FontShrinker - tool to subset a font

[meme]

@Hitch - that's what the new report will show - all characters regardless of where they are used as long as they are visible. Oh, and if you are pasting individual files into the program - why don't you just Merge all the files into one then copy and paste 1 file into it.

@JSWolf - the report will only show the individual characters that are in the book. From Jellby's description it would be up to the font shrinking program to determine which 2 or 3 pair character sets could/should be turned into a ligature. So feeding it just a list of the single characters that appear in a text would be useless for determining ligatures - you would have to feed it the actual text.

Although Jellby requested the report primarily I think for font shrinking, the report is also just another piece of information you can check to see if there are unexpected characters, etc.

As for displaying each character per font, that as Toxaris has said is a much more involved challenge. Its not something that will be in the next release. It should be possible to get the text for tags using a specific class, but checking what font is used by the class is a bit unknown. None of this would help with ligatures though without the full text.

I'll give it some thought, but without a font shrinking program to shrink the fonts it doesn't help much. And if there is already a program to do font shrinking on an epub from the beginning, there doesn't seem to be much of a need for it.

[Hitch]

Originally Posted by meme

@Hitch - that's what the new report will show - all characters regardless of where they are used as long as they are visible. Oh, and if you are pasting individual files into the program - why don't you just Merge all the files into one then copy and paste 1 file into it.

Because that was what I was trying to avoid having to do? It's just one more thing, and the list of "one more thing," particularly with the headache of SW coming online, is getting a bit long. I know I sound lazy, but trust me, we aren't--it's just ONE MORE THING. ;-) (And, on a commercial front, I'm sitting here thinking about, "do I font-shrink the original ePUB? Do I wait and do it after the [endless] revision cycle is done?" "If I wait until it's done, the client is running around with an ePUB with the whole font file embedded, and Thor alone knows where THAT will end up," "so, do I shrink each one each time, because they add hundreds of words while they're editing their work?" Thus my desire for not having to merge EVERY ePUB to test for characters/shrinking, if that explains my laziness a bit better.)

Although Jellby requested the report primarily I think for font shrinking, the report is also just another piece of information you can check to see if there are unexpected characters, etc.

As for displaying each character per font, that as Toxaris has said is a much more involved challenge. Its not something that will be in the next release. It should be possible to get the text for tags using a specific class, but checking what font is used by the class is a bit unknown. None of this would help with ligatures though without the full text.

I'll give it some thought, but without a font shrinking program to shrink the fonts it doesn't help much. And if there is already a program to do font shrinking on an epub from the beginning, there doesn't seem to be much of a need for it.

Actually, what I said was, I'd be happy even without discreet fonts per usage, just all the chars in a given book for all fonts, but, hey, I'll take discreet font/char usage any day.

H

[JSWolf]

Is there a problem with Calibre's subsetting? One thing it does very well is to remove a font if it's not actually used. So if you embed an entire font family, and the bold italic isn't used, it will be removed.

[Hitch]

Originally Posted by JSWolf

Is there a problem with Calibre's subsetting? One thing it does very well is to remove a font if it's not actually used. So if you embed an entire font family, and the bold italic isn't used, it will be removed.

I just don't like Calibre for conversion. I think it's a fabulous cataloging tool, and I use it every day, but not for conversion. And I think it would be inappropriate for me to use it in our firm, really. I can experiment with it for this aspect of sub-setting, though, and see what happens.

H

[JSWolf]

Originally Posted by Hitch

I just don't like Calibre for conversion. I think it's a fabulous cataloging tool, and I use it every day, but not for conversion. And I think it would be inappropriate for me to use it in our firm, really. I can experiment with it for this aspect of sub-setting, though, and see what happens.

H

What I do is put the ePub into Calibre, convert the ePub to ePub, pull the subsetted fonts out of the converted ePub and replace the fonts in the original ePub with the subsetted fonts. Then I fix the OPF & CSS if any of the fonts were deleted. Then I have the subsetted fonts in the ePub without it having been converted.

[Toxaris]

A quick update. I managed to open an ePUB file and retrieve all unique characters used within the whole ePUB. Next step is to see if I can read the classes in the CSS to see which ones contain font definitions. Still a lot of work to do.

[Turtle91]

You Rock!

[JSWolf]

Originally Posted by Toxaris

A quick update. I managed to open an ePUB file and retrieve all unique characters used within the whole ePUB. Next step is to see if I can read the classes in the CSS to see which ones contain font definitions. Still a lot of work to do.

Very nice! I look forward to this.

[Hitch]

Originally Posted by Turtle91

You Rock!

+1, big-time.

H

[kamanza]

Will it work on 64 bit W7?

[Toxaris]

Of course, but I cannot garantuee a specific x64 build. It would not make much sense or have value.

[Hitch]

Originally Posted by Toxaris

Of course, but I cannot garantuee a specific x64 build. It would not make much sense or have value.

It works flawlessly on mine (Win 7 64-bit), although, as I think I mentioned previously, it has a LOT of steps. ;-) For those of us who are trying to do this from a completed ePUB with chapters, it has a goodly number of steps. I love it, don't get me wrong!

Hitch

[najgori]

Got error: "The given key was not present in the dictionary".

[Toxaris]

Hitch, the newer version should have a lot less steps, but reading the CSS has some challanges...

Najgori: Is there more to the error? This doesn't ring a bell, as I don't use dictionaries.

[najgori]

check your mail


edit: it' windows 7/ 64bit fresh instalation

edit2: it's only one letter anyway - ŷ - just some nineteen century serbian classic with some accented letters. i don't think that is a problem, anyway.