14-操作PDF文件

14-操作PDF文件

例子1：读取PDF，抽取文字

首先安装PyPDF2

pip install PyPDF2

import PyPDF2
from PyPDF2 import PageObject

pdf_reader = PyPDF2.PdfReader("resources/XGBooster.pdf")
pdf_writer = PyPDF2.PdfWriter()
for page_num in range(len(pdf_reader.pages)):
    current_page = pdf_reader.pages[page_num]  # type:PageObject
    # 从页面中读取文字
    # print(current_page.extract_text())
    current_page.rotate(90)  # 旋转90度
    pdf_writer.add_page(current_page)  # 在writer里加上旋转后的页面
    pdf_writer.add_blank_page()  # 加上空白页
with open("resources/XGBooster_fixed.pdf", "wb") as file:
    pdf_writer.write(file)  # 保存

例子2：对pdf加密

import PyPDF2

pdf_reader = PyPDF2.PdfReader("resources/XGBooster.pdf")
pdf_writer = PyPDF2.PdfWriter()
for page_num in range(len(pdf_reader.pages)):
    pdf_writer.add_page(pdf_reader.pages[page_num])
pdf_writer.encrypt("abc@1234")
with open("resources/XGBooster-encrypt.pdf", "wb") as file:
    pdf_writer.write(file)

例子3：加水印

水印的pdf可以现场用斜着的文本框现场做一个

import PyPDF2

reader1 = PyPDF2.PdfReader("resources/XGBooster.pdf")
reader2 = PyPDF2.PdfReader("resources/kayotin_water.pdf")
writer = PyPDF2.PdfWriter()

water_mark = reader2.pages[0]
for page_num in range(len(reader1.pages)):
    current_page = reader1.pages[page_num]
    current_page.merge_page(water_mark)
    writer.add_page(current_page)

with open("resources/Boost_water.pdf", "wb") as file:
    writer.write(file)

读取文件夹中的文件

import os

for file in os.listdir("resources"):
    print(file, os.path.isdir("resources/" + file), os.path.isfile("resources/" + file))

💡 注意，如果判断是否是文件或文件夹，要指定路径。因为返回的文件列表是不带路径的。

例子4：创建PDF文档，插入图片

用到reportlab这个库

from reportlab.lib.pagesizes import A4
from reportlab.pdfbase import pdfmetrics
from reportlab.pdfbase.ttfonts import TTFont
from reportlab.pdfgen import canvas

pdf_canvas = canvas.Canvas("resources/demo.pdf", pagesize=A4)
width, height = A4

# 画图
image = canvas.ImageReader("resources/emp_pic.png")
pdf_canvas.drawImage(image, 20, height - 200, 400, 200)
# 图片对象，位置x轴，位置y轴， 图片宽，图片高

# 显示当前页
pdf_canvas.showPage()

# 注册字体文件，否则写不出中文
pdfmetrics.registerFont(TTFont("Font1", "resources/font/msyh.ttc"))

# 写字
pdf_canvas.setFont("Font1", 40)
pdf_canvas.setFillColorRGB(1, 0, 0, 1)  # 前三个值是rgb，最后是透明度
pdf_canvas.rotate(18)  # 旋转，此处是逆时针
pdf_canvas.drawString(width // 2 - 120, height // 2, "你好，世界")


# 保存
pdf_canvas.save()