Converting to EPUB without making a TOC

rrt · 11-27-2012, 10:05 AM

I would like to make short documents into EPUB files. They do not need a TOC: I am thinking about individual short stories, poems &c.

I can find lots of hints on how to generate a TOC, but I can't find out how NOT to generate one! I'm starting from LaTeX, and going via LaTeXML quite successfully:

latexml --dest="$book.xml" "$book.tex"
latexmlpost --nodefaultcss --dest="$book.html" "$book.xml"
ebook-convert "$book.html" "$book.epub" --language en --no-default-epub-cover "$@"

Is there some way to make ebook-convert not make a TOC?

PeterT · 11-27-2012, 10:20 AM

From http://manual.calibre-ebook.com/cli/...le-of-contents maybe

Quote:

--no-chapters-in-toc¶
Don’t add auto-detected chapters to the Table of Contents.

rrt · 11-27-2012, 10:51 AM

Thanks; I should have mentioned I already tried that. With --no-chapters-in-toc, an empty TOC is still generated.

JSWolf · 11-28-2012, 12:21 AM

An ePub has to have an NCX file even if it is empty.

rrt · 11-28-2012, 03:51 AM

Sorry, I was imprecise. When I said that Calibre generates an empty TOC, I meant that it generates a TOC with just one entry, for the entire document. Here's a shell script to remove the offending TOC entry; however, it's a horrible hack, and what I'm really after is a way to avoid generating it in the first place:

#!/bin/sh
# remove-epub-toc
# Remove the TOC from an EPUB
# (Designed to work with Calibre-generated EPUBs with a single section; see FIXMEs for generalization)
# Reuben Thomas 27th November 2012

# Shell function from atool(1)
aunpack () {
TMP=`mktemp /tmp/aunpack.XXXXXXXXXX`
atool -x --save-outdir=$TMP "$@"
DIR="`cat $TMP`"
[ "$DIR" != "" -a -d "$DIR" ] && cd "$DIR"
rm $TMP
}

book=$1
shift

# FIXME: edit the XML properly rather than grepping
# FIXME: ensure identifiers are correct, computed if necessarily

aunpack -F zip "$book"
rm *_split_000.html
sed -i -e 's/split_000/split_001/' toc.ncx
grep -v html2 content.opf > content.opf.new
mv content.opf.new content.opf
apack -F zip "$book" .
mv "$book" ..
cd ..
rm -rf "$DIR"

Agama · 11-29-2012, 03:00 AM

I think that perhaps the toc.ncx is a mandatory item in an ePub, and therefore calibre generates a default when you do not create one.

I checked a small ePub with the IDPF validator, (http://validator.idpf.org/), after removing the toc.ncx file and its reference in content.opf: it failed validation with a message that indicated the ncx was missing. Putting the toc back caused it to pass validation, comfirming that this file is required.

rrt · 11-29-2012, 07:10 AM

As you can see from my script (above), I do not remove toc.ncx. I simply want an empty TOC so that when I open the book in a reader I am taken straight to the title page. Calibre generates a TOC with a single entry, which is the title of the work, and it's ugly and pointless. (For similar reasons, my very short ebooks don't have a title page either.)

Gunnerp245 · 11-29-2012, 08:01 AM

Deleted incorrect advice...

rrt · 11-29-2012, 08:03 AM

I'm sorry, I don't understand what you mean by a "parameter" in "look and feel", or what "\" means. I re-read the "Look and feel" section of http://manual.calibre-ebook.com/conversion.html and that didn't help. Could you please elaborate?

Agama · 11-29-2012, 08:08 AM

[redundant posting: someone else took up the thread]

kovidgoyal · 11-29-2012, 08:11 AM

An empty NCX is not allowed by the EPUB spec. http://idpf.org/epub/20/spec/OPF_2.0...m#Section2.4.1 And http://www.daisy.org/z3986/2005/ncx-2005-1.dtd

rrt · 11-29-2012, 08:23 AM

I do not have an empty NCX. Here is my NCX:

<?xml version='1.0' encoding='utf-8'?>
<ncx xmlns="http://www.daisy.org/z3986/2005/ncx/" version="2005-1" xml:lang="en">
<head>
<meta content="98160948-0367-400e-ad9b-07537e37a9ef" name="dtb:uid"/>
<meta content="2" name="dtb:depth"/>
<meta content="calibre (0.8.51)" name="dtb:generator"/>
<meta content="0" name="dtb:totalPageCount"/>
<meta content="0" name="dtb:maxPageNumber"/>
</head>
<docTitle>
<text>Foo</text>
</docTitle>
<navMap>
<navPoint id="deb7899f-3f43-42e1-aba4-be8924c74e48" playOrder="1">
<navLabel>
<text>Start</text>
</navLabel>
<content src="foo_split_001.html"/>
</navPoint>
</navMap>
</ncx>

The complete ebook gives several validation errors for the actual content (apparently Calibre has some bugs), but it makes no complaint about the toc.ncx or any other part of the navigation apparatus. (To be clear: the EPUB as originally output by Calibre gives exactly the same validation errors as the EPUB I produce by editing the TOC.)

kovidgoyal · 11-29-2012, 08:36 AM

You are creating an NCX with an entry that leads nowhere. Which is a lot uglier than an entry that leads to the start of the book. And is utterly pointless.

rrt · 11-29-2012, 08:40 AM

I'm sorry, I don't understand. I haven't created (or removed) any entry in the NCX. I merely changed the link to "foo_split_001.html" (it previously said "foo_split_000.html"). Everything there was created by Calibre.

rrt · 11-29-2012, 08:47 AM

Maybe it would help if I restated my original question in a more user-oriented way:

I have some very short documents (short stories, poems) that I would like to make into EPUBs. When I try this, I open the EPUB and land on a contents page with a single entry, for the start of the poem/story. This seems redundant: if I published the poem/story on its own, it would not have a contents page. How can I tell Calibre not to generate this contents list? (Note, I am not trying to avoid having an NCX file, or in any way violate the EPUB spec, I just want to improve how the book looks to the reader.)

11-27-2012, 10:05 AM	#1
rrt Member Posts: 10 Karma: 10 Join Date: Nov 2012 Device: none	Converting to EPUB without making a TOC I would like to make short documents into EPUB files. They do not need a TOC: I am thinking about individual short stories, poems &c. I can find lots of hints on how to generate a TOC, but I can't find out how NOT to generate one! I'm starting from LaTeX, and going via LaTeXML quite successfully: latexml --dest="$book.xml" "$book.tex" latexmlpost --nodefaultcss --dest="$book.html" "$book.xml" ebook-convert "$book.html" "$book.epub" --language en --no-default-epub-cover "$@" Is there some way to make ebook-convert not make a TOC?

11-28-2012, 12:21 AM	#4
JSWolf Resident Curmudgeon Posts: 34,660 Karma: 14439052 Join Date: Nov 2006 Location: Roslindale, Massachusetts Device: Sony Reader PRS-650, iPad, nook STR	An ePub has to have an NCX file even if it is empty. __________________ Jon 2013 Reading List

11-29-2012, 08:01 AM	#8
Gunnerp245 Gadget Freak Posts: 1,035 Karma: 1025474 Join Date: Nov 2007 Location: US Device: Sony 700; Entourage Edge, Kindle 3, Pocket Edge	Deleted incorrect advice... __________________ User of both the 10" EE & 7" PE. Visit edge/pocket edge forum. calibre User Manual. Last edited by Gunnerp245; 11-29-2012 at 08:31 PM.

11-29-2012, 08:08 AM	#10
Agama Fanatic Posts: 555 Karma: 422221 Join Date: Jul 2010 Location: UK Device: Sony PRS-300 & Kindle PW	[redundant posting: someone else took up the thread] Last edited by Agama; 11-29-2012 at 08:10 AM.

11-29-2012, 08:11 AM	#11
kovidgoyal Creator of calibre Posts: 22,660 Karma: 3473290 Join Date: Oct 2006 Location: Mumbai, India Device: Various	An empty NCX is not allowed by the EPUB spec. http://idpf.org/epub/20/spec/OPF_2.0...m#Section2.4.1 And http://www.daisy.org/z3986/2005/ncx-2005-1.dtd __________________ Get calibre Notice to all: I can not provide assistance with DRM removal, for legal reasons, so please do not contact me about it.

11-27-2012, 10:51 AM	#3
rrt Member Posts: 10 Karma: 10 Join Date: Nov 2012 Device: none	Thanks; I should have mentioned I already tried that. With --no-chapters-in-toc, an empty TOC is still generated.

11-28-2012, 03:51 AM	#5
rrt Member Posts: 10 Karma: 10 Join Date: Nov 2012 Device: none	Sorry, I was imprecise. When I said that Calibre generates an empty TOC, I meant that it generates a TOC with just one entry, for the entire document. Here's a shell script to remove the offending TOC entry; however, it's a horrible hack, and what I'm really after is a way to avoid generating it in the first place: #!/bin/sh # remove-epub-toc # Remove the TOC from an EPUB # (Designed to work with Calibre-generated EPUBs with a single section; see FIXMEs for generalization) # Reuben Thomas 27th November 2012 # Shell function from atool(1) aunpack () { TMP=`mktemp /tmp/aunpack.XXXXXXXXXX` atool -x --save-outdir=$TMP "$@" DIR="`cat $TMP`" [ "$DIR" != "" -a -d "$DIR" ] && cd "$DIR" rm $TMP } book=$1 shift # FIXME: edit the XML properly rather than grepping # FIXME: ensure identifiers are correct, computed if necessarily aunpack -F zip "$book" rm *_split_000.html sed -i -e 's/split_000/split_001/' toc.ncx grep -v html2 content.opf > content.opf.new mv content.opf.new content.opf apack -F zip "$book" . mv "$book" .. cd .. rm -rf "$DIR"

11-29-2012, 03:00 AM	#6
Agama Fanatic Posts: 555 Karma: 422221 Join Date: Jul 2010 Location: UK Device: Sony PRS-300 & Kindle PW	I think that perhaps the toc.ncx is a mandatory item in an ePub, and therefore calibre generates a default when you do not create one. I checked a small ePub with the IDPF validator, (http://validator.idpf.org/), after removing the toc.ncx file and its reference in content.opf: it failed validation with a message that indicated the ncx was missing. Putting the toc back caused it to pass validation, comfirming that this file is required.

11-29-2012, 07:10 AM	#7
rrt Member Posts: 10 Karma: 10 Join Date: Nov 2012 Device: none	As you can see from my script (above), I do not remove toc.ncx. I simply want an empty TOC so that when I open the book in a reader I am taken straight to the title page. Calibre generates a TOC with a single entry, which is the title of the work, and it's ugly and pointless. (For similar reasons, my very short ebooks don't have a title page either.)

11-29-2012, 08:03 AM	#9
rrt Member Posts: 10 Karma: 10 Join Date: Nov 2012 Device: none	I'm sorry, I don't understand what you mean by a "parameter" in "look and feel", or what "\" means. I re-read the "Look and feel" section of http://manual.calibre-ebook.com/conversion.html and that didn't help. Could you please elaborate?

11-29-2012, 08:23 AM	#12
rrt Member Posts: 10 Karma: 10 Join Date: Nov 2012 Device: none	I do not have an empty NCX. Here is my NCX: <?xml version='1.0' encoding='utf-8'?> <ncx xmlns="http://www.daisy.org/z3986/2005/ncx/" version="2005-1" xml:lang="en"> <head> <meta content="98160948-0367-400e-ad9b-07537e37a9ef" name="dtb:uid"/> <meta content="2" name="dtb:depth"/> <meta content="calibre (0.8.51)" name="dtb:generator"/> <meta content="0" name="dtb:totalPageCount"/> <meta content="0" name="dtb:maxPageNumber"/> </head> <docTitle> <text>Foo</text> </docTitle> <navMap> <navPoint id="deb7899f-3f43-42e1-aba4-be8924c74e48" playOrder="1"> <navLabel> <text>Start</text> </navLabel> <content src="foo_split_001.html"/> </navPoint> </navMap> </ncx> The complete ebook gives several validation errors for the actual content (apparently Calibre has some bugs), but it makes no complaint about the toc.ncx or any other part of the navigation apparatus. (To be clear: the EPUB as originally output by Calibre gives exactly the same validation errors as the EPUB I produce by editing the TOC.)

11-29-2012, 08:36 AM	#13
kovidgoyal Creator of calibre Posts: 22,660 Karma: 3473290 Join Date: Oct 2006 Location: Mumbai, India Device: Various	You are creating an NCX with an entry that leads nowhere. Which is a lot uglier than an entry that leads to the start of the book. And is utterly pointless. __________________ Get calibre Notice to all: I can not provide assistance with DRM removal, for legal reasons, so please do not contact me about it.

11-29-2012, 08:40 AM	#14
rrt Member Posts: 10 Karma: 10 Join Date: Nov 2012 Device: none	I'm sorry, I don't understand. I haven't created (or removed) any entry in the NCX. I merely changed the link to "foo_split_001.html" (it previously said "foo_split_000.html"). Everything there was created by Calibre.

11-29-2012, 08:47 AM	#15
rrt Member Posts: 10 Karma: 10 Join Date: Nov 2012 Device: none	Maybe it would help if I restated my original question in a more user-oriented way: I have some very short documents (short stories, poems) that I would like to make into EPUBs. When I try this, I open the EPUB and land on a contents page with a single entry, for the start of the poem/story. This seems redundant: if I published the poem/story on its own, it would not have a contents page. How can I tell Calibre not to generate this contents list? (Note, I am not trying to avoid having an NCX file, or in any way violate the EPUB spec, I just want to improve how the book looks to the reader.)

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search


Advertisement Enthusiast

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
ToC detected but not transferred when converting from epub to pdf	Eoannis	Conversion	2	03-16-2012 04:41 AM
epub to epub converting: making chapters?	Joy736	Conversion	8	10-30-2011 04:47 PM
converting from HTM to ePUb - can't get TOC right	davidm99	Conversion	1	04-12-2011 10:47 PM
TOC disappears when converting from .prc to epub	baselworld	Calibre	3	09-06-2010 02:10 AM
Converting a web page to epub with TOC	philosopherdog	Calibre	5	07-23-2010 07:55 AM