[newbie] upper to lower first letter of a word

[Yvon Thoraval]

Recently, i get a vintage list (more than 500 items) with poor typo, for
example, i've :

Côte de beaune-villages

instead of :

Côte de Beaune-Villages

Crémant d'alsace

instead of :

Crémant d'Alsace

i wonder of the way to change lower to upper case and also of

a regex able to do the trick.

something like :

every letter following a " ", "-" or "'" should be upper if not
belonging to a black list of words :

black_list = %w{d de du la le sec sur entre etc...}

--
Yvon

[Mark J. Reed]

On Tue, Sep 23, 2003 at 06:29:58PM +0200, Yvon Thoraval wrote:

> Recently, i get a vintage list (more than 500 items) with poor typo, for
> example, i've :
>
> Côte de beaune-villages
>
> instead of :
>
> Côte de Beaune-Villages
>
> Crémant d'alsace
>
> instead of :
>
> Crémant d'Alsace
>
> i wonder of the way to change lower to upper case and also of
>
> a regex able to do the trick.
>
> something like :
>
> every letter following a " ", "-" or "'" should be upper if not
> belonging to a black list of words :
>
> black_list = %w{d de du la le sec sur entre etc...}

string.gsub!(/\b[a-z]+/) { |w| black_list.include?(w) ? w : w.capitalize }

-Mark

[Yvon Thoraval]

Mark J. Reed <markjreed@mail.com> wrote:

> string.gsub!(/\b[a-z]+/) { |w| black_list.include?(w) ? w : w.capitalize }

a lot of tanxs °;)
--
Yvon

[Yvon Thoraval]

Yvon Thoraval <yvon.thoravallist@-SUPPRIMEZ-free.fr.invalid> wrote:

>

> > string.gsub!(/\b[a-z]+/) { |w| black_list.include?(w) ? w : w.capitalize }

>
> a lot of tanxs °;)

it seems, it's a little bit trickier because accentuated characters are
taken as \b for example :

Vosne-romanée
becomes :
Vosne-RomanéE

then instead of \b i would have to exclude a list of chars :
[à|ä|â|é|è|ê|î|ö|ô|ü|ù]
--
Yvon

[Mark J. Reed]

On Tue, Sep 23, 2003 at 07:23:52PM +0200, Yvon Thoraval wrote:

> it seems, it's a little bit trickier because accentuated characters are
> taken as \b

Really? That's arguably a bug. What character encoding are you using?

Accented letters should be in \w, not \W, and therefore the
space between one and an adjacent letter should not match \b.
But Ruby regexes may be ASCII-only, and even if not, they're probably
Latin-1-only. So, for instance, they wouldn't work on UTF-8 strings.

> Vosne-romanée
> becomes :
> Vosne-RomanéE
>
> then instead of \b i would have to exclude a list of chars :
> [à|ä|â|é|è|ê|î|ö|ô|ü|ù]

First, you don't need the pipes (|'s) there. Pipes are for
alternation without the [...]; basically, [abc] is short for
(a|b|c). The pipe form is most useful when the alternatives are
not all single characters, for instance (alfa|bravo|charlie).

I'm not sure whether the exclude-list or the include-list would
be shorter. You could do (^|[- ']) to match "beginning of string or
dash or space or apostrophe", but then that character would be included
in the resulting string. Which means that it would be, for instance,
" d" or "-d" or "'d" instead of "d", and therefore won't be in the
blacklist and won't capitalize properly (since String#capitalize operates
on the first character, which will be the space or dash or apostrophe).
The block has to compensate for that. Something like this:

string.gsub!(/(^|[- '])([a-z]+)/) { $1 + $2.capitalize }

Except that [a-z] won't match accented characters, so it's more like this:

string.gsub!(/(^|[- '])([a-záàâçéèêíìîóòôúùû]+)/) { $1 + $2.capitalize }

And if the names aren't limited to French, then even more special characters
creep in . . .

-Mark

[Mark J. Reed]

On Tue, Sep 23, 2003 at 05:49:24PM +0000, Mark J. Reed wrote:

> string.gsub!(/(^|[- '])([a-záàâçéèêíìîóòôúùû]+)/) { $1 + $2.capitalize }

Left off the blacklist check, which should be applied to $2:

string.gsub!(/(^|[- '])([a-záàâçéèêíìîóòôúùû]+)/) {
black_list.include?($2) ? $1 + $2 : $1 + $2.capitalize
}

-Mark

[Yvon Thoraval]

Mark J. Reed <markjreed@mail.com> wrote:

> Really? That's arguably a bug. What character encoding are you using?

I'm (more-or-less) sure about that because even if i put :
l.gsub!(/\b[a-záàâçéèêíìîóòöôúùüû]+/) { |w| black_list.include?(w) ? w
: w.capitalize }

i get :
MâCon SupéRieur
when input was :
Mâcon supérieur

> Accented letters should be in \w, not \W, and therefore the
> space between one and an adjacent letter should not match \b.
> But Ruby regexes may be ASCII-only, and even if not, they're probably
> Latin-1-only. So, for instance, they wouldn't work on UTF-8 strings.

precisely i'm using utf-8 °;)
however, i'm able to do a try using iso-8859-1, my word editor (Pepper
on MacOS X) is able to transcode within 2 clicks + one cut'n paste rom
utf to iso...
sounds strange to me because Ruby is coming from Japan where "special"
chars are every-day chars ???

[snip]

> The block has to compensate for that. Something like this:
>
> string.gsub!(/(^|[- '])([a-z]+)/) { $1 + $2.capitalize }
>
> Except that [a-z] won't match accented characters, so it's more like this:
>
> string.gsub!(/(^|[- '])([a-záàâçéèêíìîóòôúùû]+)/) { $1 + $2.capitalize }
>
> And if the names aren't limited to French, then even more special characters
> creep in . . .

Yes, right, i know, for the time being, only about french and german
accentuated chars...

However because vintage are classified by area i might have to change
regex upon region...
--
Yvon

[Mark Wilson]

On Tuesday, September 23, 2003, at 12:34 PM, Yvon Thoraval wrote:

> Recently, i get a vintage list (more than 500 items) with poor typo,
> for
> example, i've :
>
> Côte de beaune-villages
>
> instead of :
>
> Côte de Beaune-Villages
>
> Crémant d'alsace
>
> instead of :
>
> Crémant d'Alsace
>
> i wonder of the way to change lower to upper case and also of
>
> a regex able to do the trick.
>
> something like :
>
> every letter following a " ", "-" or "'" should be upper if not
> belonging to a black list of words :
>
> black_list = %w{d de du la le sec sur entre etc...}

You might adapt the English language 'titlecase' program, which can be
found here:

[url]http://zem.novylen.net/ruby/titlecase.rb[/url]

Regards,

Mark

[dblack@superlink.net]

Hi --

On Wed, 24 Sep 2003, Yvon Thoraval wrote:

> Yvon Thoraval <yvon.thoravallist@-SUPPRIMEZ-free.fr.invalid> wrote:
>

> >

> > > string.gsub!(/\b[a-z]+/) { |w| black_list.include?(w) ? w : w.capitalize }

> >
> > a lot of tanxs °;)

>
> it seems, it's a little bit trickier because accentuated characters are
> taken as \b for example :
>
> Vosne-romanée
> becomes :
> Vosne-RomanéE

I believe the /s modifier to the regex will help you here by changing
the encoding, though I'm having character-rendering issues which make
it hard for me to test.... But try this, in the hope that I'm right
even though I can't see the characters:

str.gsub!(/\b[a-z]+/s) {|w| black_list.include?(w) ? w : w.capitalize}


David

--
David Alan Black
home: [email]dblack@superlink.net[/email]
work: [email]blackdav@shu.edu[/email]
Web: [url]http://pirate.shu.edu/~blackdav[/url]

[Yvon Thoraval]

Mark Wilson <mwilson13@cox.net> wrote:

> You might adapt the English language 'titlecase' program, which can be
> found here:
>
> [url]http://zem.novylen.net/ruby/titlecase.rb[/url]

Yes, tanxs, that way i'd change more easily rules versus area of
vintage...
--
Yvon

[Yvon Thoraval]

<dblack@superlink.net> wrote:

> I believe the /s modifier to the regex will help you here by changing
> the encoding, though I'm having character-rendering issues which make
> it hard for me to test.... But try this, in the hope that I'm right
> even though I can't see the characters:
>
> str.gsub!(/\b[a-z]+/s) {|w| black_list.include?(w) ? w : w.capitalize}

tanxs, i don't remember (from Perl) what's the meaning of this "s" ?


--
Yvon

[dblack@superlink.net]

Hi --

On Wed, 24 Sep 2003, Yvon Thoraval wrote:

> <dblack@superlink.net> wrote:
>

> > I believe the /s modifier to the regex will help you here by changing
> > the encoding, though I'm having character-rendering issues which make
> > it hard for me to test.... But try this, in the hope that I'm right
> > even though I can't see the characters:
> >
> > str.gsub!(/\b[a-z]+/s) {|w| black_list.include?(w) ? w : w.capitalize}

>
> tanxs, i don't remember (from Perl) what's the meaning of this "s" ?

It's different in Perl and Ruby. In Perl, it means: treat the string
as a single line, so that '.' matches newline. In Ruby, it affects
the encoding.... I wish I could give a more knowledgeable account,
but I've never actually used it myself and can't seem to dig up
documentation.


David

--
David Alan Black
home: [email]dblack@superlink.net[/email]
work: [email]blackdav@shu.edu[/email]
Web: [url]http://pirate.shu.edu/~blackdav[/url]

[Mark J. Reed]

On Wed, Sep 24, 2003 at 06:07:08AM +0900, [email]dblack@superlink.net[/email] wrote:

> On Wed, 24 Sep 2003, Yvon Thoraval wrote:

> > tanxs, i don't remember (from Perl) what's the meaning of this "s" ?

> It's different in Perl and Ruby. In Perl, it means: treat the string
> as a single line, so that '.' matches newline. In Ruby, it affects
> the encoding.... I wish I could give a more knowledgeable account,
> but I've never actually used it myself and can't seem to dig up
> documentation.
>

According to the Pickaxe, or at least the online version thereof
(my dead-trees vesion is at home), /s means to use the SJIS
(Shift-Japanese Information Systems or something like that) multibyte
text encoding. Similarly, /e means to use EUC, and /u means to use
UTF-8. So /u is probably a better bet than /s for Yvon.


[url]http://www.rubycentral.com/book/ref_c_regexp.html#Regexp.new[/url]

-Mark

[Robert Klemme]

"Yvon Thoraval" <yvon.thoravallist@-SUPPRIMEZ-free.fr.invalid> schrieb im
Newsbeitrag
news:1g1r8u8.1hzv3mvupjeizN%yvon.thoravallist@-SUPPRIMEZ-free.fr.invalid...

> Mark J. Reed <markjreed@mail.com> wrote:
>

> > Really? That's arguably a bug. What character encoding are you

using?

>
> I'm (more-or-less) sure about that because even if i put :
> l.gsub!(/\b[a-záàâçéèêíìîóòöôúùüû]+/) { |w| black_list.include?(w) ? w
> : w.capitalize }

I'd omit the "\b" at the beginning since "é" then still matches a word
boundry:

l.gsub!(/[a-záàâçéèêíìîóòöôúùüû]+/) { |w| black_list.include?(w) ? w
: w.capitalize }

Alternatively:

l.gsub!(/[^\s!?.;:-]+/) {|w| black_list.include?(w) ? w : w.capitalize }

Regards

robert

[Yvon Thoraval]

Robert Klemme <bob.news@gmx.net> wrote:

> I'd omit the "\b" at the beginning since "é" then still matches a word
> boundry:
>
> l.gsub!(/[a-záàâçéèêíìîóòöôúùüû]+/) { |w| black_list.include?(w) ? w
> : w.capitalize }

yes, fine, i discovered also that capitalization don't work on
accentuated chars (as é)

then i've done another step for those "special" chars being as the first
letter of a xord

> Alternatively:
>
> l.gsub!(/[^\s!?.;:-]+/) {|w| black_list.include?(w) ? w : w.capitalize }

ok, however in my list no punctuation as ?!;:... only " " and "-"
--
Yvon

[Carlos]

> yes, fine, i discovered also that capitalization don't work on

> accentuated chars (as é)

You can use an old library named unicode:

irb(main):001:0> $KCODE="u"
=> "u"
irb(main):002:0> require "unicode"
=> true
irb(main):003:0> Unicode.capitalize("Ã*ëÃ*ôů")
=> "ÀëÃ*ôů"

[url]http://raa.ruby-lang.org/list.rhtml?name=unicode[/url]

[Yvon Thoraval]

Carlos <angus@quovadis.com.ar> wrote:

>
> You can use an old library named unicode:
>
> irb(main):001:0> $KCODE="u"
> => "u"
> irb(main):002:0> require "unicode"
> => true
> irb(main):003:0> Unicode.capitalize("àëíô?")
> => "Àëíô?"
>
> [url]http://raa.ruby-lang.org/list.rhtml?name=unicode[/url]

tanxs for all !
--
Yvon

[Thomas A. Reilly]

I would appreciate it if someone could give me the regexp that it would
split the following:
for example -
"clonidine300 mg" into "clonidine 300 mg"

I have a bunch of drug data where the dose had been typed together.

Thanks

[Jim Freeze]

On Saturday, 27 September 2003 at 12:17:19 +0900, Thomas A. Reilly wrote:

> I would appreciate it if someone could give me the regexp that it would
> split the following:
> for example -
> "clonidine300 mg" into "clonidine 300 mg"
>
> I have a bunch of drug data where the dose had been typed together.

There's probably more than one way to do this. Here's one way:

irb(main):001:0> s="clonidine300 mg"
=> "clonidine300 mg"
irb(main):005:0> s.scan(/[a-zA-Z]+|\d+/) { |i| p i }
"clonidine"
"300"
"mg"



--
Jim Freeze
----------
Anybody can win, unless there happens to be a second entry.

[Jim Freeze]

On Saturday, 27 September 2003 at 12:44:11 +0900, Jim Freeze wrote:

> On Saturday, 27 September 2003 at 12:17:19 +0900, Thomas A. Reilly wrote:

> > I would appreciate it if someone could give me the regexp that it would
> > split the following:
> > for example -
> > "clonidine300 mg" into "clonidine 300 mg"
> >
> > I have a bunch of drug data where the dose had been typed together.

>
> There's probably more than one way to do this. Here's one way:
>

And yet another way:


rb(main):018:0> m = /(\w+?)(\d+)\s+(\w+)/.match(s)
=> #<MatchData:0x81c2200>
irb(main):019:0> m[1]
=> "clonidine"
irb(main):020:0> m[2]
=> "300"
irb(main):021:0> m[3]
=> "mg"

--
Jim Freeze
----------
"There is no reason for any individual to have a computer in their
home."
-- Ken Olson, President of DEC, World Future Society
Convention, 1977