How to split docx / pdf file pages in Laravel 9

I have an online library application that allows users to search the content of books by keywords.

To add a book to my application, I want to be able to upload the docx file of the book, then extract its content page by page, and add it to the database separately.
I couldn't find any specific library in Laravel/PHP to do that for me, so I tried to unzip the Docx file and look for a specific separator to detect pages based on that, but apparently, there is no such thing!

I even tried to read pages from the PDF file, but it returns weird characters that are not readable! I have to add the books are not written in English. They are written in Persian and Arabic which is a right-to-left language (please see bellow screenshot). Here is the link for a real data sample in pdf format: https://www.dropbox.com/s/g939q5oot14ib1w/test-for-stack-over-flow.pdf?dl=0.

Please let me know if you have any idea how can I add each page of the book to the database separately (as a new row in the pages table!), no matter what format I have to use, docx, pdf, txt, etc!

Ok I am told that for word output to PDF then importing to MS Word is likely to give best results, but PHP itself is not an MSOffice User Workstation.

Comments and installation feedback from @Kmaj ,are at the end.

The closest I know is far from perfect but possibly can be bettered. So command line export PDF to plain text will look like this for the first page, where a number of issues with plain text become apparent, the major hurdle is word order can often be wrong at strategic input/output points.

 = symbolic characters like emojis

So using Poppler PDFtotext gives this output for page 1 and the 1 can be in a loop from 1-100+.

pdftotext -nopgbrk -layout -f 1 -l 1 SO-75712922.pdf

‫انتشارات تست‪ /‬شمارۀ ‪53‬‬




‫کتاب تست‬
‫تست این کتاب‪‬‬



‫تست اصحاب‬
‫وصی و فرستادۀ ی‪‬‬

‫(خداوند در زمین تمکینش دهد)‬




‫مترجم‬
‫گروه مترجمان انتشارات تست ‪‬‬

And without -layout it should be more compact, but then Page 2 NEEDS it to be -Layout to not encourage line splits. So documents like this are highly difficult to guess such that output with layout and an alternative may need some automated selector for best of 2 or 3 runs.

Page 1 without Layout looks about 90% usable

Page 2 with Layout has the lines aligned for same height (but reversed order for tabular)

For windows users the 64 bit pre-compiled binaries (currently latest=2023-01) are at https://github.com/oschwartz10612/poppler-windows

From @Kmaj for Mac users

> I first installed poppler on my mac using "brew install poppler". Then, I could successfully run the command you mentioned above in a loop and generate each txt page separately with a dynamic name.

> For those who have the same issue, I ran the following command in my mac terminal:

 for i in {1..3};
 do pdftotext-nopgbrk -layout -f ${i} -l ${i} ~/mypath-to-pdf/pdf-name.pdf ~/mypath-to-output-folder/output-name-${i}.txt;
 done