bug/reopen temp file (pdf high_res)

[FAbrahamDev]

Describe the bug
In line:

unstructured-inference/unstructured_inference/inference/layout.py

Line 362 in d4785df

with tempfile.NamedTemporaryFile() as tmp_file:

is a tmp file created to pass as filename to process_file_with_model -> DocumentLayout.from_file -> load_pdf -> extract_pages (pdf_miner).
The extract_pages tries to read the file again with open_filename(pdf_file, "rb") as fp:.

Which results in a PermissionError: [Errno 13] Permission denied: 'C:\\Users\\...\\AppData\\Local\\Temp\\tmpf9flca30' under windows.

Same error here:
https://github.com/Unstructured-IO/unstructured/blob/d3a404cfb541dae8e16956f096bac99fc05c985b/unstructured/partition/pdf_image/ocr.py#L79

To Reproduce

import tempfile

# print operating system name
import os
print(os.name)


# Create a temporary file
with tempfile.NamedTemporaryFile() as tmp_file:
    # Write some data to the file
    tmp_file.write(b'Hello, world!')
    tmp_file.flush()  # Flush the buffer to make sure data is written

    # Get the name of the file
    file_name = tmp_file.name

    # Since the file is closed after the with block, we need to open it again for reading
    with open(file_name, 'r') as file:
        # Read the data from the file
        content = file.read()
        print("Content of the temp file:", content)

Expected behavior
I expect it not to crash :)

Additional context
Possible solution taken from here: https://stackoverflow.com/questions/39983886/python-writing-and-reading-from-a-temporary-file

def process_data_with_model(
    data: BinaryIO,
    model_name: Optional[str],
    **kwargs,
) -> DocumentLayout:
    """Processes pdf file in the form of a file handler (supporting a read method) into a
    DocumentLayout by using a model identified by model_name."""

    with tempfile.TemporaryDirectory() as td:
        f_name = os.path.join(td, "tmp_file")
        with open(f_name, "w") as tmp_file:
            tmp_file.write(data.read())
            tmp_file.flush()

        layout = process_file_with_model(
            f_name,
            model_name,
            **kwargs,
        )

    return layout

or another solution by gpt:

import tempfile

with tempfile.NamedTemporaryFile(delete=False) as tmp_file:
    tmp_file.write(b'Hello, world!')
    # Get the name of the file before closing
    file_name = tmp_file.name

# Now the file is closed, you can open it again
with open(file_name, 'r') as file:
    content = file.read()
    print("Content of the temp file:", content)

# Optionally, delete the file if you don't need it anymore
import os
os.remove(file_name)

Not sure which is better.
The latter one probably requires a try catch final with the removal and then reraise the error.