Copy notes from EPUB to computer?

[Shohreh]

Hellj

Is there a simple tool (for Windows), that could grab the notes from an EPUB file once it's connected through USB, and append them to a local TXT file?

I found that, after adding a note, each .epub file has a corresponding file in "\Digital Editions\Annotations\blah.epub.annot", but that XML file only contains links to annotions, not the annotations themselves:

<annotation>
	<dc:identifier>urn:uuid:de120535-d147-4f21-bbb6-3da9ba9dc7a9</dc:identifier>
	<dc:date>2026-01-11T02:58:19Z</dc:date>
	<dc:creator>urn:uuid:374c8101-17fe-41c5-b9c7-0f4d8a6833f5</dc:creator>
	<target>
		<fragment start="OEBPS/ch01.html#point(/1/2/1/18/9/1:260)" end="OEBPS/ch01.html#point(/1/2/1/18/9/1:260)" progress="0.0922619" color="0"/>
	</target>
</annotation>

Also, under \.kobo, the following SQLite files can be found:

• BookReader.sqlite
• fonts.sqlite
• KoboReader.sqlite

I currently use the medieval solution of… shooting screenshots with my smartphone and copy/pasting the PNGs to my computer before possibly grabbing the actual text from the EPUB after searching the file through CTRL+F and pasting it to a TXT if I need to :-/

Thank you.

[JSWolf]

Using calibre, you give ths Annotations and/or Kobo Utilities plugins a try to get your annotations.

Annotations - https://www.mobileread.com/forums/sh...d.php?t=241206
Kobo Utilities - https://www.mobileread.com/forums/sh...d.php?t=366110

[Shohreh]

Thanks. Can it be used through the CLI, so I can call it through a batch file instead of loading the GUI?

[JSWolf]

Originally Posted by Shohreh

Thanks. Can it be used through the CLI, so I can call it through a batch file instead of loading the GUI?

No idea. I don't use calibre via the CLI.

[Shohreh]

I'll see if it supports it, thx

[Shohreh]

It looks like all the highlights ("annotations") are appended to XML files under \Digital Editions\Annotations\, so there's no need to use an external app like Calibre or KOreader — a simple Python script with BeautifulSoup will take care of business:

\Digital Editions\Annotations\myfile.epub.annot:

<annotation>
	<dc:identifier>urn:uuid:4fe72a8f-81d1-4b15-a5ea-7552f68295eb</dc:identifier>
	<dc:date>2026-01-14T22:36:13Z</dc:date>
	<dc:creator>urn:uuid:374c8101-17fe-41c5-b9c7-0f4d8a6833f5</dc:creator>
	<target>
		<fragment start="text/part0015.html#point(/1/4/65/4:1)" end="text/part0015.html#point(/1/4/66:0)" progress="0.122371" color="4">
			<text>stuff I highlited</text>
		</fragment>
	</target>
</annotation>

--
Edit: First draft

import wx
from bs4 import BeautifulSoup
import shortuuid
import pathlib
import sys

DEFDIR = 'd:/Digital Editions/Annotations/'

# Zack Gainsforth 2014
def FileDialog(DefaultDir=DEFDIR, Title='Select File:', Extension='*.annot'):
	dlg = wx.FileDialog (None, Title, DefaultDir, '', Extension, wx.FD_DEFAULT_STYLE | wx.FD_FILE_MUST_EXIST)
	if dlg.ShowModal() == wx.ID_OK:
		FileName = dlg.GetPath()
	else:
		FileName = None
	dlg.Destroy()

	return FileName

if 'app' not in locals():
	app = wx.App(None)

INPUT = FileDialog()
BASENAME = pathlib.Path(INPUT).stem
uuid = shortuuid.ShortUUID().random(length=8) 
OUTPUTDIR = sys.path[0]
OUTPUTFILE = f"{OUTPUTDIR}/{BASENAME}.{uuid}.annot"

soup = BeautifulSoup(open(INPUT, 'r',encoding='utf-8'), features='xml')
title = soup.find("dc:title").text if soup.find("dc:title") else "No title"
print(title)
if soup.find("annotation"):
	for annot in soup.find_all("annotation"):
		#<fragment start="text/part0015.html#point(/1/4/65/4:1)" end="text/part0015.html#point(/1/4/66:0)" progress="0.122371" color="4">
		#assume highlight located in single file within EPUB
		location= annot.target.fragment["start"].split(".html")[0]
		print(location)
		text = annot.target.text.strip() + "\n========"
		print(text)

		with open(f"{OUTPUTFILE}","a") as f:
			f.write(f"{location}\n")
			f.write(f"{text}\n")
else:
	print("No highlight")

--
Edit: Oddly, the highlights are now gone from the *.annot files under \Digital Editions\Annotations\

And when running "select count(*) from Bookmark;" from the \.kobo\KoboReader.sqlite database, SQLite ends with "Runtime error: database disk image is malformed (11)".

They can be found after recovering the DB through "sqlite3 KoboReader.sqlite .recover >data.sql ; sqlite3 fixed.sqlite <data.sql", but 1) I shouldn't need to do this, and 2) the lost_and_found table contains 101 columns from c0 to c100 :-/

select rootpgno,trim(c9) from lost_and_found where c1 LIKE '%My Author%' AND c9 IS NOT NULL AND TRIM(c9,' ') != '';

--
Edit: Per this post, I rebooted before copying the DB file to the computer… and the SELECT worked to list highlights:

sqlite3 KoboReader.sqlite
select VolumeID, Text from Bookmark WHERE Text IS NOT NULL AND Text != '';

In "\Digital Editions\Annotations", I notice that they are available for one book but not another, which used to work. It looks like Kobo doesn't export highlights reliably between the DB and those *.epub.annot files, so it's better to read from the DB directly.

To check the integrity of a SQLite DB:

c:\temp\sqlite3 KoboReader.sqlite
PRAGMA integrity_check;

--
Edit: There's a solution that 1) doesn't require installing anything, and 2) exports annotations into text files, so you can then remove the book and they'll still be available — which isn't true of the KoboReader.sqlite file: I noticed that once a book is removed from the Kobo, annotations will also be gone.

1. Connect the Kobo the computer with a (reliable) USB cable
2. Open "\.kobo\Kobo\Kobo eReader.conf" in a text editor, eg. Notepad
3. Search for the following section, add it if not found:

[FeatureSettings]
ExportHighlights=true

4. Save file, disconnect Kobo, and reboot
5. Highlight a section in a book, close it, long-press the book in the list, choose "Export annotations": They will exported as a .txt file
6. Re-connect the Kobo to the computer: Annotations are located in \Exported Annotations\

Sources:
https://www.reddit.com/r/kobo/commen...mment/lk6k4ro/
https://wiki.mobileread.com/wiki/Kob...ration_Options

--
Edit: One book with more hightlights was semi-SQLite garbage (see attachment). I had to reboot the Kobo and re-export the annotations for the text file to be kosher. I'll give KoReader a try