Copy with bibtex / citation

[anandkumar89]

I often read articles and I want to copy selected text along with bibtex/citation not just [1,2] etc. but append the citation as -

some copied paragraph has [1,2] citations that I've selected and copied

[1] ....
[2] ...

I need help on following -

1. Getting text of pdf (for parsing references): Does sioyek parses and makes available text of pdf. I plan to write a python code that finds and extracts references from the article. I'd save it somewhere in database. This function can be called once for each article, gets stored in a database

2. select a rectangle and then OCR: how can I get screenshot of selected rectangle. Is it possible directly via sioyek or any easy way.

Thanks in advance. Totally new to sioyek, any heads up or suggestions will be helpful.

[rossikos]

One way to get the document text is with the sioyek python package:

from sioyek.sioyek import Sioyek, Document

sioyek = Sioyek(SIOYEK_PATH, LOCAL_DATABASE_FILE, SHARED_DATABASE_FILE)
pdf = Document(pdf_path, sioyek)
for page in range(pdf.doc.page_count):
    ptext = pdf.doc[page].get_text()
...

However, I would only recommend this if you are planning on using some of the other sioyek methods. You can look through them here.

The library gives an amazing level of control, but since sioyek uses the MuPDF engine, you can just do:

import pymupdf

with pymupdf.open(fpath) as doc:
    for page in range(doc.page_count):
        ptext = doc[page].get_text()

As for the OCR, I'm not sure why you need that. If your pdf lacks a text layer, you can run it through something like ocrmypdf.

[anandkumar89]

I need OCR for text with mathematical objects - It'd be useful for me to grab the screenshot of rectangle selected run it through OCR.

(1) helps me to create a list of references.

Thanks a ton for your response. This helps me get started well.

[rossikos]

There is a script mentioned in discussions/1218 that seems to do what you're looking for. You can adapt it for your particular OCR solution.

You can also get screenshots with the _copy_screenshot_to_clipboard command.

[anandkumar89]

Thanks a ton! This does what I need.