[Plugin] DOCXImport

[BeckyEbook]

@DiapDealer:
I thought you could add extra error handling, but it turns out that in qtdialogs.py you have to check for smap and css files first, and if they are not there (and the checkboxes are checked) the whole dialogbox returns default values, so the DOCX file also is "None".

Since the existence of smap and css files is not crucial to the operation of the plugin, you can probably just add messages directly in qtdialogs.py instead of "return".
The user will be better informed and the plugin will still work.

You don't need to change anything at tkdialogs.py.

Spoiler:

        if self.checkbox_smap.isChecked():
            if len(self.cust_smap_path.text()):
                self.prefs['useSmapPath'] = self.cust_smap_path.text()
                _DETAILS['smap'] = (self.checkbox_smap.isChecked(), self.cust_smap_path.text())
            else:
                # Message box that no file is selected
                # return
                print("Not selected smap file!")
        self.prefs['useCss'] = self.checkbox_css.isChecked()
        if self.checkbox_css.isChecked():
            if len(self.cust_css_path.text()):
                self.prefs['useCssPath'] = self.cust_css_path.text()
                _DETAILS['css'] = (self.checkbox_css.isChecked(), self.cust_css_path.text())
            else:
                # Message box that no file is selected
                # return
                print("Not selected css file!")

Perhaps you will find a better solution.

[DiapDealer]

I'll look into it. Thanks!

[kawaberitan]

I really enjoy using docximport, but I have a small issue. I have to manually adjust the indentations and spacing of each quoted sentence or paragraph because it is recognized as an ordinary body paragraph. How can I prevent this from happening?

[DiapDealer]

Originally Posted by kawaberitan

How can I prevent this from happening?

By creating a custom style for the quoted paragraphs in Word which defines them as unordinary, and then mapping that style to the desired css in the style map the plugin uses. There are examples of style mapping in the samples.

[Quoth]

No auto-generation of CSS from styles then? I have to predefine CSS for each style?

[DiapDealer]

Originally Posted by Quoth

No auto-generation of CSS from styles then?

Nope.

Originally Posted by Quoth

I have to predefine CSS for each style?

Yes. As it was always intended. I've always envisioned this plugin as an authorial tool that could be tailored to a user's particular--and hopefully consistent--docx style practices. Not a conversion tool that would automatically generate css from a myriad of users' varied docx style practices.

[kawaberitan]

Originally Posted by DiapDealer

By creating a custom style for the quoted paragraphs in Word which defines them as unordinary, and then mapping that style to the desired css in the style map the plugin uses. There are examples of style mapping in the samples.

thanks for the answer. I had a hard time with multi-quote books. I solve it by marking the tags that docximport supports and importing them back into the blockquote tag.

[DiapDealer]

Glad you got it working, but you're probably better off using div tags with css styling rather than blockquote tags. Naked text inside blockquote tags will fail EPUBCheck validation, and it might prove difficult to map your docx style to something like <blockquote><p>blah</p></blockquote>.

Not sure what you mean by "tags docximport supports," though. The plugin should be fairly tag agnostic. It has been in my experience.

[phillipgessert]

You can nest elements in those mammoth map files. Something like p[style-name='Quote'] => blockquote > p:fresh.

[Interrobang]

@DiapDealer:

First off I wanted to thank you for all the work you’ve put into docximport. I’ve had it on my To Be Tried pile for a while and have finally started to use it. As a result, I’ve come across a few things.

When importing a .docx file, it creates a new epub with the .xhtml and .css files at the root of the file structure and not under the OEBPS folder. I thought the epub spec required the files to be nested under that folder. Correct?

Mammoth supports the command include_default_style_map=False to disable default style mappings. Is there a way to pass that command with docximport? If not, could it be added?

Finally, with regards to split markers, what elements can be defined in the style map to be mapped to it besides the horizontal line? It would be great if one could use the Word page break character.

Thanks again for your hard work on the plugin.

Tim

[DNSB]

Originally Posted by Interrobang

When importing a .docx file, it creates a new epub with the .xhtml and .css files at the root of the file structure and not under the OEBPS folder. I thought the epub spec required the files to be nested under that folder. Correct?

Nope. You must have an uncompressed mimetype file in the root of the .zip container. You must have a META-INF directory which must contain a container.xml file. The META-INF directory can contain other files such as encryption.xml, manifest.xml, metadata.xml, rights.xml and signatures.xml though other than encryption.xml, you will seldom see the other files in the wild. None of those files will be listed in the manifest as they strictly speaking are not part of the Epub document nor can they be encrypted.

Everything else can be placed in the root of the .zip file though using an OEBPS directory containing Text, Images, Styles, Misc and the content.opf and toc.ncx is pretty common and what Sigil uses when you click on Restructure Epub to Sigil Norm. Early versions of Sigil forced that structure.

If you are suffering from insomnia, you can check the Open Container Format for more information.

[Interrobang]

Originally Posted by DNSB

If you are suffering from insomnia, you can check the Open Container Format for more information.

Mmmm, midnight fuel. Thanks for the link. I did know that some files and the META-INF folder are at the root.

My comment was based on a simple Google search, in which every answer appeared to confirm the fact that other content related files and folders are stored under /OEBPS. Even the Wiki page on EPUB states:

Apart from mimetype and META-INF/container.xml, the other files (OPF, NCX, XHTML, CSS and images files) are traditionally put in a directory named OEBPS.

So I interpreted that as a requirement, but it appears it's just strongly suggested. Do you concur?

Thanks again,
Tim

[DiapDealer]

I ran into too much trouble trying to accommodate a structure with subdirectories, so I took the easy way out of creating a flat structure on import. I did this, knowing that it's trivial in Sigil to Restructure to Sigil Norm. It can even be included in a Toolchain after running the plugin. But really; any reading system or rendering engine that can't deal with a flat epub structure should be considered broken.

As for include_default_style_map=False, there is no current way, but I'll try to look into it.

I don't remember offhand how the split markers are generated, but I would assume that anything that CAN be mapped by Mammoth could be mapped to sigil's split marker. I've not altered Mammoth in that regard.

[DiapDealer]

Originally Posted by Interrobang

Finally, with regards to split markers, what elements can be defined in the style map to be mapped to it besides the horizontal line? It would be great if one could use the Word page break character.

The following should work in your style map:

br[type='page'] => hr.sigil_split_marker

I will incorporate that into my samples. It makes more sense than adding horizontal lines to the original document.

[Interrobang]

Originally Posted by DNSB

...what Sigil uses when you click on Restructure Epub to Sigil Norm.

Originally Posted by DiapDealer

...knowing that it's trivial in Sigil to Restructure to Sigil Norm.

Thank you both for that tidbit of information. I'm learning Sigil as I work on this project and didn't know that command was available.

Thanks
Tim