【Python】Docx解析

# -*- coding: utf-8 -*-

import os

import docx

from win32com import client as wc

docs = []

def traverse(f):

    fs = os.listdir(f)

    for f1 in fs:

        tmp_path = os.path.join(f,f1)

        if not os.path.isdir(tmp_path):

            #print('文件: %s'%tmp_path)

            if  os.path.splitext(tmp_path)[-1].lower() == ".doc" or os.path.splitext(tmp_path)[-1].lower() == ".docx":

                #print('文件: %s'%tmp_path)

                docs.append(tmp_path)

        else:

            #print('文件夹：%s'%tmp_path)

            traverse(tmp_path)

def parseDoc(f):

    doc = docx.Document(f)

    parag_num = 0

    for para in doc.paragraphs :

        print("----------------------------------------------------")

        print(para.text)

        print("----------------------------------------------------")

        parag_num += 1

    print ('This document has ', parag_num, ' paragraphs')

def doc2docx(full_path):

    #dirname = os.path.dirname(full_path)

    #filename = os.path.basename(full_path)

    #newpath = full_path.replace('doc','docx')

    newpath = full_path + "x"

    if os.path.exists(newpath):

        return

    # 首先将doc转换成docx

    word = wc.Dispatch("Word.Application")

    # 找到word路径 + 文件名 ，即可打开文件

    doc = word.Documents.Open(full_path)

    # 使用参数16表示将doc转换成docx，保存成docx后才能 读文件

    doc.SaveAs(newpath,16)

    doc.Close()

    word.Quit()

path = 'E:/NLP/Docs/'

traverse(path)

for k,v in enumerate(docs):

    if k < 1:

        print(k,v)

        parseDoc(v)

        #doc2docx(v)