Python 使用 Streamlit 提取 PDF 文档文字

Python 基于 Streamlit 实现 PDF 文档文字提取

本文介绍如何使用 Python 和 Streamlit 构建一个简单的工具，用于从 PDF 文档中提取文本内容，并支持加密 PDF 的密码输入处理。

环境依赖

安装所需的 Python 库：

pip install streamlit pdfplumber

核心代码

以下是完整的 Main.py 脚本示例。该程序实现了文件上传、文本提取（含加密处理）及简单的问答展示功能。

import os
import tempfile
import traceback
import streamlit as st
from typing import Optional, List
import pdfplumber

class DocumentQASystem:
    """私有化 PDF 文档问答系统"""
    def __init__(self):
        st.set_page_config(
            page_title="PDF 文档问答系统",
            layout="wide",
            initial_sidebar_state="collapsed"
        )
        if "doc_text" not in st.session_state:
            st.session_state.doc_text = ""
        self._set_custom_style()

    def _set_custom_style(self):
        st.markdown("""
            <style>
                .stButton>button {width: 100%; margin-top: 10px;}
                .stTextInput>div>div>input {padding: 8px;}
                .doc-preview {max-height: 400px; overflow-y: auto; border: 1px solid #eee; padding: 10px; border-radius: 5px;}
            </style>
        """, unsafe_allow_html=True)

    def _extract_pdf_text(self, pdf_path: ) -> :
        
        full_text = 
        :
            :
                 pdfplumber.(pdf_path)  pdf:
                     page_num, page  (pdf.pages, ):
                        text = page.extract_text()
                         text:
                            full_text += 
             pdfplumber.utils.PDFEncryptionError:
                st.warning()
                pdf_password = st.text_input(
                    , =, key=,
                    =
                )
                 pdf_password:
                    :
                         pdfplumber.(pdf_path, password=pdf_password)  pdf:
                             page_num, page  (pdf.pages, ):
                                text = page.extract_text()
                                 text:
                                    full_text += 
                     Exception:
                        st.error()
                         
                :
                    st.info()
                     
             full_text.strip()
         ImportError:
            st.error()
             
         Exception  e:
            st.error()
             

     () -> :
         

     ():
        st.title()
        st.divider()
        col1, col2 = st.columns([, ])
        
         col1:
            st.subheader()
            uploaded_file = st.file_uploader(, =[])
            
             uploaded_file   :
                 st.spinner():
                     tempfile.NamedTemporaryFile(delete=, suffix=)  tmp_file:
                        tmp_file.write(uploaded_file.read())
                        tmp_path = tmp_file.name
                    
                    st.session_state.doc_text = ._extract_pdf_text(tmp_path)
                    
                    :
                        os.unlink(tmp_path)
                     Exception  e:
                        st.warning()
                    
                     st.session_state.doc_text:
                        st.success()
                    
                     st.button(, =):
                        st.session_state.doc_text = 
                        st.rerun()
            
            st.divider()
            question = st.text_input(
                ,
                placeholder=,
                disabled= st.session_state.doc_text
            )
            
            submit_btn = st.button(
                ,
                =,
                disabled= (question  st.session_state.doc_text)
            )
        
         col2:
            st.subheader()
             submit_btn:
                 st.spinner():
                    answer = ._mock_llm_answer(question, st.session_state.doc_text)
                    st.markdown()
                    st.write(answer)
            
            st.divider()
             st.session_state.doc_text:
                st.markdown()
                st.markdown(
                    ,
                    unsafe_allow_html=
                )
            :
                st.info()
        
        st.divider()
        st.caption()

 ():
    qa_system = DocumentQASystem()
    qa_system.render_ui()

 __name__ == :
    main()