Sometimes it may be tedious to type all the matter from a non-searchable files. IndianScriptures is a great site to get any scriptures in both English and Sanskrit. Puranas are also available on WestBengal Dspace but only in English and that too in non-text searchable format. I found two methods to post answers from non-searchable documents.
1. Google Drive OCR (accurate and the best method):
In this method, upload the PDF (size must be below 10MB) to drive. Right Click on file - Open with Google docs.
In next tab, document will be opened in text searchable format. We can scan Devanagari with this method and accuracy is very good. No other OCR tools can scan Devanagari, even the paid software. Limitations of this method is, it will only allow upto 10MB file and if you have chunk of big file you need to split into a smaller file by using Splitting tools. PDFsam Basic can do this work for windows and you can use other tools for other platforms. Also, Google Drive OCR won't scan image files and you need to convert the image to PDF using ImagetoPdf converter tools.
2. Microsoft Onenote OCR :
This method can be used to scan Puranas from dspace. Accuracy of recognition is good but not great. This method cannot recognize Devanagari. To do this, take screenshot of text you want to scan using snipping tool (Windows)- copy and paste the image in Microsoft OneNote, which is part of Microsoft office. Right Click and use the option "copy text from Picture".
There are many free software to this. See these questions. Free OCR software that makes a PDF searchable (with searchable text at the right place) and Which tools can turn scanned paper documents into searchable text pdf on Mac?.
I personally use Google Drive OCR to answer questions from Shiva Purana (Indian Scriptures) which is not searchable. To copy small passages, i use OneNote OCR, though italicized must be edited. But you face no problems in Google Drive OCR.