Solution for visualising PDFs via URL

[ArthurTech123]

version information:

• streamlit_pdf_viewer: 0.0.19
• python: 3.11.7
• streamlit: 1.39.0

problem

I try to input my s3 pdf file path, but get the error from terminal

Traceback (most recent call last):
  File "c:\Users\Demo\Desktop\GitHub\demo\.venv\Lib\site-packages\streamlit\runtime\scriptrunner\exec_code.py", line 88, in exec_func_with_error_handling
    result = func()
             ^^^^^^
  File "c:\Users\Demo\Desktop\GitHub\demo\.venv\Lib\site-packages\streamlit\runtime\scriptrunner\script_runner.py", line 579, in code_to_exec
    exec(code, module.__dict__)
  File "C:\Users\Demo\Desktop\GitHub\demo\app.py", line 539, in <module>
    Flow(
  File "C:\Users\Demo\Desktop\GitHub\demo\app.py", line 462, in __init__
    self.set_content()
  File "C:\Users\Demo\Desktop\GitHub\demo\app.py", line 505, in set_content
    ).run()
      ^^^^^
  File "c:\Users\Demo\Desktop\GitHub\demo\.venv\Lib\site-packages\streamlit\navigation\page.py", line 293, in run
    self._page()
  File "C:\Users\Demo\Desktop\GitHub\demo\app.py", line 395, in portfolio
    pdf_viewer(
  File "c:\Users\Demo\Desktop\GitHub\demo\.venv\Lib\site-packages\streamlit_pdf_viewer\__init__.py", line 93, in pdf_viewer
    with open(input, 'rb') as fo:
         ^^^^^^^^^^^^^^^^^
OSError: [Errno 22] Invalid argument: 'https://demo.com/0.pdf'

solution

file location: site-packages/streamlit_pdf_viewer/init.py

modify the code

import requests

    if type(input) is not bytes:
        try: # give a try except on original function
            with open(input, 'rb') as fo:
                binary = fo.read()
        except:
            try: # try to get url file with requests
                response = requests.get(input)
                response.raise_for_status()
                binary = response.content
            except: # if still failed raise error message
                raise TypeError(f"input must be a file path, URL, or binary data. {input} is not a valid file path or URL.")
    else:
        binary = input

[lfoppiano]

Thanks! We had an issue opened for this (#33), but we've decided to leave this to the developer because it might require to cache the PDF otherwise at each click the pdf viewer will try to download it.