[Editor Plugin] EpubCheck

[Doitsu]

This plugin allows you to check books with EPUBCheck 5.x Make sure that you have Java installed on your machine before installing and running this plugin.

To test whether Java is installed on your machine open a command prompt/terminal window and enter the following command:

java -version

If you don't get a version number, Java is either not installed or the Java binary can't be found. If Java is not installed, go to the Oracle website to download the latest Java version. (Note that you'll need to re-boot your machine after the installation.)

Credits: Kovid Goyal who helped me with my Qt5 and Calibre API questions and gave me useful pointers. jackie_w helped me with the subprocess code and thiago.eec provided very helpful pointers.

Disclaimer: Since I'm not a programmer you might want to save your books before having them checked by this plugin. (The plugin will also automatically create a checkpoint.)

Troubleshooting: If you get a WindowsError: [Error 2] unhandled exception message, the Java binary couldn't be found. Installing/updating Java and re-booting your machine should take care of this problem.

Settings: If you're an advanced user, you can change several settings by creating an EpubCheck.json file with one or more of the following entries in the Calibre plugins folder (Preferences > Miscellaneous > Open calibre configuraton directory > plugins):

{
  "locale" : "xx",
  "usage" : true,
  "java_path": "path/to/java/or/java.exe,"
  "is32bit": false, 
  "github": true, 
  "check_interval": 7 
}

1. If you want to change the language of EPUBCheck messages add a locale entry with one of the following language codes: en, de, es, fr, it, ja, nl, pt-BR, ko-KR.
2. If you want EPUBCheck to also display USAGE messages, add a usage entry.
3. If you want to use a custom Java path, add a java_path entry.
4. The plugin should automatically detect 32bit JVMs. If it doesn't, change the value of the is32bit entry to true.
5. By default, the plugin will automatically check the EPUBCheck Github website for updates. If you'd rather manually update the EPUBCheck Java files bundled with this plugin, change the value of the github entry to false.
6. The value of the check_interval entry controls how often the plugin will contact the Github EPUBCheck website for updates. The default value is 7 days = once a week.
   

Please note that if you specify multiple values, each entry (except for the last one) must be terminated by a comma. (If you only add one entry, don't terminate it with a comma.)

If you still can't get the plugin to work, download Pagina Epub-Checker (freeware) and use it with the Calibre Open With plugin.

Version History

Spoiler:

Version 0.2.5 - 15 July 2023
Updated command line parameters for EPUBCheck 5.1.0
Version 0.2.4 - 2 Jan 2022
PyQt6 support added
Version 0.2.3 - 10 Jul 2020
EPUBCheck 4.2.4 bundled; better Python 3 support
Version 0.2.2 - 27 Dec 2019
Calibre dark mode support added
Version 0.2.1 - 9 Sep 2019
Minor bug fixes and enhancements
Version 0.2.0 - 9 Sep 2019
Minor bug fixes and enhancements
Version 0.1.9 - 19 Jul 2019
Updated EPUBCheck v4.2.1 to EPUBCheck v4.2.2.; added auto-update feature
Version 0.1.8 - 23 Apr 2019
Updated EPUBCheck v4.2.0 to EPUBCheck v4.2.1.
Version 0.1.7 - 23 Apr 2019
Updated EPUBCheck v4.1.0 to EPUBCheck v4.2.0.
Version 0.1.6 - 23 Jan 2019
Updated EPUBCheck v4.1.0 to EPUBCheck v4.1.1 & minor bug fixes.
Version 0.1.5 - 23 Dec 2018
Fixed INFO message bug and added support for usage and java_path parameters
Version 0.1.4 - 26 Nov 2018
Updated EpubCheck v4.0.2 to EPUBCheck v4.1.0.
Version 0.1.3 - 5 Nov 2018
The plugin will no longer display the Windows CMD window.
Version 0.1.2 - 22 Jun 2017
Added support for Windows temp folders on drives other than C:.
Version 0.1.1 - 25 Dec 2016
Disabled option to automatically close the Check Book tab.
Version 0.1.0 - 25 Dec 2016
Initial release.

[jackie_w]

Hi Doitsu,

Thank you for your work

Some feedback. I didn't have any problem installing it but when I run it in the Editor it always removes my docked panel (bottom left on my setup) for the standard calibre Check Book. It's easy enough to go back into View to recheck 'Check Book' but I'd rather not have to do that if possible. Running Check Book also puts it back in place.

Also EpubCheck automatically docks itself along the whole of the top of the Editor window. It allows itself to be docked elsewhere (I put it alongside 'Check Book') but when it's run it always moves itself back along the top. Is this deliberate?

[Doitsu]

Hi jackie_w,

All the issues that you mentioned are actually by design, because I first run Check Book and then EpubCheck, but this can be easily disabled. I attached a new version, which will no longer close the Check Book window, to my first post.

The original behavior can be restored by adding the following entry to EpubCheck.json:

{
  "close_cb": true
}

D.

[BetterRed]

@Doitsu - thanks for the Christmas surprise.

I adjusted your thread title to make it consistent with similar plugins.

The Notes at the top of the Plugins Index thread describe what you need to do to get your PI included in the Index. Including it in the Index will ensure users get notified when you release a new version.

BR

[Doitsu]

Originally Posted by BetterRed

I adjusted your thread title to make it consistent with similar plugins.

Thanks!

Originally Posted by BetterRed

The Notes at the top of the Plugins Index thread describe what you need to do to get your PI included in the Index. Including it in the Index will ensure users get notified when you release a new version.

Here's the required information:

[*][Editor Plugin] EpubCheck 
Simple EpubCheck wrapper
Version: 0.1.1; Released: 12-25-2016; Calibre: 2.70.0; Author: Doitsu; History: Yes; 
Platforms: Windows, OSX, Linux

[BetterRed]

@Doitsu - it should get copied to here Index of calibre plugins within the next hour or so, and then it will show up in Get New plugins

BR

[jackie_w]

Originally Posted by Doitsu

Hi jackie_w,

All the issues that you mentioned are actually by design, because I first run Check Book and then EpubCheck, but this can be easily disabled. I attached a new version, which will no longer close the Check Book window, to my first post.

The original behavior can be restored by adding the following entry to EpubCheck.json:

{
  "close_cb": true
}

Thanks for the extra option

[jackie_w]

Hi Doitsu,

I'd like to follow-up on a comment you made here, namely:

The Sigil version will also try to move the cursor to the column position reported by epubcheck. However, the column position that epubcheck reports often points to the closing tag.

Is there any prospect of the calibre Editor plugin auto-moving to the file/line error when clicked?

At the moment if I click on an item in the EpubCheck error pane all I ever see is as shown in the attached image.

Manually going to the location reported in the epubcheck error message is normally enough for me to figure out what epubcheck doesn't like but an automated go-to option would be nice .

[Doitsu]

@jackie_w: There appears to be a bug in the plugin, because the error message should only be displayed, if epubcheck didn't report a file name.
Also file names should be displayed without path information. For an example, have a look at the attached screenshot, which was generated by running the Linux version of the Calibre epubcheck plugin.
Most likely the structure of your epub is different from the structure of my test file or it's an OS-specific bug. Can you please test my test file on your Linux/macOS(?) machine?
If it works as designed, please PM me a link to the file that caused the bug or another Public Domain file that causes the same bug.
Alternatively, saving the epub in Calibre Editor before running the epubcheck plugin might also take care of this problem.

[jackie_w]

Thanks for a speedy reply.

I'm a Windows 10 user. I tried your test broken.epub. I see exactly the same problem as before - see attached image. Perhaps it is a Windows-only problem or possibly something specific in my Windows setup, but I'm not aware of having done anything 'unusual' to my Windows settings.

[jackie_w]

In case it helps I've just run epubcheck.jar 'raw' from a Windows CMD box. This was the output:

java -jar epubcheck.jar broken.epub
Validating using EPUB version 2.0.1 rules.
FATAL(RSC-016): broken.epub/OEBPS/Text/Section0001.xhtml(10,29): Fatal Error while parsing file 'The element type "b" must be terminated by the matching end-tag "</b>".'.
ERROR(RSC-005): broken.epub/OEBPS/Text/Section0001.xhtml(-1,-1): Error while parsing file 'The element type "b" must be terminated by the matching end-tag "</b>".'.

Check finished with errors

epubcheck completed

[Doitsu]

Unfortunately, I wasn't able to reproduce this problem on my 64bit Windows 10 machine. Can you please check your Java version?

I got the following results:

java -version
java version "1.8.0_91"
Java(TM) SE Runtime Environment (build 1.8.0_91-b15)
Java HotSpot(TM) Client VM (build 25.91-b15, mixed mode, sharing)

Are you using a Windows 10 VMWare/VirtualBox image by any chance or did you install Calibre on a drive other than C:?
If you installed Calibre on a drive other than C: could you please temporarily install Calibre on your C: drive to test whether this fixes the bug?
If it does, I'll have to update the plugin code to search for drives other than C:.

[jackie_w]

Originally Posted by Doitsu

Unfortunately, I wasn't able to reproduce this problem on my 64bit Windows 10 machine. Can you please check your Java version?

I got the following results:

java -version
java version "1.8.0_91"
Java(TM) SE Runtime Environment (build 1.8.0_91-b15)
Java HotSpot(TM) Client VM (build 25.91-b15, mixed mode, sharing)

Here's my Java version output:

java -version
java version "1.8.0_131"
Java(TM) SE Runtime Environment (build 1.8.0_131-b11)
Java HotSpot(TM) 64-Bit Server VM (build 25.131-b11, mixed mode)

Originally Posted by Doitsu

Are you using a Windows 10 VMWare/VirtualBox image by any chance or did you install Calibre on a drive other than C:?

Calbre is installed in the default location, C:\Program Files\Calibre2
No, I'm a plain old Windows user. I'm not using a Windows 10 VMWare/VirtualBox. In fact I don't think I'd know how to do this.

P.S. I know very little about Java other than to run an update when I get a notification.

[Doitsu]

I looks like my Java version wasn't up-to-date, but even after installing the latest Java version, I couldn't reproduce your bug.

Originally Posted by jackie_w

Calbre is installed in the default location, C:\Program Files\Calibre2

I'm assuming that you installed the 64bit Calibre version and that you're using a 64bit Windows 10 machine. If not this might be 32bit Calibre/Windows issue.

Originally Posted by jackie_w

P.S. I know very little about Java other than to run an update when I get a notification.

I don't know anything about Java either.

I'm afraid I can't help you with your problem, since I can't reproduce it. However, since you appear to be a much better Python programmer than me and the plugin code is really very simple, you might be able to locate the problem yourself.

Try to comment out the following sections in main.py:

1. lines 154-157

                    if iswindows:
                        drive_letter = os.getenv("SystemDrive")    
                        line = re.sub(drive_letter, 'C^',  line)
                        line = re.sub(drive_letter.lower(), 'C^',  line)

and lines 191 and 192

                    if iswindows:
                        filename = os.path.basename(re.sub('C^', drive_letter, filename))

There's a very slim chance that these changes might fix the problem, but I'm not overly optimistic about it.

It that doesn't work, uncomment the above sections and insert the following line after line #219

                    error_messages.append((filepath, linenumber, err_code, message))
                    QMessageBox.information(self.gui, "Debug", "filepath: " + str(filepath) + '\n' + "linenumber: " + str(linenumber) + '\n' + "err_code: " + str(err_code) + '\n' + "message: " + str(message))

Then report the output.

BTW, I got:

filepath: OEBPS/Text/Section0001.xhtml
linenumber: 10
err_code: FATAL(RSC-016)
message: Section0001.xhtml Line: 10 Col: 29 FATAL(RSC-016): Fatal Error while parsing file 'The element type "b" must be terminated by the matching end-tag "</b>".'.

and

filepath: OEBPS/Text/Section0001.xhtml
linenumber: None
err_code: ERROR(RSC-005)
message: Section0001.xhtml ERROR(RSC-005): Error while parsing file 'The element type "b" must be terminated by the matching end-tag "</b>".'.

[jackie_w]

Originally Posted by Doitsu

I'm assuming that you installed the 64bit Calibre version and that you're using a 64bit Windows 10 machine. If not this might be 32bit Calibre/Windows issue.

Yes it's the 64bit Calibre version.

Originally Posted by Doitsu

... since you appear to be a much better Python programmer than me ...

Appearances can be deceptive

Originally Posted by Doitsu

Try to comment out the following sections in main.py

Commenting out those sections didn't help. The contents of the extra QMessageBoxs for the 2 errors are attached below.

My output text does seem to be very different from yours. No wonder there are problems.

We could use some feedback from another Windows user to see if it's Windows in general or something about my PC. Either way, don't worry as the plugin as-is lets me know what needs fixing.