Spell checker
From HIPpie
Contents |
Syntax
http://library.hud.ac.uk/hippie/
< library ID >/
spellchecker/
option/option/option.../
< word or phrase >
Info
The spell checker uses GNU Aspell - http://aspell.net
The HIPpie code is currently in its testing phase and anyone working in a library or educational establishment can use it freely. However, there may come a day when you'll need to either register or apply for a free API key.
Examples
Options
The detaults options are to return plain text output using the "ultra" suggestion mode (view text samples)
- xml - return XML output (view sample)
- xmlcount - include number count in XML output (requires xml)
- best - only return the "best guess" for each word
- secondbest - if the suggestion matches the original keywords, then make a second best guess
- whole - only include whole words in the output (does not return multi-word suggestions)
- limit3 - only return up to 3 suggestions for each word
- limit5 - only return up to 5 suggestions for each word
- limit10 - only return up to 10 suggestions for each word
- ultra (default) - use the "ultra" suggestion mode (see below)
- fast - use the "fast" suggestion mode (see below)
- normal - use the "normal" suggestion mode (see below)
- badspellers - use the "bad-spellers" suggestion mode (see below)
Suggestion Modes
From http://aspell.net/0.50-doc/man-html/4_Customizing.html#suggestion
- ultra
This method will use the fastest method available to come up with decent suggestions. This currently means that it will look for soundslikes within one edit distance apart without doing any typo analysis. It is slower than Ispell by a factor of 1.5 to 2 when a single word list is used. It speed is only minor affected by the size of the word list, if at all, but it is strongly effected by the number of word lists use. In this mode Aspell gets about 87% of the words from my small test kernel of misspelled words. (Go to http://aspell.net/testfor more info on the test kernel as well as comparisons of this version of Aspell with previous versions and other spell checkers.)
- fast
This method is like ultra except that it also performs typo analysis unless it is turned off by setting the keyboard to none. The typo analysis brings words which are likely to be due to typos to the beginning of the list but slows things down by a factor of about two. This mode should get around the same number of words that the ultra method does.
- normal
This method looks for soundslikes within two edit distance apart and perform typo-analysis unless it is turned off. Is is around 10 times slower than fast mode with the english word list but returns better suggestions. Its speed is directly proportional to the size of the word list. This mode gets 93% of the words.
- bad-spellers
This method also looks for soundslikes within two edit distances apart but is more tailored for the bad speller where as fast or normal are more tailed to strike a good balance between typos and true misspellings. This mode never performs typo-analysis and returns a huge number of words for the really bad spellers who can't seam to get the spelling anything close to what it should be. If the misspelled word looks anything like the correct spelling it is bound to be found somewhere on the list of 100 or more suggestions. This mode gets 98% of the words.
