svm

import spacy
nlp = spacy.load('en_core_web_lg')
#from scipy import spatial
from scipy.spatial.distance import cosine
import numpy as np
from sklearn.manifold import TSNE
#import matplotlib.pyplot as plt
import pylab as pl
#%pylab inline
from sklearn import svm
import random

# from random import randint
# import numpy


oldf = open('/Users/apple/PycharmProjects/helloworld/wordlistmini', 'r')
newf = open('/Users/apple/PycharmProjects/helloworld/wordlistmini1', 'w')
n = 0
resultList = random.sample(range(0, 673), 4)  # sample(x,y)函数的作用是从序列x中，随机选择y个不重复的元素。
print("resultList---", resultList)
lines = oldf.readlines()
matrix = []
for i in resultList:
    newf.write(lines[i])
    matrix.append(lines[i])

oldf.close()
newf.close()
print("..........")

for i in range(4):
    matrix[i] = matrix[i].strip('\n')

print("matrix---", matrix)

print("...................")
print()

"""
dog = nlp.vocab["love"]
cat = nlp.vocab["hate"]
apple = nlp.vocab["apple"]
orange = nlp.vocab["orange"]
print(dog.vector)
print(dog.similarity(cat))

def vector_similarity(x, y):
    return 1 - cosine(x, y)

print(vector_similarity(dog.vector, cat.vector))
"""
embedding = np.array([])
word_list = []
#text = "apple banana pink dog"
text = " ".join(matrix)
doc = nlp(text)

for token in doc:
    if not(token.is_punct) and not(token.text in word_list):
        word_list.append(token.text)

print(doc[0].vector)

for word in word_list:
    embedding = np.append(embedding, nlp.vocab[word].vector)

print(embedding.shape)
embedding = embedding.reshape(len(word_list), -1)
print(embedding.shape)
print(embedding[0])

def vector_similarity(x, y):
    return 1 - cosine(x, y)

print(vector_similarity(embedding[0],embedding[1]))

tsne = TSNE(init='pca',random_state=1)#

low_dim_embedding = tsne.fit_transform(embedding)

x = low_dim_embedding#
#np.r_[np.random.randn(2, 2) - [2, 2], np.random.randn(2, 2) + [2, 2]]
y = [0] * 2 + [1] * 2  # 标签
print(x)
print(y)


print(vector_similarity(x[0], x[1]))

clf = svm.SVC(kernel='linear')  # linear rbf  poly
clf.fit(x, y)  # 训练模型
print(clf.support_)  # 支持向量对应的下表
print()
#print(clf.support_vectors_)  # 所有的支持向量

w = clf.coef_[0]
a = -w[0] / w[1]

xx = np.linspace(-120, 120)  # 产生-5到5的线性连续值，间隔为1
yy = a * xx - (clf.intercept_[0]) / w[1]
# clf.intercept_[0]是w3.即为公式a1*x1+a2*x2+w3中的w3。
# (clf.intercept_[0])/w[1]即为直线的截距
#X = [(x[0] + x[1]) / 2]
#Y = [0]
#X = np.array(X)
#print(X)
# 得出支持向量的方程
b = clf.support_vectors_[0]
yy_down = a * xx + (b[1] - a * b[0])  # (b[1]-a*b[0])就是简单的算截距
b = clf.support_vectors_[-1]
yy_up = a * xx + (b[1] - a * b[0])

print("w:", w)  # 打印出权重系数
print("a:", a)  # 打印出斜率
#print("suport_vectors_:", clf.support_vectors_)  # 打印出支持向量
# print("clf.coef_:",clf.coef_) #打印出权重系数，还是w



"""def plot_with_labels(low_dim_embs, labels, filename='tsne.pdf'):
    assert low_dim_embs.shape[0] >= len(labels), "More labels than embeddings"
    plt.figure(figsize=(18, 18))  # in inches
    for i, label in enumerate(labels):
        x, y = low_dim_embs[i, :]

        plt.scatter(x, y)
        plt.annotate(label,
                 xy=(x, y),
                 xytext=(5, 2),
                 textcoords='offset points',
                 ha='right',
                 va='bottom')
    plt.savefig(filename)
plot_with_labels(low_dim_embedding, word_list)

"""
# 这个就是画出来
pl.plot(xx, yy, 'k-')
#pl.plot(xx, yy_down, 'k--')
#pl.plot(xx, yy_up, 'k--')




# 支持向量
pl.scatter(clf.support_vectors_[:, 0], clf.support_vectors_[:, 1], s=100)
# 全部的点
pl.scatter(x[:, 0], x[:, 1], c=y, cmap=pl.cm.Paired, marker='o')

#pl.scatter(X[:, 0], X[:, 1], c='y', cmap=pl.cm.Paired, marker='x')

pl.savefig('LinearSVC.png')



pl.axis('tight')
pl.show()