Joseph Brenner > Text-Capitalize-1.3 > Text::Capitalize

Download:
Text-Capitalize-1.3.tar.gz

Dependencies

Annotate this POD

CPAN RT

New  3
Open  3
View/Report Bugs
Module Version: 1.3   Source  

NAME ^

Text::Capitalize - capitalize strings ("to WORK AS titles" becomes "To Work as Titles")

SYNOPSIS ^

   use Text::Capitalize;

   print capitalize( "...and justice for all" ), "\n";
      ...And Justice For All

   print capitalize_title( "...and justice for all" ), "\n";
      ...And Justice for All

   print capitalize_title( "agent of SFPUG", PRESERVE_ALLCAPS=>1 ), "\n";
      Agent of SFPUG

   print capitalize_title( "the ring:  symbol or cliche?",
                           PRESERVE_WHITESPACE=>1 ), "\n";
      The Ring:  Symbol or Cliche?
      (Note, double-space after colon is still there.)

   # To work on international characters, may need to set locale
   use Env qw( LANG );
   $LANG = "en_US";
   print capitalize_title( "über maus" ), "\n";
      Ãœber Maus

   use Text::Capitalize qw( scramble_case );
   print scramble_case( 'It depends on what you mean by "mean"' );
      It dEpenDS On wHAT YOu mEan by "meAn".

ABSTRACT ^

  Text::Capitalize is for capitalizing strings in a manner
suitable for use in titles.

DESCRIPTION ^

Text::Capitalize provides some routines for title-like formatting of strings.

The simple capitalize function just makes the inital character of each word uppercase, and forces the rest to lowercase.

The capitalize_title function applies English title case rules (discussed below) where only the "important" words are supposed to be capitalized. There are also some customization features provided to allow the user to choose variant rules.

Comparing capitalize and captialize_title:

  Input:             "lost watches of splitsville"
  capitalize:        "Lost Watches Of Splitsville"
  capitalize_title:  "Lost Watches of Splitsville"

Some examples of formatting with capitalize_title:

  Input:             "KiLLiNG TiMe"
  capitalize_title:  "Killing Time"

  Input:             "we have come to wound the autumnal city"
  capitalize_title:  "We Have Come to Wound the Autumnal City"

  Input:             "ask for whom they ask for"
  captialize_title:  "Ask for Whom They Ask For"

Text::Capitalize also provides some functions for special effects such as scramble_case, which typically would be used for this sort of transformation:

  Input:            "get whacky"
  scramble_case:    "gET wHaCkY"  (or something similar)

EXPORTS ^

default exports

capitalize

Makes the inital character of each word uppercase, and forces the rest to lowercase.

The original routine by Stanislaw Y. Pusep.

capitalize_title

Applies English title case rules (See BACKGROUND) where only the "important" words are supposed to be capitalized.

The one required argument is the string to be capitalized.

Some customization options may be passed in as pairs of names and values following the required argument.

The following customizations are allowed:

Boolean:

  PRESERVE_WHITESPACE
  PRESERVE_ALLCAPS
  PRESERVE_ANYCAPS

Array reference:

  NOT_CAPITALIZED

See "Customizing the Exceptions to Capitalization".

optional exports

@exceptions

The list of minor words that don't usually get capitalized in titles (used by capitalize_title). Defaults to:

     a an the
     and or nor for but so yet
     to of by at for but in with has
     de von
%defaults_capitalize_title

Defines the default arguments for the capitalize_title function Initially, this is set-up to shut off the features PRESERVE_WHITESPACE, PRESERVE_ALLCAPS and PRESERVE_ANYCAPS; it also has @exceptions as the NOT_CAPITALIZED list.

scramble_case

This routine provides a special effect: sCraMBliNg tHe CaSe

The algorithm here uses a modified probability distribution to get a weirder looking effect than simple randomization such as with random_case.

For a discussion of the algorithm, see "SPECIAL EFFECTS".

random_case

Randomizes the case of each character with a 50-50 chance of each one becoming upper or lower case.

zippify_case

Function to provide a special effect: "RANDOMLY upcasing WHOLE WORDS at a TIME".

This uses a similar algorithm to scramble_case, though it also ignores words on the @exceptions list, just as capitalize_title does.

BACKGROUND ^

The capitalize_title function tries to do the right thing by default: adjust an arbitrary chunk of text so that it can be used as a title. But as with many aspects of the human languages, it is extremely difficult to come up with a set of programmatic rules that will cover all cases.

Words that don't get capitalized

This web page:

  http://www.continentallocating.com/World.Literature/General2/LiteraryTitles2.htm

presents some admirably clear rules for capitalizing titles:

  ALL words in EVERY title are capitalized except
  (1) a, an, and the,
  (2) two and three letter conjunctions (and, or, nor, for, but, so, yet),
  (3) prepositions.
  Exceptions:  The first and last words are always capitalized even
  if they are among the above three groups.

But consider the case:

  "It Waits Underneath the Sea"

Should the word "underneath" be downcased because it's a preposition? Most English speakers would be surprised to see it that way. Consequently, the default list of exceptions to capitalization in this module only includes the shortest of the common prepositions (to of by at for but in).

The default entries on the exception list are:

     a an the
     and or nor for but so yet
     to of by at for but in with has
     de von

The observant may note that the last row is not composed of English words. The honorary "de" has been included in honor of "Honoré de Balzac". And "von" was added for the sake of equal time.

Customizing the Exceptions to Capitalization

If you have different ideas about the "rules" of English (or perhaps if you're trying to use this code with another language with different rules) you might like to substitute a new exception list of your own:

  capitalize_title( "Dude, we, like, went to Old Slavy, and uh, they didn't have it",
                     NOT_CAPITALIZED => [ qw( uh duh huh wha like man you know ) ] );

This should return:

   Dude, We, like, Went To Old Slavy, And uh, They Didn't Have It

Less radically, you might like to simply add a word to the list, for example "from":

   use Text::Capitalize 0.2 qw( capitalize_title @exceptions );
   push @exceptions, "from";

   print capitalize_title( "fungi from yuggoth",
                           NOT_CAPITALIZED => \@exceptions);

This should output:

    Fungi from Yuggoth

All Uppercase Words

In order to work with a wide range of input strings, by default capitalize_title presumes that upper-case input needs to be adjusted (e.g. "DOOM APPROACHES!" would become "Doom Approaches!"). But, this doesn't allow for the possibilities such as an acronym in a title (e.g. "RAM Prices Plummet" ideally should not become "Ram Prices Plummet"). If the PRESERVE_ALLCAPS option is set, then it will be presumed that an all-uppercase word is that way for a reason, and will be left alone:

   print capitalize_title( "ram more RAM down your throat",
                           PRESERVE_ALLCAPS => 1 );

This should output:

      Ram More RAM Down Your Throat

Preserving Any Usage of Uppercase for Mixed-case Words

There are some other odd cases that are difficult to handle well, notably mixed-case words such as "iMac", "CHiPs", and so on. For these purposes, a PRESERVE_ANYCAPS option has been provided which presumes that any usage of uppercase is there for a reason, in which case the entire word should be passed through untouched. With PRESERVE_ANYCAPS on, only the case of all lowercase words will ever be adjusted:

   print capitalize_title( "TLAs i have known and loved",
                       PRESERVE_ANYCAPS => 1 );

This should output:

   TLAs I Have Known and Loved

   print capitalize_title( "the next iMac: just another NeXt?",
                            PRESERVE_ANYCAPS => 1);

This should output:

   The Next iMac: Just Another NeXt?

Handling Whitespace

By default, the capitalize_title function presumes that you're trying to clean up potential title strings. As an extra feature it collapses multiple spaces and tabs into single spaces. If this feature doesn't seem desirable and you want it to literally restrict itself to adjusting capitalization, you can force that behavior with the PRESERVE_WHITESPACE option:

   print capitalize_title( "it came from texas:  the new new world order?",
                           PRESERVE_WHITESPACE => 1);

This should output:

      It Came From Texas:  The New New World Order?

(Note: the double-space after the colon is still there.)

Comparison to Text::Autoformat

As you might expect, there's more than one way to do this, and these two pieces of code perform very similar functions:

   use Text::Capitalize 0.2;
   print capitalize_title( $t ), "\n";

   use Text::Autoformat;
   print autoformat { case => "highlight", right => length( $t ) }, $t;

Note: with autoformat, supplying the length of the string as the "right margin" is much faster than plugging in an arbitrarily large number. There doesn't seem to be any other way of turning off line-breaking (e.g. by using the "fill" parameter) though possibly there will be in the future.

As of this writing, "capitalize_title" has some advantages:

  1. It works on characters outside the English 7-bit ASCII range, for example with my locale setting (en_US) the ISO-8859-1 International characters are handled correctly, so that "über maus" becomes "Über Maus".
  2. Minor words following leading punctuation become upper case:
       "...And Justice for All"
  3. It works with multiple sentence input (e.g. "And sooner. And later." should probably not be "And sooner. and later.")
  4. The list of minor words is more extensive (i.e. includes: so, yet, nor), and is also customizable.
  5. There's a way of preserving acronyms via the PRESERVE_ALLCAPS option and similarly, mixed-case words ("iMac", "NeXt", etc") with the PRESERVE_ANYCAPS option.
  6. capitalize_title is roughly ten times faster.

Another difference is that Text::Autoformat's "highlight" always preserves whitespace something like capitalize_title does with the PRESERVE_WHITESPACE option set.

However, it should be pointed out that Text::Autoformat is under active maintenance by Damian Conway. It also does far more than this module, and you may want to use it for other reasons.

Still more ways to do it

Late breaking news: The second edition of the Perl Cookbook has just come out. It now includes: "Properly Capitalizing a Title or Headline" as recipe 1.14. You should familiarize yourself with this if you want to become a true master of all title capitalization routines.

(And I see that recipe 1.13 includes a "randcap" program as an example, which as it happens does something like the random_case function described below...)

SPECIAL EFFECTS ^

Some functions have been provided to make strings look weird by scrambling their capitalization ("lIKe tHiS"): random_case and scramble_case. The function "random_case" does a straight-forward randomization of capitalization so that each letter has a 50-50 chance of being upper or lower case. The function "scramble_case" performs a very similar function, but does a slightly better job of producing something "weird-looking".

The difficulty is that there are differences between human perception of randomness and actual randomness. Consider the fact that of the sixteen ways that the four letter word "word" can be capitalized, three of them are rather boring: "word", "Word" and "WORD". To make it less likely that scramble_case will produce dull output when you want "weird" output, a modified probability distribution has been used that records the history of previous outcomes, and tweaks the likelihood of the next decision in the opposite direction, back toward the expected average. In effect, this simulates a world in which the Gambler's Fallacy is correct ("Hm... red has come up a lot, I bet that black is going to come up now."). "Streaks" are much less likely with scramble_case than with random_case.

Additionally, with scramble_case the probability that the first character of the input string will become upper-case has been tweaked to less than 50%. (Future versions may apply this tweak on a per-word basis rather than just on a per-string basis).

There is also a function that scrambles capitalization on a word-by-word basis called "zippify_case", which should produce output like: "In my PREVIOUS life i was a LATEX-novelty REPAIRMAN!"

EXPORT ^

By default, this version of the module provides the two functions capitalize and capitalize_title. Future versions will have no further additions to the default export list.

Optionally, the following functions may also be exported:

scramble_case

A function to scramble capitalization in a wEiRD loOOkInG wAy. Supposed to look a little stranger than the simpler random_case output

random_case

Function to randomize capitalization of each letter in the string. Compare to "scramble_case"

zippify_case

A function like "scramble_case" that acts on a word-by-word basis (Somewhat LIKE this, YOU know?).

It is also possible to export the following variables:

@exceptions

The list of minor words that capitalize_title uses by default to determine the exceptions to capitalization.

%defaults-capitalize_title

The hash of allowed arguments (with defaults) that the capitalize_title function uses.

BUGS ^

1. In capitalize_title, quoted sentence terminators are treated as actual sentence breaks, e.g. in this case:

     'say "yes but!" and "know what?"'

The program sees the ! and effectively treats this as two separate sentences: the word "but" becomes "But" (under the rule that last words must always be uppercase, even if they're on the exception list) and the word "and" becomes "And" (under the first word rule).

2. There's no good way to automatically handle names like "McCoy". Consider the difficulty of disambiguating "Macadam Roads" from "MacAdam Rode". If you need to solve problems like this, consider using the case_surname function of Lingua::En::NameParse.

3. In general, Text::Capitalize is a very parochial English oriented module that looks like it belongs in the "Lingua::En::*" tree.

4. There's currently no way of doing a PRESERVE_ANYCAPS that *also* adjusts capitalization of words on the exception list, so that "iMac Or iPod" would become "iMac or iPod".

SEE ALSO ^

Text::Autoformat

"The Perl Cookbook", second edition, recipes 1.13 and 1.14

Lingua::En::NameParse

About "scramble_case": http://obsidianrook.com/devnotes/talks/esthetic_randomness/

VERSION ^

Version 0.9

AUTHORS ^

   Joseph M. Brenner
      E-Mail:   doom@kzsu.stanford.edu
      Homepage: http://obsidianrook.com/map

   Stanislaw Y. Pusep  (who wrote "capitalize")
      E-Mail:   stanis@linuxmail.org
      ICQ UIN:  11979567
      Homepage: http://sysdlabs.hypermart.net/

And many thanks (for feature suggestions and code examples) to:

    Belden Lyman, Yary Hcluhan, Randal Schwartz

COPYRIGHT AND LICENSE ^

Copyright 2003 by Joseph Brenner. All rights reserved.

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

syntax highlighting: