Python3, 5 lines of code, Chatxxx can perform a series of operations such as rotating, extracting, and merging PDF files. After reading this article, the 80-year-old grandmother can walk without supporting the wall.

1 Introduction

Little Diaosi : Brother Yu, what have you been up to lately?
Xiaoyu : Recently? How close do you mean?
Little Diaosi : Just these few days?
Xiaoyu : These days I am moving bricks.
Little Diaosi : How about a few days ago?
Xiaoyu : A few days ago, during my May Day holiday, I was also moving bricks.
Little Diaosi : Brother Yu, you...

Xiaoyu : I really do.
Little Diaosi : Then change the subject, how is ChatGPT doing?
Xiaoyu : ChatGPT is not all big companies are doing it. ChatGPT can answer what you want to know.
Xiao Diaosi : Brother Yu, you...!
Xiaoyu : This is the strength of ChatGPT, and it is true.
Little Diaosi : Okay, then I want to extract the content of the PDF document.
Xiaoyu : This article: " Python3, 9-line batch extraction of specified content of PDF files, this kind of operation is guaranteed to be loved by everyone... "
Xiao Diaosi : I am a RMB player, and I want a more advanced one.
Xiaoyu : Here... Let me take a look at
Xiao Diaosi : What are you looking at?
Xiaoyu : See how much is left in your account, should it be recharged?
Little Diaosi : Brother Yu, you...!
Xiaoyu : Full, giving you the luxury you want...

Little Diaosi : Hehe... full of...

2. Code combat

2.1 Principle

When it comes to ChatPDF, the first reaction of most students is, sorry, I don't know.
However, when it comes to ChatGPT, you may say, I must know this, the product of OPenAI company, and it is so popular now, how can I not know.

Now that you know ChatGPT, ChatPDF is not difficult to understand.
In fact, ChatPDF is a derivative of ChatGPT.

Xiao Diaosi : Since it is a derivative product of ChatGPT, what is the working principle?
Xiaoyu : The working principle is not difficult, there are only 2 steps to put the elephant in the refrigerator.

• 1. ChatPDF reads the contents of PDF files and converts them into text (can be .txt) format;
• 2. ChatPDF cleans and standardizes the extracted text content, such as: segment, sentence, etc.;
• 3. Use OpenAI's Embeddings API to convert each segment into a vector, which will encode the semantics in the text for easy comparison with the question's vector;
• 4. Use OpenAI's Embeddings API to convert the question into a vector and compare with each segment's vector to find the most similar segment. This similarity calculation can be performed using common methods such as cosine similarity;
• 5. Use the most similar segment and question found as a prompt, call OpenAI's Completion API, let ChatGPT learn the segment content, and then answer the corresponding question;
• 6. The answer generated by ChatGPT will be returned to the user to complete a query.

Xiao Diaosi : Unexpectedly, the implementation process is really easy.
Xiaoyu : The big river bends to the east, Niu Niu rushes forward~~

2.2 Installation

In the previous chapter, we knew what ChatPDF is and the operation process.
Next, we have to actually operate it.

Of course, when it comes to third-party libraries, the old rules, installation starts.

pip install chatpdf

Then just wait for the installation.

For other installation methods, see these two articles directly:

• " Python3, choose Python to automatically install third-party libraries, and say goodbye to pip from now on! ! "
• 《Python3: I import all Python libraries with only one line of code in a low-key manner! "

Because chatPDF takes a while to install (I wouldn't say the process is " short ").

2.2 Examples

After the installation is complete, let's see what chatPDF can do.

2.2.1 Creating PDF files

code example

# -*- coding:utf-8 -*-
# @Time   : 2023-05-06
# @Author : Carl_DJ

'''
实现功能：
    使用chatPDF的基本方法，创建PDF文件
'''

'--------->创建PDF文件<---------'
from chatpdf import ChatPDF

#文件名字
file_name = './data/TestDemo.pdf'
pdf = ChatPDF()
#添加页数
pdf.add_page()
#设置字体
pdf.set_font("Arial", size=12)
#设置内容
pdf.cell(200,10,txt='Hello, Python')
#输出内容
pdf.output(file_name)

2.2.2 Rotating PDF files

code example

# -*- coding:utf-8 -*-
# @Time   : 2023-05-06
# @Author : Carl_DJ

'''
实现功能：
    使用chatPDF的基本方法，实现旋转PDF文件内容
'''

'--------->旋转PDF文件内容<---------'
from chatpdf import rotate_pages

#pdf源文件
pdf_file = './data/input.pdf'

#输出的文件
output_file = './data/output.pdf'

#旋转的页码
pages = [1, 3]

#旋转角度设置
rotation_angle = 270

rotate_pages(pdf_file, output_file, pages, rotation_angle)

2.2.3 Split PDF file

code example

# -*- coding:utf-8 -*-
# @Time   : 2023-05-06
# @Author : Carl_DJ

'''
实现功能：
    使用chatPDF的基本方法，实现拆分PDF文件内容
'''

'--------->拆分PDF文件<---------'
from chatpdf import split

#需要拆分的pdf源文件
pdf_file = 'input_demo.pdf'

#拆分后的pdf保存的文件夹
output_folder = './data/output'

split(pdf_file,output_folder)

2.2.4 Merge PDF files

code example

# -*- coding:utf-8 -*-
# @Time   : 2023-05-06
# @Author : Carl_DJ

'''
实现功能：
    使用chatPDF的基本方法，实现合并PDF文件内容
'''

'--------->合并PDF文件<---------'
#获取所有文件
file1 = './data/demo1.pdf'
file2 = './data/demo2.pdf'
file3 = './data/demo3.pdf'

#列表展示所有需要合并的pdf文件
pdf_file_list = [file1,file2,file3]
#合并后输出文件名称
output_file = 'output_demo.pdf'

merge(pdf_file_list,output_file)

2.2.5 Extract PDF file content

code example

# -*- coding:utf-8 -*-
# @Time   : 2023-05-06
# @Author : Carl_DJ

'''
实现功能：
    使用chatPDF的基本方法，实现提取PDF文件内容
'''

'--------->提取PDF文件<---------'
from chatpdf import extract_pages

#待提取的pdf源文件
pdf_file = 'input.pdf'

#输出提取的pdf文件内容
output_file = './data/output.pdf'
#提取的源文件的页码
pages = [1,3,5,7,10]

extract_pages(pdf_file, output_file, pages)

Little Diaosi : Brother Yu, I remember you also wrote a blog post about extracting the contents of PDF documents.
Xiaoyu : Well, your memory is quite good, and you did write it, which is the following article.

• " Python3, 9 lines to extract the specified content of PDF files in batches, this operation is guaranteed to be loved by everyone... "

Of course, regarding the operation of PDF documents, you can also read other blog posts of Xiaoyu:

• " Python3: For 200 yuan of living expenses, I cracked my wife's encrypted PDF file, but when I saw the content, I... panicked~~ "
• 《Python3, 2 lines of code, multiple methods, directly convert web page content into PDF documents and pictures. "
• 《Python3, 2 pieces of code, add watermark to pdf file, it turns out that watermark can also play like this . "

3. Summary

Seeing this, the introduction of the ChatPDF library is complete.
In fact, the functions of the ChatPDF library are not only the functions I mentioned, but also:

• PDF file encryption ;
• PDF file decryption ;
• …

At the moment when ChatGPT is exploding, as a technical er, we more or less need to know some AI knowledge.
Even if it is white ~ whoring ~ the number of times ChatGPT is used, it is also possible.
Of course, as a small fish, I will definitely not do (bai piao) things, but the pace of learning has never stopped.

I am a small fish :

• CSDN blog expert ;
• Aliyun expert blogger ;
• 51CTO blog expert ;
• 51 Certified Instructor ;
• Certified gold interviewer ;
• workplace training planner ;

Follow me and let you learn more and more interesting Python knowledge.