unstructured-io/.github

.github's Issues

partition_docx - does not pick up 'subject' document property as document text

I am working with partition_docx() and the document I am processing has the 'subject' document property inserted several places throughout the document.

It appears that partition_docx() does not pick up the text of this property, and it is therefore dropped from the output text.

You can easily recreate this:

• create a new MS Word document
• add the subject property via file->info->properties->advanced properties
• add some text to the document and then insert the subject property into the document via insert->quick parts->field
• parse the document via partition_docx()
• the subject property text is omitted from the output of partition_docx()

skip ocr on certain elements

There are some element types (such as images or pictures) that may be required to pass through a custom classifier model to determine if it needs ocr. As such by specifying a argument to not ocr those element types, we can skip ocr of those element types. This feature would probably only be applicable to pdfs and images since docx, pptx, etc are not able to capture element types like image