Issue with Arabic Text Extraction Using PDF-Extract API

[abhi_0910]

I am using the PDF-Extract API to extract text from a PDF document that contains both English and Arabic text. The PDF is scannable (non-image, non-OCR). However, I have noticed an issue with how the Arabic script is extracted.

In the original PDF, the Arabic names appear correctly within a paragraph as:

محمد مصطفى سالم منصور
منصور ال محمد مصطفى سالم
محمد مصطفى سالم

However, the Output is 

ﺍ
ﻝ
ﺭ
ﻣﺤﻤﺪ ﻣﺼﻄﻔﻰ ﺳﺎﻟﻢ ﻣﻨﺼﻮ
ﺭ
ﺍ
ﻝ
ﻣﻨﺼﻮﻣﺤﻤﺪ ﻣﺼﻄﻔﻰ ﺳﺎﻟﻢ
ﻣﺤﻤﺪ ﻣﺼﻄﻔﻰ ﺳﺎﻟﻢ

Additionally, the extracted Arabic text moves to the beginning of the paragraph instead of retaining its original position in the document.

I would like to understand:

1. What is the cause of this issue?
2. Is there a fix or workaround available to correctly extract non-Latin text while maintaining its formatting and position?

[Joel Geraci]

I'm not sure how you are getting anything at all from the Arabic text.  Those areas should be identified as "Figures". The API is currently optimized for English language content.