[Bug]: UnstructuredReader does not respect encoding and errors in the SimpleDirectoryReader inputs

[Alavi1412]

Bug Description

I am reading some files from my local using this code:

loader = SimpleDirectoryReader(
    local_folder,
    encoding='utf-8',
    file_extractor={
        '.pdf': UnstructuredReader(),
        '.png': UnstructuredReader(),
        '.jpg': UnstructuredReader(),
        '.pptx': UnstructuredReader(),
        '.ppt': UnstructuredReader(),
        '.docx': UnstructuredReader(),
        '.doc': UnstructuredReader(),
        '.xlsx': UnstructuredReader(),
        '.xls': UnstructuredReader(),
        '.txt': UnstructuredReader(),
    },
    required_exts=required_exts,
    recursive=True
)
documents = loader.load_data()
return documents

The problem is, inside the partition_text method of unstructured the encoding is null, and even the errors parameter is not even passed into the place where the file is being opened.

I have attached the debugger from unstructured.partition.text.partition_text which indicates the encoding is none

Version

0.10.42

Steps to Reproduce

Set the encoding in the SimpleDirectoryReader and use UnstructuredReader as file_extractor, and you'll see that the encoding wouldn't even be used.

Relevant Logs/Tracbacks

No response

[logan-markewich]

These should be args on the actual UnstructuredReader class, but currently, it's not supported. Would need a PR to add that