Streamlit实现多PDF文件动态分Tab展示问题求助

需求说明 #

开发Streamlit应用，实现以下功能：

• 支持上传多个PDF文件
• 选中文件时，以文件名为标题动态创建标签页展示PDF内容
• 取消选中文件时，自动关闭对应标签页

原代码问题 #

原代码未实现动态添加/移除标签页的预期效果，核心问题：

1. st.tabs使用方式错误，循环绑定内容的逻辑不符合Streamlit组件渲染规则
2. 未处理文件取消选中的状态同步，也未利用Session State保存上传文件的状态

原代码如下：

import streamlit as st
import os
from PyPDF2 import PdfReader
from io import BytesIO

# Function to read PDF and return content
def read_pdf(file_path):
    # Replace with your PDF reading logic
    pdf = PdfReader(file_path)
    return pdf

# Function to display PDF content
def display_pdf(pdf):
    num_pages = pdf.page_count

    # Display navigation input for selecting a specific page
    page_number = st.number_input(
        "Enter page number", value=1, min_value=1, max_value=num_pages
    )

    # Display an image for the selected page
    image_bytes = pdf[page_number - 1].get_pixmap().tobytes()
    st.image(image_bytes, caption=f"Page {page_number} of {num_pages}")

# Main Streamlit app code
st.title("PDF Viewer App")

# Get uploaded files
selected_files = st.file_uploader("Upload PDF files", type=["pdf"], accept_multiple_files=True)

# List to store content of all pages for each file
all_files_content = []

# Dictionary to store selected file content
selected_file_content = {}

# Iterate over uploaded files
for uploaded_file in selected_files:
    file_content = uploaded_file.read()

    temp_file_path = f"./temp/{uploaded_file.name}"
    os.makedirs(os.path.dirname(temp_file_path), exist_ok=True)

    with open(temp_file_path, "wb") as temp_file:
        temp_file.write(file_content)

    if uploaded_file.type == "application/pdf":
        # Read PDF and store content
        pdf = read_pdf(temp_file_path)
        all_files_content.append(pdf)

        # Display PDF content
        selected_file_content[uploaded_file.name] = pdf

        # Cleanup: Remove the temporary file
        os.remove(temp_file_path)

# Create tabs dynamically for each file
with st.tabs(list(selected_file_content.keys())):
    for file_name, pdf in selected_file_content.items():
        display_pdf(pdf)

修正后的代码 #

import streamlit as st
from PyPDF2 import PdfReader
from io import BytesIO

# 初始化Session State，保存已上传的PDF文件
if "uploaded_pdfs" not in st.session_state:
    st.session_state.uploaded_pdfs = {}

# 读取PDF文件（直接使用BytesIO，无需临时文件）
def read_pdf(file_bytes):
    return PdfReader(BytesIO(file_bytes))

# 展示PDF内容
def display_pdf(pdf, tab_key):
    num_pages = len(pdf.pages)
    page_number = st.number_input(
        "输入页码", value=1, min_value=1, max_value=num_pages, key=f"page_num_{tab_key}"
    )
    # 获取页面图片字节流（需确保已安装pdf2image依赖）
    page = pdf.pages[page_number - 1]
    pix = page.get_pixmap()
    st.image(pix.tobytes(), caption=f"第 {page_number} 页 / 共 {num_pages} 页")

st.title("PDF 标签页查看器")

# 上传文件组件
uploaded_files = st.file_uploader("上传PDF文件", type=["pdf"], accept_multiple_files=True)

# 处理上传的文件，存入Session State
for file in uploaded_files:
    if file.name not in st.session_state.uploaded_pdfs:
        file_bytes = file.read()
        pdf = read_pdf(file_bytes)
        st.session_state.uploaded_pdfs[file.name] = pdf

# 显示已上传文件的多选框，控制要展示的文件
available_files = list(st.session_state.uploaded_pdfs.keys())
selected_files = st.multiselect("选择要查看的PDF文件", available_files, default=available_files)

# 动态创建标签页
if selected_files:
    tabs = st.tabs(selected_files)
    for idx, tab in enumerate(tabs):
        file_name = selected_files[idx]
        with tab:
            display_pdf(st.session_state.uploaded_pdfs[file_name], file_name)
else:
    st.info("请上传并选择PDF文件以查看")

关键改进点 #

• Session State 持久化：用st.session_state保存已上传的PDF，避免页面刷新后丢失文件状态
• 动态标签页逻辑：通过st.tabs返回的标签对象逐个绑定对应PDF的展示内容，符合Streamlit组件渲染规则
• 取消选中自动关闭标签：通过st.multiselect控制选中文件列表，未选中的文件不会生成标签页，实现自动关闭效果
• 移除临时文件：直接用BytesIO读取上传文件的字节流，无需创建本地临时文件，更高效安全

内容的提问来源于stack exchange，提问作者Mejdi Dallel