CSS to Approximate a Telegram?

[enuddleyarbl]

Originally Posted by Turtle91

...

p.first::first-line {font-variant:small-caps; font-size:1.15em}

...

Usually, manually converting a couple of ALL CAPS words at the start of each sentence to proper sentence case isn't too much of an issue. But, I was just playing with a book with 130+ chapters with many ALL CAP words and that was just too much.

I found a thread of yours on the Sigil forum from about 10 years ago where you asked about converting these clumps of ALL CAP words to sentence case. Have you found anything that works well in Calibre?

[Turtle91]

Not quite. I think the creators of Sigil are contemplating having some kind of sentence case/title case option... Right now I think Calibre has the regex functions which could do it but I haven't practiced enough with it to feel competent. I use the following regex to get it mostly done:

find: <p class="first">([A-Z])([A-Z ]+)*
replace: <p class="first">\1\L\2\E

I put the class="first" to designate the first paragraph in a chapter...you'll have to adjust.

I roll through the ePub one at a time in case it tries to change proper names or 'I' to lower case.

Not completely automatic...but 130 chapters shouldn't take more than a few minutes.

[Tex2002ans]

Originally Posted by enuddleyarbl

I found a thread of yours on the Sigil forum from about 10 years ago where you asked about converting these clumps of ALL CAP words to sentence case. Have you found anything that works well in Calibre?

I wrote about this back in:

• 2014: "Upper to Lower Case Regex - I'm stuck!"

and more recently in:

• 2020: "Add title="" to h* based on existing TOC -- suggestion for new feature (or plugin?)"

There were a few other discussions in there:

• Regex method
• Calibre's "Function mode"

But honestly, proper Title Casing is a very hard problem, and all tools I've tested have flaws. "Edge cases" wrong like:

• Names
• Acronyms
• Accents
• Roman Numerals
• Letters after punctuation

(For example, I do not recommend CapitalizeMyTitle anymore. Whatever algorithm they used years ago has progressively gotten worse, I believe ever since they got purchased by that rotten Grammarly.)

There are a few python libraries, like:

• titlecase

but this, too, has serious flaws.

Calibre does a decent job with ALL CAPS -> Titlecase... definitely one of the better ones.

What you would need is a custom TitleCasing plugin where you can check options (like what to do after hyphens).

- - -

Side Note: Personally, what I do when having to work on hundreds/thousands of chapter titles, is:

• Export a giant list of all titles.
  
  • For example, ripping out every <h2>ALL CAPS TITLE</h2>
• Run entire list through a capitalize tool.
• Run original list through secondary capitalize tool.
• Compare both lists together.
• Skim down list to make manual corrections.
  
  • Words like "of", "and", ...

Then I could easily map:

<h2>ALL CAPS TITLE</h2>

to the equivalent line in:

<h2>All Caps Title</h2>

and do a mass Search/Replace.

This means I could also have an "All Caps" title in the TOC while still keeping the "ALL CAPS" visually in the book itself (just like I described in Post #15).

Originally Posted by Turtle91

Not completely automatic...but 130 chapters shouldn't take more than a few minutes.

Heh. Yeah, I still do a lot of manual stuff.

To this date, I still haven't found a satisfactory tool for me.

One of the best capitalizers I've found is in TeXStudio. It has an Edit > Text Operations > To Titlecase (smart) which is quite good.

So sometimes, I even copy/paste text into there, then run it, then paste back where needed.

[enuddleyarbl]

Originally Posted by Turtle91

...
find: <p class="first">([A-Z])([A-Z ]+)*
replace: <p class="first">\1\L\2\E

I don't recognize \L or \E in regex (neither does Calibre). Is that particular to Sigil?

EDIT: An example of what I get as a result:

W\LHAT EMBLEMS\E

EDIT 2: According to:

https://regex101.com/

\L and \E are specific to "PCRE (PHP <7.3)" and Calibre's editor is Python based. I can't find an equivalent.

[Turtle91]

If you are already using Calibre then I would just look up Regex Functions in the Calibre User's Guide. There is an example in there for Title Case, but that could just be slightly modified for Sentence case.

Sorry... I haven't used Calibre's Editor since Sigil came alive again, so I'm not as proficient with it. If you have issues you may find some smart Calibre people on their forum...although there are a few that poke their head's in here occasionally!

[Turtle91]

I did a google and found this...might help point you in the general direction if you don't want to try the regex functions.

find: ([\.\r\n][ ]*)([A-Za-z\r])([^\.^\r^\n]+)
replace: \1\U\2\L\3

[Quoth]

Not all style guides agree on nuances of Title case. Also German is different to English anyway with nouns in sentences and if doing pre-20th C book or replicated style they had different rules then.

[Karellen]

Those full uppercase first few words of a chapter are pretty annoying. They are ugly and very intrusive when trying to read.

I pieced together this regex that works well in most cases, except for the usual "I" and proper nouns. I can flick through a hundred chapters in a couple of minutes, only stopping when I have to re-capitalise a name.

Find... </span>\b\s?[A-Z’]+(?:\s+[A-Z\.’,\s]+)*\b

Replace... regex function - Lower case (ignore tags)

[Hitch]

Originally Posted by bookman156

There is nothing simpler than using c2sc and smcp. If they are unsupported or a different font is used by the reader you get a better result as fallback. What's difficult about that?

If the font is changed or c2sc and smcp aren't supported, the full caps will stay as full caps, whereas the smcp can be backed up with font-variant-caps: small-caps; so you still get small caps if that is supported.

You've done it the best as is possible.

Yeah, because these function tags are built INTO the font--which 99% of the time, aren't. effectively, you may as well be embedding a proper SC face into the eBook, with now-unnecessary media-query fallbacks if it doesn't work.

And it still requires that you have a device that can work with embedded fonts and certain CSS.

Hitch

[Tex2002ans]

Just wanted to put this example from my latest conversion here:



I think it can speak for itself.

[Hitch]

Originally Posted by Tex2002ans

Just wanted to put this example from my latest conversion here:

Attachment 196500

I think it can speak for itself.

I'm sure it does, but apparently I'm deaf. What is it precisely that you're demo'ing here, Tex?

H

[Turtle91]

Smallcaps can be used with lowercase AND uppercase letters - and the uppercase glyph is used while still maintaining a font-size difference. It is also using a font that looks like a telegram or an old typewriter.

It looks really nice actually!

[Hitch]

Originally Posted by Turtle91

Smallcaps can be used with lowercase AND uppercase letters - and the uppercase glyph is used while still maintaining a font-size difference. It is also using a font that looks like a telegram or an old typewriter.

It looks really nice actually!

Yes, it does look nice, but...never mind. I officially give up.

Hitch

[JSWolf]

To do smallcaps, the characters have to be lowercase. Hence text-transform: lowercase;. If you are converting smallcaps done with a font size to real smallcaps, then you don't need to edit the uppercase text to make it lowercase.

[Turtle91]

Originally Posted by JSWolf

To do smallcaps, the characters have to be lowercase. Hence text-transform: lowercase;. If you are converting smallcaps done with a font size to real smallcaps, then you don't need to edit the uppercase text to make it lowercase.

You are incorrect Jon.

Please take the time to actually read what was written previously before making these blanket statements that are demonstrably false.

Here's an exercise for you:

Show us the html you would use to create the following (make sure you keep the difference in font size between the upper and lower case letters) and you must use the upper-case glyph for all the lower-case letters:

"Now Is The Time For All Good Men To Come To The Aid Of Their Countries."

Here's mine:

.sc {font-variant:small-caps}

<p class="sc">Now Is The Time For All Good Men To Come To The Aid Of Their Countries.</p>