[Plugin] ODTImport

[Doitsu]

Originally Posted by Leonatus

I have been using the plugin very frequently without issue. But this time, I receive this error message: [...]

Unfortunately, I wasn't able to reproduce this problem on my Windows 11 machine. The file selection dialog didn't have problems with spaces or umlauts in the path or the file name. Maybe the file path contained some invisible control characters.

[Leonatus]

As well as in the folder, there had been a "ú" in the name of the file itself ("Júlia"). When the plugin didn't work for the first time, I changed the name of both ("Julia"), suspecting that this letter might have caused the problem. But the issue continued. So, I tried to choose a totally different name and save it to a different location. That solved the problem.

[Doitsu]

Originally Posted by philja

On reading the latest "user-manual-Writer2xhtml17.odt" with special reference to anything mentioning lists, I saw the option "list_formatting" with three possible values - ignore_all, convert_labels, convert_all. This is at odds with line 20 in the config.xml supplied with the plugin 0.4.1 which quotes "value=css1"

In writer2xhtml 1.7.x, list_formatting has 4 possible values:

convert_all (default) old value hard_labels
convert_label_styles old value css1
convert_labels old value css1_hack
ignore_all

I found a sample config.xml file on the writer2xhtml Github page:
https://github.com/writer2latex/writ...cleanxhtml.xml

and slightly updated it for epub3 output and removed all redundant entries where the value matched the default value.
If you change the value in the following line from:

  <option name="list_formatting" value="ignore_all" />		<!-- default: convert_all -->

to

  <option name="list_formatting" value="convert_label_styles" />		<!-- default: convert_all -->

You should get the same results as in older versions.

I also found out by chance that writer2xhtml 1.7.x supports page numbers in epub3 files. Unfortunately, it'll add a masterpage attribute to page break definitions, which EPUBCheck doesn't like.

I commented out one source code line to change this behavior and created a new experimental build that is attached to this post.

If you generate epub2 books, you'll need to change the following config.xml entry from:

to:

  <option name="original_page_numbers" value="true" />		<!-- default: false -->

to:

<option name="original_page_numbers" value="false" />		<!-- default: false -->

If you're not interested in generating page numbers in epub3 books, you can ignore this build.

[roger64]

Today,when trying to create a new ePub, ODT Import failed to launch.
OS:Archlinux.
java version "1.8.0_431"
Java(TM) SE Runtime Environment (build 1.8.0_431-b10)
Java HotSpot(TM) 64-Bit Server VM (build 25.431-b10, mixed mode)


I have the latest version of Sigil (2.3.1.). Please find below the relevant error message.

[KevinH]

Your Arch Python is too new. It has dropped support for the imghdr.py module. This was a really stupid decision by the maintainers of python as the module is small but heavily used and should not have been dropped in any point release no matter if deprecated or not earlier.

Luckily, you can extract the pure python imghdr.py file from the previous version of python and add it to the plugin zip archive and all should work.

You can grab it from here:

https://github.com/python/cpython/bl.../Lib/imghdr.py

Since python breaking plugins is not in the best interest for any platform I may add it to the Sigil python plugin standard library if its license is at all compatible.

[roger64]

Thanks for your informed reply. I'll try it and report asap.

[roger64]

ODT Import is now working as usual.

I added imghdr.py precisely here (see below)

For info, my python version is 3.13.1-1 and has been -massively- updated on the 22nd of December 2024.

Thanks again!

[Sean Fenian]

Doitsu,
THANK YOU THANK YOU THANK YOU for Writer2xhtml. I've spent uncounted hours over the past eighteen months trying to develop a better EPUB3 publishing workflow after trying to post-fix EPUBs exported directly from LibreOffice and seeing first-hand what a godawful flaming dumpster fire libepubgen spews out.

Writer2xhtml gets me closer, in one step, to a publication-ready ebook than anything else I have yet tried. You WILL be getting an acknowledgment in my next published book.

[Doitsu]

Originally Posted by Sean Fenian

Doitsu,
THANK YOU THANK YOU THANK YOU for Writer2xhtml.

I'm glad that you like the plugin, but I'm not the author of Writer2xhtml!
You should thank the developer, Henrik Just, instead. He deserves all the credit.

(The plugin is only a very simple wrapper.)

The author also has a Github page.

[Leilu]

Hi,

Status: failed

Traceback (most recent call last):
  File "C:\Program Files\Sigil\plugin_launchers\python\launcher.py", line 142, in launch
    self.exitcode = target_script.run(container)
                    ~~~~~~~~~~~~~~~~~^^^^^^^^^^^
  File "C:\Users\leilu\AppData\Local\sigil-ebook\sigil\plugins\ODTImportQt\plugin.py", line 107, in run
    sync_prefs_with_plugin_folder(plugin_folder, prefs_folder)
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\leilu\AppData\Local\sigil-ebook\sigil\plugins\ODTImportQt\plugin.py", line 35, in sync_prefs_with_plugin_folder
    for name in os.listdir(pref_dir):
                ~~~~~~~~~~^^^^^^^^^^
FileNotFoundError: [WinError 3] The system cannot find the path specified: 'C:\\Users\\leilu\\AppData\\Local\\sigil-ebook\\sigil\\plugins_prefs\\ODTImportQt'
Error: [WinError 3] The system cannot find the path specified: 'C:\\Users\\leilu\\AppData\\Local\\sigil-ebook\\sigil\\plugins_prefs\\ODTImportQt'

Non QT works fine.

Where to add the

"output_format": "epub3",

line? Thank you

[Doitsu]

I must have introduced a new bug in the latest version. Will release an updated version tonight or tomorrow.

Where to add the [...] line?

In ODTImportQt.json

[Doitsu]

Originally Posted by Leilu

[CODE]Status: failed [...]

I attached a new version to the first post.

Originally Posted by Leilu

Where to add the

"output_format": "epub3",

line?

If you create a blank epub3 book in Sigil before running the plugin for the first time, it should automatically add this preference to ODTImportQt.json.

[Leilu]

Nice. Thank you.

*

My ODT file is quite messy because it was exported from ABBYY FineReader OCR. For example, I have about a dozen different character styles that are actually identical only for italicized text. When imported into Sigil, they appear as <span class="CharStyle78">some italic text</span> in some places and <span class="CharStyle32">some other italic text</span> in others.

I want to avoid importing these character styles this way and rely solely on paragraph styles, except for italics and bold, which should appear in Sigil as <em> and <strong> tags (similar to how Pandoc handles conversions). I would prefer to stick with the ODTImport plugin because, as mentioned in another thread, I need to preserve the physical book's page references.

[Doitsu]

Originally Posted by Leilu

I want to avoid importing these character styles this way and rely solely on paragraph styles, except for italics and bold, which should appear in Sigil as <em> and <strong>[...]

If ABBY Fine Reader assigned the same character styles throughout the book, you might be able to map them to tags in the config.xml file in the plugin folder.

For numerical character styles such as CharStyle78, adding the following line to config.xml should theoretically output them as <em>..</em> tags.

<xhtml-style-map css="(none)" element="em" family="text" name="CharStyle78" />

If they're character overrides the following line might work:

<xhtml-style-map css="(none)" element="em" family="attribute" name="CharStyle78" />

For more information, see the official writer2xhtml user guide.

If there are simply too many of these numerical styles, you could also replace them in bulk in Sigil with Saved Search groups or DiapDealer's excellent TagMechanic Plugin

BTW, you might find my Incremental IDs Plugin useful for manipulating page number placeholders. You could now insert generic placeholder <span class="page"></span> tags and add sequential ids later.

[philja]

I've just installed v0.5.2 to replace v0.5.0 and took care to make sure my customised config.xml and epub.css files were in the plugins preferences directory.

I thought I understood from your red print notice in the first post that the updated plugin would copy the customised files from preferences into the plugins directory.

If you've previously used this plugin and customized config.xml and/or epub.css make sure to create a backup of the plugin folder and also copy these files to the plugin preferences folder. (The latest version of the plugin will always copy these files from the plugin preferences folder to the plugin folder.)

Neither on linux nor windows has the 0.5.2 plugin copied the preferences versions to the plugin folder.

Edit: my mistake

I just ran a trial import, and the plugin copied the files as promised. I hadn't understood that the copy would occur on first use rather than on installation.