csv('data/test_15_ds.csv')[0:200]
X_train = []
Y_train = []
X_dev = []
Y_dev = []
Y_dev_true = []
X_test = []
Y_test = []
Y_test_true = []
def reindex_index(indexname, suffix):
"""Reindex a named index.
"""
if not hasattr(settings, 'ENABLE_SEARCHIFY') or not settings.ENABLE_SEARCHIFY:
return
models = _index_models.get(indexname, None)
if models is None:
raise KeyError("Index %r is not known" % indexname)
try:
# Get the index-wide settings.
index_settings = {}
def merge_dicts(path, a, b):
for (k, v) in b.iteritems():
if k not in a:
a[k] = v
continue
if isinstance(v, dict):
merge_dicts('%s.%s' % (path, k), a[k], v)
continue
if a[k] == v:
continue
raise ValueError("Conflicting values in index_settings (at %s)" % path[1:])
for model in models:
indexer = get_indexer(model)
merge_dicts('.', index_settings, indexer.index_settings)
created = False
for model in models:
print "Indexing %s to %s, using suffix %s" % (model, indexname, suffix)
indexer = get_indexer(model)
try:
indexer.client.set_suffix(suffix)
if not created:
#print "Creating index with settings %r" % index_settings
indexer.client.create_index(index_settings)
created = True
indexer.apply_mapping()
indexer.index_all(with_cascade=False)
finally:
indexer.client.set_suffix()
indexer.client.flush()
# Get the old value of the alias.
try:
old_index = client.get_alias(indexname)[0]
except IndexError:
old_index = None
if old_index == indexname:
# Old index wasn't an alias; we have to delete it and then set the
# new alias for it.
print "Warning: no alias in use, so must delete in-use index"
old_index = None
client.delete_index(indexname)
print "Setting alias to make new index live"
client.set_alias(indexname, indexname + suffix)
except:
try:
client.delete_index(indexname + suffix)
except Exception:
# Ignore any normal exceptions, so we report the original error.
pass
raise
if old_index:
print "Removing old index: %s" % old_index
client.delete_index(old_index)
self.indexer = Indexer()
def get_indexer(self):
return self.indexer
def extract_features(self, ex):
feature_vector = np.zeros(len(self.indexer))
for word in ex.text:
index = self.indexer.index_of(word)
feature_vector[index] += 1
return feature_vector
filename = sys.argv[1]
indexer = get_indexer('data/indexer_' + filename)
train_set = get_train_data_from_csv('data/train_' + filename)
dev_set = get_dev_data_from_csv('data/dev_' + filename)
test_set = get_test_data_from_csv('data/test_' + filename)
p = PerceptronClassifier(indexer, FeatureExtractor())
p.train(train_set)
y_pred = []
y_true = []
for ex in dev_set:
y_true.append(ex.label)
y_pred.append(p.predict(ex))
print("Dev Set Results: ")
print("Accuracy: ", accuracy_score(y_true, y_pred))
