Improving concordance method

[thiagogomesverissimo]

@yooper,

First, thank you for your efforts in this library, it is so great.
@Euak and I are working in a project and we did some changes that we believe improve the scope of the concordance method from Corpus class.
Changes:

• Switched from stripos to preg_match_all to find $needle positions
• The change above allow us to find $needle in any part of a word
• It possible now choose if $needle is case insensitive or sensitive

Some things that can be better worked in the future:

• substr and mb_substr did not handle well with latin accent (like ~, ^ etc) so we did utf8_decode before everything and utf8_encode in return
• We inserted spaces before and after $text to capture first and last word, maybe we need to change the regex

Minor changes:

• A public domain Brazilian book added to tests
• A new test added to concordance method using this book

[yooper]

Thanks for your contributions. I will pull it down and test it out and merge within the next couple days.

[Euak]

Thanks, @yooper . It's very kind of you merging it so fast.