main.py

#coding:utf-8
import win32com
import win32con
import win32gui
import codecs
from win32com.client import Dispatch
import pythoncom
from gensim.models import Word2Vec
from pprint import pprint
import re, collections
import string
import os
import matplotlib.pyplot as plt

protocol = '38101-1'
tmpfile = 'tmp.txt'
txtfile = protocol + '.txt'
docfile = protocol + '.doc'
path = 'D:/Github/Vis4Protocol/procotol/'

class MSOffice2txt():
    def __init__(self, fileType=['doc']):
        self.docCom = None
        pythoncom.CoInitialize()
        if type(fileType) is not list:
            return 'Error, please check the fileType, it must be list[]'
        for ft in fileType:
            if ft == 'doc':
                self.docCom = self.docApplicationOpen()

    def close(self):
        self.docApplicationClose(self.docCom)

    def docApplicationOpen(self):
        docCom = win32com.client.Dispatch('Word.Application')
        docCom.Visible = 1
        docCom.DisplayAlerts = 0
        docHwnd = win32gui.FindWindow(None, 'Microsoft Word')
        win32gui.ShowWindow(docHwnd, win32con.SW_HIDE)
        return docCom

    def docApplicationClose(self,docCom):
        if docCom is not None:
            docCom.Quit()

    def doc2Txt(self, docCom, docFile, txtFile):
        doc = docCom.Documents.Open(FileName=docFile,ReadOnly=1)
        doc.SaveAs(txtFile, 2)
        doc.Close()

    def translate(self, filename, txtFilename):
        if filename.endswith('doc') or filename.endswith('docx'):
            if self.docCom is None:
                self.docCom = self.docApplicationOpen()
            self.doc2Txt(self.docCom, filename, txtFilename)
            return True
        else:
            return False

    def punctuationDelete(self, txtFilename1, txtFilename2):
    	with open(txtFilename1, 'r') as f:
    		with open(txtFilename2, 'a') as t:
    			for line in f:
    				t.write(re.sub('[%s]' % re.escape(string.punctuation), '', line) + '\n')
    	os.remove(txtFilename1)


class MyCorpus(object):
	def __init__(self, fname):
		self.fname = fname

	def __iter__(self):
		for line in open(self.fname):
			yield line.lower().split()


if __name__ == '__main__':
	msoffice = MSOffice2txt()
	if msoffice.translate(path + docfile, path + tmpfile):
	    print('Successed!')
	else:
	    print('Failed!')
	msoffice.punctuationDelete(path + tmpfile, path + txtfile)
	msoffice.close()

	with open(path + txtfile, 'r') as f:
		file = f.read()
		words = [word for word in file.split() if re.findall(r'\w', word)]
	print(len(words))

	sentences = MyCorpus(path + txtfile)
	model = Word2Vec(sentences, size=2, min_count=1)

	x, y = 0, 0
	for word in words:
		say_vector = model[word.lower()]  # get vector for word
		x += say_vector[0]
		y += say_vector[1]
	x /= len(words)
	y /= len(words)


	with open('data.txt', 'a') as f:
		f.write(protocol + ' ' + str(x) + ' ' + str(y) + ' ' + str(len(words)) + '\n')

	'''	
	plt.figure()
	plt.scatter(x, y, s=len(words), facecolors='none', edgecolors='r')
	plt.scatter(x, y, s=10, color='r')
	plt.text(x, y, '38211-002')
	plt.show()
	
	'''