python UI编程之实现word页码统计
用python制作了一个统计word页码的工具，能够统计选定目录下所有word文件的页码。用PyPDF实现了word页码统计功能，用PyQt设计实现了工具UI界面。
主窗口界面代码main.py
# -*- coding: utf-8 -*-
# Form implementation generated from reading ui file 'Main.ui'
#
# Created by: PyQt5 UI code generator 5.15.4
#
# WARNING: Any manual changes made to this file will be lost when pyuic5 is
# run again. Do not edit this file unless you know what you are doing.
from PyQt5 import QtCore, QtGui, QtWidgets
class Ui_Dialog(object):
def setupUi(self, Dialog):
Dialog.setObjectName("Dialog")
Dialog.resize(808, 808)
Dialog.setAutoFillBackground(True)
Dialog.setStyleSheet("")
self.pushButton = QtWidgets.QPushButton(Dialog)
self.pushButton.setGeometry(QtCore.QRect(0, 0, 231, 41))
self.pushButton.setIconSize(QtCore.QSize(30, 30))
self.pushButton.setObjectName("pushButton")
font = QtGui.QFont()
font.setFamily("Arial") # 括号里可以设置成自己想要的其它字体
font.setPointSize(10) # 括号里的数字可以设置成自己想要的字体大小
self.pushButton.setFont(font)
self.retranslateUi(Dialog)
QtCore.QMetaObject.connectSlotsByName(Dialog)
def retranslateUi(self, Dialog):
_translate = QtCore.QCoreApplication.translate
Dialog.setWindowTitle(_translate("Dialog", "Dialog"))
self.pushButton.setText(_translate("Dialog", "开始统计页码"))
主窗口功能代码WordHelper.py
import sys
from PyQt5.QtGui import QColor, QBrush, QPixmap
from PyQt5.QtWidgets import QApplication, QMainWindow, QFileDialog
import os
from PageWindow import PageWindow
from mainwindow import *
from Page import *
class MainWindow(QMainWindow, Ui_Dialog):
def __init__(self):
super(MainWindow, self).__init__()
self.setupUi(self)
self.setGeometry(100,100,1024,600)
self.setWindowTitle('Word助手')
palette = QtGui.QPalette()
palette.setBrush(self.backgroundRole(), QBrush(QPixmap("./win10.jpg").scaled(self.size(), QtCore.Qt.IgnoreAspectRatio, QtCore.Qt.SmoothTransformation)))
self.setPalette(palette)
self.setAutoFillBackground(True)
self.setFixedSize(1024,600)
if __name__ == '__main__':
app = QApplication(sys.argv)
main = MainWindow()
main.show()
pageWindow = PageWindow()
main.pushButton.clicked.connect(pageWindow.openPage)
sys.exit(app.exec_())
统计窗口界面代码Page.py
# -*- coding: utf-8 -*-
# Form implementation generated from reading ui file 'Page.ui'
#
# Created by: PyQt5 UI code generator 5.15.4
#
# WARNING: Any manual changes made to this file will be lost when pyuic5 is
# run again. Do not edit this file unless you know what you are doing.
from PyQt5 import QtCore, QtGui, QtWidgets
class Ui_pageWindow(object):
def setupUi(self, pageWindow):
pageWindow.setObjectName("pageWindow")
pageWindow.resize(792, 811)
self.label = QtWidgets.QLabel(pageWindow)
self.label.setGeometry(QtCore.QRect(40, 30, 71, 21))
font = QtGui.QFont()
font.setPointSize(10)
self.label.setFont(font)
self.label.setObjectName("label")
self.listWidget = QtWidgets.QListWidget(pageWindow)
self.listWidget.setGeometry(QtCore.QRect(40, 110, 681, 231))
self.listWidget.setObjectName("listWidget")
self.tableWidget = QtWidgets.QTableWidget(pageWindow)
self.tableWidget.setGeometry(QtCore.QRect(40, 410, 681, 281))
self.tableWidget.setObjectName("tableWidget")
self.tableWidget.setColumnCount(0)
self.tableWidget.setRowCount(0)
self.totalLabel = QtWidgets.QLabel(pageWindow)
self.totalLabel.setGeometry(QtCore.QRect(40, 720, 251, 21))
font = QtGui.QFont()
font.setPointSize(10)
self.totalLabel.setFont(font)
self.totalLabel.setObjectName("totalLabel")
self.line = QtWidgets.QFrame(pageWindow)
self.line.setGeometry(QtCore.QRect(20, 350, 721, 21))
self.line.setFrameShape(QtWidgets.QFrame.HLine)
self.line.setFrameShadow(QtWidgets.QFrame.Sunken)
self.line.setObjectName("line")
self.line_2 = QtWidgets.QFrame(pageWindow)
self.line_2.setGeometry(QtCore.QRect(10, 20, 20, 331))
self.line_2.setFrameShape(QtWidgets.QFrame.VLine)
self.line_2.setFrameShadow(QtWidgets.QFrame.Sunken)
self.line_2.setObjectName("line_2")
self.line_3 = QtWidgets.QFrame(pageWindow)
self.line_3.setGeometry(QtCore.QRect(730, 20, 20, 331))
self.line_3.setFrameShape(QtWidgets.QFrame.VLine)
self.line_3.setFrameShadow(QtWidgets.QFrame.Sunken)
self.line_3.setObjectName("line_3")
self.line_4 = QtWidgets.QFrame(pageWindow)
self.line_4.setGeometry(QtCore.QRect(730, 370, 20, 331))
self.line_4.setFrameShape(QtWidgets.QFrame.VLine)
self.line_4.setFrameShadow(QtWidgets.QFrame.Sunken)
self.line_4.setObjectName("line_4")
self.line_5 = QtWidgets.QFrame(pageWindow)
self.line_5.setGeometry(QtCore.QRect(10, 370, 20, 331))
self.line_5.setFrameShape(QtWidgets.QFrame.VLine)
self.line_5.setFrameShadow(QtWidgets.QFrame.Sunken)
self.line_5.setObjectName("line_5")
self.line_6 = QtWidgets.QFrame(pageWindow)
self.line_6.setGeometry(QtCore.QRect(20, 700, 721, 21))
self.line_6.setFrameShape(QtWidgets.QFrame.HLine)
self.line_6.setFrameShadow(QtWidgets.QFrame.Sunken)
self.line_6.setObjectName("line_6")
self.line_7 = QtWidgets.QFrame(pageWindow)
self.line_7.setGeometry(QtCore.QRect(20, 10, 721, 21))
self.line_7.setFrameShape(QtWidgets.QFrame.HLine)
self.line_7.setFrameShadow(QtWidgets.QFrame.Sunken)
self.line_7.setObjectName("line_7")
self.layoutWidget = QtWidgets.QWidget(pageWindow)
self.layoutWidget.setGeometry(QtCore.QRect(40, 50, 661, 30))
self.layoutWidget.setObjectName("layoutWidget")
self.horizontalLayout = QtWidgets.QHBoxLayout(self.layoutWidget)
self.horizontalLayout.setContentsMargins(0, 0, 0, 0)
self.horizontalLayout.setObjectName("horizontalLayout")
self.label_2 = QtWidgets.QLabel(self.layoutWidget)
font = QtGui.QFont()
font.setPointSize(8)
self.label_2.setFont(font)
self.label_2.setObjectName("label_2")
self.horizontalLayout.addWidget(self.label_2)
self.lineEdit = QtWidgets.QLineEdit(self.layoutWidget)
self.lineEdit.setObjectName("lineEdit")
self.horizontalLayout.addWidget(self.lineEdit)
self.openButton = QtWidgets.QPushButton(self.layoutWidget)
font = QtGui.QFont()
font.setPointSize(7)
self.openButton.setFont(font)
self.openButton.setObjectName("openButton")
self.horizontalLayout.addWidget(self.openButton)
self.layoutWidget1 = QtWidgets.QWidget(pageWindow)
self.layoutWidget1.setGeometry(QtCore.QRect(40, 370, 681, 30))
self.layoutWidget1.setObjectName("layoutWidget1")
self.horizontalLayout_2 = QtWidgets.QHBoxLayout(self.layoutWidget1)
self.horizontalLayout_2.setContentsMargins(0, 0, 0, 0)
self.horizontalLayout_2.setObjectName("horizontalLayout_2")
self.label_3 = QtWidgets.QLabel(self.layoutWidget1)
font = QtGui.QFont()
font.setPointSize(10)
self.label_3.setFont(font)
self.label_3.setObjectName("label_3")
self.horizontalLayout_2.addWidget(self.label_3)
spacerItem = QtWidgets.QSpacerItem(40, 20, QtWidgets.QSizePolicy.Expanding, QtWidgets.QSizePolicy.Minimum)
self.horizontalLayout_2.addItem(spacerItem)
self.resultButton = QtWidgets.QPushButton(self.layoutWidget1)
font = QtGui.QFont()
font.setPointSize(7)
self.resultButton.setFont(font)
self.resultButton.setObjectName("resultButton")
self.horizontalLayout_2.addWidget(self.resultButton)
self.retranslateUi(pageWindow)
QtCore.QMetaObject.connectSlotsByName(pageWindow)
def retranslateUi(self, pageWindow):
_translate = QtCore.QCoreApplication.translate
pageWindow.setWindowTitle(_translate("pageWindow", "Form"))
self.label.setText(_translate("pageWindow", "源"))
self.totalLabel.setText(_translate("pageWindow", "合计页码："))
self.label_2.setText(_translate("pageWindow", "请选择word文档所在目录："))
self.openButton.setText(_translate("pageWindow", "打开目录"))
self.label_3.setText(_translate("pageWindow", "结果"))
self.resultButton.setText(_translate("pageWindow", "开始统计"))
统计窗口功能代码PageWindow.py
import _thread
import os
import threading
from PyQt5.QtWidgets import QMainWindow, QFileDialog, QTableWidgetItem
import mytool.wordtopdf as wordtopdf
from Page import *
class PageWindow(QMainWindow, Ui_pageWindow):
def __init__(self):
super(PageWindow, self).__init__()
self.setupUi(self)
self.setWindowTitle("统计页码")
self.tableWidget.setColumnWidth(0,600)
self.tableWidget.setColumnWidth(1,100)
self.tableWidget.setColumnCount(2)
item = QtWidgets.QTableWidgetItem()
item.setText("文件名")
self.tableWidget.setHorizontalHeaderItem(0, item)
item = QtWidgets.QTableWidgetItem()
item.setText("页数")
self.tableWidget.setHorizontalHeaderItem(1, item)
self.openButton.clicked.connect(self.openDir)
self.resultButton.clicked.connect(self.calculateClick)
self.file_list = []
def openDir(self):
dirStr = QFileDialog.getExistingDirectory(self, "打开目录或文件", "E:\资料\论文");
self.lineEdit.setText(dirStr)
self.getAllWordFiles(self.lineEdit.text())
self.listWidget.addItems(self.file_list)
def openPage(self):
self.show()
def getAllWordFiles(self, dirStr):
for file in os.listdir(dirStr):
suffix = file.split('.')[-1]
if suffix=='doc' or suffix=='docx':
file_path = os.path.join(dirStr, file)
self.file_list.append(file_path)
def calculateClick(self):
self.totalLabel.setText("正在统计......")
_thread.start_new_thread(self.calculate, ())
def calculate(self):
value_list = wordtopdf.wordtopdf1(self.file_list)
print (value_list)
total_pages = str(value_list[0])
self.totalLabel.setText("合计页码为：" + total_pages)
result_list = value_list[1]
self.tableWidget.setRowCount(len(result_list))
self.tableWidget.setColumnCount(2)
for i in range(self.tableWidget.rowCount()):
for j in range(self.tableWidget.columnCount()):
content = result_list[i][j]
newItem = QTableWidgetItem(content)
self.tableWidget.setItem(i, j, newItem)
PyPDF统计页面代码--word2pdf.py
# -*- coding:utf-8 -*-
import os
from time import sleep
from win32com.client import Dispatch, DispatchEx # 导入pywin32模块的client包下的函数
from win32com.client import constants # 导入pywin32模块的client包下的保存COM常量的类
from win32com.client import gencache # 导入pywin32模块的client包下的gencache函数
from PyPDF2 import PdfFileReader # 获取页码用
import pythoncom # 导入封装了OLE自动化API的模块，该模块为pywin32的子模块
totalPages = 0 # 记录总页数的全局变量
returnlist = [] # 保存文件列表的全局变量
# Word转换为PDF(多个文件)
def wordtopdf(filelist,targetpath):
valueList = []
try:
pythoncom.CoInitialize() # 调用线程初始化COM库，解决调用Word 2007时出现“尚未调用CoInitialize”错误的问题
gencache.EnsureModule('{00020905-0000-0000-C000-000000000046}', 0, 8, 4)
# 开始转换
w = Dispatch("Word.Application")
for fullfilename in filelist:
temp = fullfilename.split('\\')
path = temp[0]
softfilename = os.path.splitext(temp[1])
filename = temp[1]
os.chdir(path)
doc = os.path.abspath(filename)
# filename, ext = os.path.splitext(doc)
os.chdir(targetpath)
pdfname = softfilename[0] + ".pdf"
output = os.path.abspath(pdfname)
pdf_name = output
# 文档路径需要为绝对路径，因为Word启动后当前路径不是调用脚本时的当前路径。
try:
doc = w.Documents.Open(doc, ReadOnly=1)
doc.ExportAsFixedFormat(output, constants.wdExportFormatPDF, \
Item=constants.wdExportDocumentWithMarkup,
CreateBookmarks=constants.wdExportCreateHeadingBookmarks)
except Exception as e:
print(e)
if os.path.isfile(pdf_name):
valueList.append(pdf_name)
else:
print('转换失败！')
return False
w.Quit(constants.wdDoNotSaveChanges)
return valueList
except TypeError as e:
print('出错了！')
print(e)
return -1
# Word转换为PDF并提取页码
def wordtopdf1(filelist):
# global totalPages # 全局变量
totalPages = 0
valueList = []
try:
pythoncom.CoInitialize() # 调用线程初始化COM库，解决调用Word 2007时出现“尚未调用CoInitialize”错误的问题
gencache.EnsureModule('{00020905-0000-0000-C000-000000000046}', 0, 8, 4)
# 开始转换
w = Dispatch("Word.Application")
for fullfilename in filelist:
temp = fullfilename.split('\\')
path = temp[0]
filename = temp[1]
os.chdir(path)
doc = os.path.abspath(filename)
filename, ext = os.path.splitext(doc)
output = filename + '.pdf'
a = os.path.join(path ,"pdf")
pdf_name = output
# 文档路径需要为绝对路径，因为Word启动后当前路径不是调用脚本时的当前路径。
try:
sleep(1)
doc = w.Documents.Open(doc, ReadOnly=1)
doc.ExportAsFixedFormat(output, constants.wdExportFormatPDF, \
Item=constants.wdExportDocumentWithMarkup,
CreateBookmarks=constants.wdExportCreateHeadingBookmarks)
except Exception as e:
print(e)
if os.path.isfile(pdf_name):
# 获取页码
pages = getPdfPageNum(pdf_name) # 获取页码
valueList.append([fullfilename,str(pages)])
totalPages += pages # 累加页码
os.remove(pdf_name) # 删除生成的PDF文件
else:
print('转换失败！')
return False
w.Quit(constants.wdDoNotSaveChanges)
return totalPages,valueList
except TypeError as e:
print('出错了！')
print(e)
return -1
####################### 统计页码 ############################################
def getPdfPageNum(path):
with open(path, "rb") as file:
doc = PdfFileReader(file)
pagecount = doc.getNumPages()
return pagecount
####################### 提取目录 ############################################
def getPdfOutlines(pdfpath,listpath,isList):
print("提取目录")
with open(pdfpath, "rb") as file:
doc = PdfFileReader(file)
outlines = doc.getOutlines() # 获取大纲
global returnlist # 全局变量，保存大纲的列表
returnlist = [] # 创建一个空列表
mylist = getOutline(outlines,isList) # 递归获取大纲
w = DispatchEx("Word.Application") # 创建Word文档应用程序对象
w.Visible = 1
w.DisplayAlerts = 0
doc1 = w.Documents.Add()# 添加一个Word文档对象
range1 = doc1.Range(0,0)
for item in mylist: # 通过循环将获取的目录列表插入到Word文档对象中
range1.InsertAfter(item)
outpath = os.path.join(listpath,'list.docx') # 连接Word文档路径
doc1.SaveAs(outpath) # 保存文件
doc1.Close() # 关闭Word文档对象
w.Quit() # 退出Word文档应用程序对象
return outpath
def getOutline(obj,isList): # 递归获取大纲
global returnlist
for o in obj:
if type(o).__name__ == 'Destination':
# mypage = getRealPage(doc, pagecount, o.get('/Page').idnum)
if isList: # 包括页码
returnlist.append( o.get('/Title') + "\t\t" + str(o.get('/Page') + 1) + "\n")
else: # 不包括页码
returnlist.append(o.get('/Title') + "\n")
elif type(o).__name__ == 'list':
getOutline(o,isList) # 递归调用获取大纲
return returnlist
最终效果
