Modifying the Spellchecker

Hi,

Has anyone worked to make the LUA-based spellchecker function more like the proprietary spellchecker, in terms of options? Right now it's a bit beyond me, but I figured perhaps others have already tackled it. I know that Nick was kind enough to add within the script a way to tell it to ignore capitalized words, but it'd be very useful if the script had some of the other options from the propiertary spellchecker in it:

Ignore All-Caps words
Ignore words with numbers
Ignore words with mixed-case
Report doubled words (e.g. the the)
Case-sensitive

Ideally, it'd be done in such a way that you could uncomment the ones you want to use. Not sure if that's doable, though.

Thanks.

That shouldn't be too hard to do. Just let me know how the old spellchecker handled these things, and I'll get to it. I mostly need to know you used to set the flags in the old one. I never use spellcheckers in my mud, although I probably should start with area building I'm doing now.

In the proprietary spellchecker, there was a GUI with checkboxes for the options you wanted ennabled.

Here's a URL to a screenshot:

http://tinypic.com/view.php?pic=2uzeds7

Well, I can't get that dialog exactly, but I'm sure I can approximate it with some of the provided MUSHclient functions.

That'd be cool. It doesn't really so much if it's done via a GUI or something. For example, in the script you can define IGNORE_CAPITALIZED as true, and the script will then take that into account when evaluating text to be spellchecked. I thought approaching it that way wouldn't be a bad idea. I tried to do an IGNORE_MIXED based on IGNORE_CAPITALIZED, with some help on the regexp, but it doesn't quite seem to work (possibly I've got the REGEXP wrong).

For reference, my try at ignoring mixed case words:

if IGNORE_MIXED then
-- ignore words featuring mixed case
if string.find (word, "^([a-z].*[A-Z])|^(.*([a-z].*[A-Z])|([A-Z].*[a-z]))") then
return word, "ignore"
end -- this round, ignore this word
end -- if IGNORE_MIXED

The Lua regular expressions do not support the "or" feature that the MUSHclient ones do. However this will work:





It already ignores words which are all numbers.

However if you mean to take something like "route66" as a single word, and then ignore it, you need to do something like this:

First, change the "word" regexp near the start of the spellcheck file from:


-- this is the pattern we use to find "words" in the text to be spell-checked
local pattern = "%a+'?%a+" -- regexp to give us a word with a possible single imbedded quote



to:


-- this is the pattern we use to find "words" in the text to be spell-checked
local pattern = "%a+'?[%a%d]+" -- regexp to give us a word with a possible single imbedded quote


Now it recognises imbedded numbers in a word.

Further down, (where you had the IGNORE_MIXED) stuff, add a test that now drops those words:



In other words, the first change makes numbers be considered as part of a word, the second change makes us ignore such words.