What Features or Tools does Sigil Still Need Yet?

[Hitch]

Originally Posted by DiapDealer

[Snip helpful explanation of how to actually use Sigil's splitting functions....]

As for the rest of your post: don't let the door hit you. I'm the resident arrogant butthead around here, not Kevin. He's the reasonable polite one.

Perhaps we can all hope that it's simply a language issue and not whingy wankerism.

Hitch

[KevinH]

Originally Posted by elibrarian

And then one small gripe: When I regenerate a TOC, Sigil will preserve the header and the "<nav epub:type='toc' id='toc'>", but removes "role='doc-toc'". It is of course a small thing to re-insert (I always run ACE as one of the last things before publishing my books, and it will tell me if I forget).

This was easy enough to do immediately so aria role for both the toc and any page-list in the nav are now set on creation and recreation. I do not think there is an equivalent aria role for the epub:type="landmarks" so I did not add one there.

Are you aware of an equivalent for that part of the nav?

This change will be in the next release.

Thanks!

[elibrarian]

Originally Posted by KevinH

This was easy enough to do immediately so aria role for both the toc and any page-list in the nav are now set on creation and recreation. I do not think there is an equivalent aria role for the epub:type="landmarks" so I did not add one there.

Are you aware of an equivalent for that part of the nav?

This change will be in the next release.

Thanks!

Not to my knowledge, no (I usually consult this list: https://idpf.github.io/epub-guides/epub-aria-authoring/

And of course: Thanks! One thing less to remember

Regards,

Kim

[KevinH]

To see if using SavedSearches would fit the bill to provide the Atlantis-like replacement capability, I have added the ability for Saved Searches to import both *.txt files - (tab delimited list one per line) and *.csv (comma delimted list one per line). No export yet as exporting can be done to ini files.

Given Sigil SavedSearches can already support Regex find and replace, we get that added capability for free as long as current Find & Replace it set that way.

I am not sure about handling csv files that quote strings data as that would create an issue on import, so for the time being I am assuming all string data is exported as from Excel as csv without strings being quoted.

Either way, I have pushed a very simplistic approach for this to master, that we can play around with and try to see if it will be enough to do what people want for serial lists of replacements.

If anyone builds their own, please pull from master and give it a try and let us know.

KevinH

Originally Posted by Doitsu

On second thought, there actually is no need to create a brand-new new plugin, because users already can create a Saved Search groups.
All that's missing is an import filter for comma/tab-delimited files.

How about simply adding .csv and .txt files as file types to the Import Search Entries dialog box? (.txt would be for tab-delimited files)

[KevinH]

Originally Posted by The_book

2. Yes. Normal/Regex/Case sensitive/Do not change, current/all/selected/do not change.

Okay I took a look at this. It is doable but is counter to how the user manual says to use this feature.

According to the user manual, typically, you are supposed to open Find & Replace to determine your search type and settings. Once set the way you want your right click to get the SavedSearches dialog. You then select the saved search and run it with whatever the current settings are in Find & Replace. That way none of that additional metadata needs to be saved or be part of the ini.

I am not sure changing this is worthwhile for most users, especially those who use it like the manual describes, as it would make all current saved search files obsolete and could end up confusing people.

So unless there is strong support for changing this from current users, it is not something I am in favour of.

Thanks

[elibrarian]

I'm not quite sure that I understand this correctly, but I often find myself fetching a saved search I made as a regex, run it - and it finds nothing. Because the last search/replace I ran was a Normal search, and the Saved search don't remember the type, so happily try to run my Regex search as a Normal search.

If this is what is being discussed, I for one would support a change, so that the saved searches remember at least the search type (but maybe not the range and Up/Down settings?)

Regards,

Kim

[KevinH]

As I mentioned earlier, that would break backwards compatibility with others existing search ini files and require a redesign of the user interface to carry that info. It is possible but since you are in Find & Replace just set it to regex (or not) before loading the saved search.

Let's hear from others first. If there is a strong consensus in its favour, I will consider adding it.

[Vanguard3000]

Great work, guys! Overall, I'm very happy with Sigil. There are a few nice-to-haves I have in mind. They may be specific to my personal workflows and use-cases, but hey, you asked.

Find in: Open Tabs
I've mentioned this before at some point, but I still think it'd be a pretty valuable option. As an example use case, I often find myself fixing up ePubs that have been converted and put together poorly, resulting not only in a lot of header-defined "sgc-x" styles, but inconstent ones. So, some files have sgc-1 as bold and sgc-2 as italics, while others are reversed.

What I'd ideally like to do is search for

p.sgc-1 {font-style: italic}

opening all files with that definition. Then, do a find & replace in all opened tabs for

<p class="sgc-1">(.*?)</p> (or similar)

and replace with traditional italics tags or whatever.

As it is now, I have to find those files, then select them in the sidebar, then use the find in "Selected HTML Files" option. This works, but the selection is lost very easily. If I click a tab, for example.

"Wrappable"/regex Clips
EDIT: This is already in Sigil! My mistake!
I'd love for defining clips to have a similar option to regex replacing using "\1" it would be great to be able to define a clip such as:

<span class="banana">\1</span>

and be able to select some text, add the clip with a click or hotkey, and have it surround the selected text, similar to using the bold/italic/underline buttons and hotkeys.

Multi-Paragraph Justification Changing
It would be nice to be able to make a selection over two or more paragraphs, and use the justification buttons to change all paragraphs touched by the selection highlight. This would allow the use to quickly and easily centre an entire title page, for example, rather than having to select each paragraph individually.

Search "Ignore Tags" option
This could be a checkbox option alongside "DotAll", "minimal Match", etc. Basically, it would allow the use to search for text while ignoring any html tags. For example, in:

<p>The quick <i>brown</i> fox</p>
<p>jumps over the lazy dog.</p>

I could search for "brown fox jumps" and get a hit.

Often when OCRing or fixing up weird ePubs, there may be weird styling added to certain words, or even mid-word, and it would be nice to have the option to find text before I've rooted all that weirdness out.

That's all I can think of at the moment. nothing huge, but these things would definitely improve my process when fixing up ebooks. Thanks again!

[BeckyEbook]

Thank you both for the years of developing the Sigil application.

What am I missing?

• The ability to rename files in BookBrowser via plugins.

Explanation: I KNOW this requires checking all linkages to every other file, but since Rename in the GUI does this all the time, the plug would do it too.

• Ability to enable/disable adding Sigil version to metadata in the OPF file.

Explanation: I would not like to offend the creators and developers of Sigil with this proposal, I simply express my subjective opinion. I am proud to use Sigil, but like having clean metadata. Theoretically, the output plugin could help, though invoking a plugin and simply hitting "Save" makes a difference.

• Templates for the table of contents/index.

Explanation: I mean the possibility to "build" a different layout, because the current one has a hardcoded <div> structure, even if I adapt it to my needs, another TOC refresh will restore the default layout again.
This is where a plugin could help, if there was a function that would give the TOC/index "ingredients" to styleaccording to the current table of contents settings, including "sigil_not_in_toc", "sigil_index_marker" etc. Sure, you could write a plugin to do the table of contents in a different form, but the plugin would do 99% of what Sigil is already doing, adding only the final layout of the TOC.xhtml file.

----------

As for Saved Searches:
In my opinion, the existing solution is not perfect. The fact that it forces you to pre-set all the options and then load saved searches is the weakest point. It also means that - although I like this option - I rarely use it. It may seem strange, but I have quite a few different searches where the key information is whether they are normal or regex searches.
If I save them in Saved Searches in a folder, I always have to remember that these are normal, and these are regex ...
I would like to be able to have any searches in the folder so that it could be performed sequentially, e.g.

1. regex
2. regex
3. normal
4. case sensitive
5. regex, enabled "DotAll"
6. normal
7. regex

etc.

I see different solutions, always taking into account backward compatibility.

Sample solution:
We ignore backward compatibility Just kidding.
For backward compatibility, we ignore the existing sigil_searches.ini file and from version 1.5 only save sigil_searches_new.ini
The file structure may be the same:

[search_entries]
1\Name=XXX
1\Find=YYY
1\Replace=ZZZ

but we add

1\Options=QQQ

If someone insists and downgrades Sigil:
* will have its old file
* if he wants to use newly created searches he can make a copy of the new ini file and name it sigil_searches.ini and IMHO should work with previous versions of Sigil.

[Doitsu]

Originally Posted by Vanguard3000

"Wrappable"/regex Clips
I'd love for defining clips to have a similar option to regex replacing using "\1"[...]

That exact feature already exists. If you highlight text and select a clip with a "\1" in it, the text will be wrapped in it.

Originally Posted by BeckyEbook

For backward compatibility, we ignore the existing sigil_searches.ini file and from version 1.5 only save sigil_searches_new.ini.

I like the idea, but we might as well replace all .ini files with .json files and offer a stand-alone converter for the first Sigil version that no longer uses .ini files.
IIRC, current sigil.ini files are no longer 100% compatible with older versions anyway and .json files must be encoded as UTF-8 files by default. (I know that UTF-16/32 are also allowed, but I've yet to see one "in the wild.")

[KevinH]

There is a lot of machinery around Qt settings .ini files that json does not support. For example, a file watcher is specified and an inter-process communication is used to synchronize settings shared across all of the current instances of the main process.

This machinery would all have to be recreated if we move away completely from ini files. We tried that by moving to utf-8 but since a unicode character cannot be written in a single access and is made up of multiple bytes collisions and data corruption happened.

So moving away from ini to pure standalone json files is probably not going to happen. We lost 3 releases to ini file corruption when we tried just to change it to utf-8 encoding.

[KevinH]

Originally Posted by BeckyEbook

What am I missing?

• The ability to rename files in BookBrowser via plugins.

Explanation: I KNOW this requires checking all linkages to every other file, but since Rename in the GUI does this all the time, the plug would do it too.

The gui does it by invoking a large amount of code in SourceUpdates that handles searching for and updating any links all using multiple threads.

The plugin interface is a file based interface and can not callback to invoke that code. So all of that code would need to be duplicated and rewritten in python and then maintained in parallel for all future versions.

• Ability to enable/disable adding Sigil version to metadata in the OPF file.

Explanation: I would not like to offend the creators and developers of Sigil with this proposal, I simply express my subjective opinion. I am proud to use Sigil, but like having clean metadata. Theoretically, the output plugin could help, though invoking a plugin and simply hitting "Save" makes a difference.

Most software : Word, Excel, PDF, etc includes metadata on what software created it and modification date and time. Sigil deserves the same respect.

With your skills you could easily comment out the offending lines and make your own builds. I think you regularly do that anyway, right? It is just metadata. There are plugin approaches that can do this quite easily too.

• Templates for the table of contents/index.

Explanation: I mean the possibility to "build" a different layout, because the current one has a hardcoded <div> structure, even if I adapt it to my needs, another TOC refresh will restore the default layout again.
This is where a plugin could help, if there was a function that would give the TOC/index "ingredients" to styleaccording to the current table of contents settings, including "sigil_not_in_toc", "sigil_index_marker" etc. Sure, you could write a plugin to do the table of contents in a different form, but the plugin would do 99% of what Sigil is already doing, adding only the final layout of the TOC.xhtml file.

There are already stylesheets you can add to help refine how things are displayed. So I am not sure exactly what you want to change in your "styling". Please generate a complete list of the aspects of the html toc that would need to be changed so that I have some idea of how to modify the creation code to handle those options.


Thanks.

[KevinH]

Originally Posted by Vanguard3000

Find in: Open Tabs
I've mentioned this before at some point, but I still think it'd be a pretty valuable option. As an example use case, I often find myself fixing up ePubs that have been converted and put together poorly, resulting not only in a lot of header-defined "sgc-x" styles, but inconstent ones. So, some files have sgc-1 as bold and sgc-2 as italics, while others are reversed.

What I'd ideally like to do is search for

opening all files with that definition. Then, do a find & replace in all opened tabs for

and replace with traditional italics tags or whatever.

As it is now, I have to find those files, then select them in the sidebar, then use the find in "Selected HTML Files" option. This works, but the selection is lost very easily. If I click a tab, for example.

Not sure how doable that will be but I will look into it.

Multi-Paragraph Justification Changing
It would be nice to be able to make a selection over two or more paragraphs, and use the justification buttons to change all paragraphs touched by the selection highlight. This would allow the use to quickly and easily centre an entire title page, for example, rather than having to select each paragraph individually.

I did not know it worked like that. I think PageEdit handles it differently. Injecting an inline style is probably not the best approach here anyway. Perhaps the ability to select multiple paragraphs and add the same new class attribute to each one would be better. Then in css a single definition can handle things.

I will look into this as well.

Search "Ignore Tags" option
This could be a checkbox option alongside "DotAll", "minimal Match", etc. Basically, it would allow the use to search for text while ignoring any html tags. For example, in:

I could search for "brown fox jumps" and get a hit.

Searching for text across tags is reasonably easy to implement (I hope those are not famous last words!) whereas replacing text that crosses tags is very hard. But since you just asked about search, if the text is found across tags where should the cursor end up once the search is performed and how much of the search term if any should be highlighted if pieces are not adjacent?

Thanks

[Thasaidon]

Originally Posted by DiapDealer

As for the rest of your post: don't let the door hit you. I'm the resident arrogant butthead around here, not Kevin.

Do not beat around the bush. Say what you mean bluntly and do not use so many circumlocutions.

[Thasaidon]

Not a request just an idea for ruminating on. If Sigil already has this feature please let me know as I am not aware of it.

When I am fixing html I often find styling which refers to a single paragraph. This may be inline but usually uses a CSS selector(s).

Normally I just delete this styling manually which means highlighting and deleting pieces of code individually.

As I believe the latest version of Sigil now detects the open and closing tags would it be easily doable to add a feature so that if the cursor is on a tag pressing a hotkey (or clicking on an icon) would remove the tag pair concerned?

This would speed up cleaning these paragraphs.