docbook-apps

While working on my large set of docbook source files, which I've
converted from the original html source, I discovered the need for a
tool to check the spelling. I bashed together a Perl wrapper around
the program 'hunspell' so the program can be run against a set of
files, read the output, eliminate unnecessary lines, and use a hash to
eliminate duplicates.

I'm sure it can be improved, but I leave that as an exercise for the
user. I uploaded it to the DocBook wiki at the bottom of this page:

http://wiki.docbook.org/topic/ConvenienceTools#preview

The Perl program is called "check_spelling.pl" and a limited help
message is shown if you execute it without any arguments.

Regards,

-Tom

Thomas M. Browder, Jr.
Niceville, Florida
USA

Am 26.06.2010 15:24, schrieb Tom Browder:
> While working on my large set of docbook source files, which I've
> converted from the original html source, I discovered the need for a
> tool to check the spelling. I bashed together a Perl wrapper around
> the program 'hunspell' so the program can be run against a set of
> files, read the output, eliminate unnecessary lines, and use a hash to
> eliminate duplicates.
>
> I'm sure it can be improved, but I leave that as an exercise for the
> user. I uploaded it to the DocBook wiki at the bottom of this page:
>
> http://wiki.docbook.org/topic/ConvenienceTools#preview
>
> The Perl program is called "check_spelling.pl" and a limited help
> message is shown if you execute it without any arguments.

You mean like :)
aspell check --mode=sgml doc.xml

Stefan


>
> Regards,
>
> -Tom
>
> Thomas M. Browder, Jr.
> Niceville, Florida
> USA
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: docbook-apps-unsubscribe@lists.oasis-open.org
> For additional commands, e-mail: docbook-apps-help@lists.oasis-open.org
>

On Sun, Jun 27, 2010 at 04:43, Stefan Kost <ensonic@hora-obscura.de> wrote:
> Am 26.06.2010 15:24, schrieb Tom Browder:
...
> You mean like :)
> aspell check --mode=sgml doc.xml

Something like that, but that line doesn't work for me.

The reason for the wrapper is I couldn't find any way to handle
multiple files non-interactively. Of course a better solution is
always sought--suggestions welcome.

Regards,

-Tom

On Sun, Jun 27, 2010 at 05:32, Tom Browder <tom.browder@gmail.com> wrote:
> On Sun, Jun 27, 2010 at 04:43, Stefan Kost <ensonic@hora-obscura.de> wrote:
>> Am 26.06.2010 15:24, schrieb Tom Browder:
> ...
>> You mean like :)
>> aspell check --mode=sgml doc.xml

I get it, I could have used aspell instead of hunspell and specifying
the English system dictionary,

Hunspell was the first thing I saw when running apropos on my machine
to look for a spelling program--didn't think to look further. I
forgot about aspell--been a loooong time since I used an external
spell checker

At any rate, aspell should work with my wrapper with suitable command
line changes. It will still help eliminate the cruft and simplify
spell checking a large set of files.

Thanks.

Regards,

-Tom

On Sun, Jun 27, 2010 at 12:47 PM, Tom Browder <tom.browder@gmail.com> wrote:
> On Sun, Jun 27, 2010 at 05:32, Tom Browder <tom.browder@gmail.com> wrote:
>> On Sun, Jun 27, 2010 at 04:43, Stefan Kost <ensonic@hora-obscura.de> wrote:
>>> Am 26.06.2010 15:24, schrieb Tom Browder:
>> ...
>>> You mean like :)
>>> aspell check --mode=sgml doc.xml
>
> I get it, I could have used aspell instead of hunspell and specifying
> the English system dictionary,
>
> Hunspell was the first thing I saw when running apropos on my machine
> to look for a spelling program--didn't think to look further.  I
> forgot about aspell--been a loooong time since I used an external
> spell checker
>
> At any rate, aspell should work with my wrapper with suitable command
> line changes.  It will still help eliminate the cruft and simplify
> spell checking  a large set of files.

Here is my two cents:

xmllint --postvalid --xinclude --nonet ${input_file} | aspell
list -p path/to/aspell.en.pws --mode=sgml --lang=en --encoding=utf-8

Because this trigger quite a lot of false positive I also append:

foreach( skip in `acronym application author code
hardware filename markup programlisting
productname screen sgmltag`)
"--add-f-sgml-skip=${skip}"
endforeach()

HTH
--
Mathieu

On Sun, Jun 27, 2010 at 10:16, Mathieu Malaterre
<mathieu.malaterre@gmail.com> wrote:
> On Sun, Jun 27, 2010 at 12:47 PM, Tom Browder <tom.browder@gmail.com> wrote:
>> On Sun, Jun 27, 2010 at 05:32, Tom Browder <tom.browder@gmail.com> wrote:
>>> On Sun, Jun 27, 2010 at 04:43, Stefan Kost <ensonic@hora-obscura.de> wrote:
>>>> Am 26.06.2010 15:24, schrieb Tom Browder:
...
> Here is my two cents:
>
>      xmllint --postvalid --xinclude --nonet ${input_file} | aspell
...

I don't do that because I want to treat the spelling check as a separate task.

-Tom

Am 27.06.2010 18:16, schrieb Mathieu Malaterre:
> On Sun, Jun 27, 2010 at 12:47 PM, Tom Browder <tom.browder@gmail.com> wrote:
>> On Sun, Jun 27, 2010 at 05:32, Tom Browder <tom.browder@gmail.com> wrote:
>>> On Sun, Jun 27, 2010 at 04:43, Stefan Kost <ensonic@hora-obscura.de> wrote:
>>>> Am 26.06.2010 15:24, schrieb Tom Browder:
>>> ...
>>>> You mean like :)
>>>> aspell check --mode=sgml doc.xml
>>
>> I get it, I could have used aspell instead of hunspell and specifying
>> the English system dictionary,
>>
>> Hunspell was the first thing I saw when running apropos on my machine
>> to look for a spelling program--didn't think to look further. I
>> forgot about aspell--been a loooong time since I used an external
>> spell checker
>>
>> At any rate, aspell should work with my wrapper with suitable command
>> line changes. It will still help eliminate the cruft and simplify
>> spell checking a large set of files.
>
> Here is my two cents:
>
> xmllint --postvalid --xinclude --nonet ${input_file} | aspell
> list -p path/to/aspell.en.pws --mode=sgml --lang=en --encoding=utf-8
>
> Because this trigger quite a lot of false positive I also append:
>
> foreach( skip in `acronym application author code
> hardware filename markup programlisting
> productname screen sgmltag`)
> "--add-f-sgml-skip=${skip}"
> endforeach()
>
> HTH

I use a local dictionary for the exceptions. This is the whole rule in my makefile:

check-local:
@echo "Spellchecking C/$(DOC_MODULE).xml..."
if test 0 -ne `cat $(srcdir)/C/$(DOC_MODULE).xml | aspell list --lang=en_US -p
$(PWD)/$(srcdir)/C/aspell.pws --mode=sgml | wc -l`; then \
echo; \
echo "C/$(DOC_MODULE).xml doesn't pass spellchecking, please fix with" >&2; \
echo " aspell check --lang=en_US -p ./C/aspell.pws --mode=sgml
C/$(DOC_MODULE).xml" >&2; \
echo; \
false; \
else \
echo " done"; \
fi