def ana_input(self,input_data):
self.logger.info("in ana_input")
flows = input_data['flows']
for flow in flows:
args = flow.split(' ')
src = args[0]
dst = args[1]
try:
port = int(args[2])
except:
port = None
old = tools.str_to_list(args[3])
new = tools.str_to_list(args[4])
self.logger.info(new)
trans_type = args[5]
ratio = float(args[6])
bw = int(args[7])
flow = FlowDesGlobal(src,dst,port,old,new,None,trans_type)
flow.ratio = ratio
flow.bw = bw
self.logger.info(flow.new)
self.logger.info(flow.trans_type)
self.flows_to_schedule.update({src+dst+str(port):flow})
self.logger.info(self.flows_to_schedule)
self.logger.info("---------------------start!---------------------")
self.logger.info(nowTime())
def embedding(csv_path):
train = pd.read_csv(csv_path)
x_data = []
for sentence in train.new_article.values:
data = str_to_list(sentence)
x_data.append(data)
w2v_model = gensim.models.word2vec.Word2Vec(size=100,
window=5,
min_count=2)
w2v_model.build_vocab(x_data)
words = w2v_model.wv.vocab.keys()
vocab_size = len(words)
print("Vocab size", vocab_size)
# Train Word Embeddings
w2v_model.train(x_data, total_examples=len(x_data), epochs=100)
return w2v_model
import os
from model import base_model
from pathlib import Path
from tools import str_to_list, text2sequence, glove_word2vec
from tensorflow.keras.callbacks import ModelCheckpoint, CSVLogger
DATA_PATH = Path("data")
max_len = 100
epochs = 10
batch_size = 128
train = pd.read_csv(DATA_PATH / "prep_news_train.csv")
word2vec = gensim.models.Word2Vec.load('./data/news_min0.embedding')
train_data = [ str_to_list(sentence) for sentence in train.new_article.values ]
train_X, vocab_size, tokenizer = text2sequence(train_data, max_len = max_len)
train_y = train['info']
word_index = tokenizer.word_index
embedding_matrix = np.zeros((vocab_size, max_len))
for word, index in word_index.items():
if word in word2vec:
embedding_vector = word2vec[word]
embedding_matrix[index] = embedding_vector
else:
print("word2vec에 없는 단어입니다.")
break
import pandas as pd
import numpy as np
from pathlib import Path
from tools import str_to_list, text2sequence
DATA_PATH = Path("data")
train = pd.read_csv(DATA_PATH / "prep_news_train.csv")
x_data = []
for sentence in train.new_article.values:
data = str_to_list(sentence)
x_data.append(data)
train_X, vocab_size, tokenizer = text2sequence(x_data, max_len = 100)
train_y = train['info']
word_index = tokenizer.word_index
for i in word_index.items():
print(i)
break