I have got the Jazzy code I downloaded today working, using the dictionary provided. But one of my test cases shows the on-line demo on SF to perform much better than the downloaded code.
The specific test is "nurs", which should have "nurse" as the best suggestion, or at elast one of them. That's so on line, but my version suggests the following:
furs nuns nuts curs ours
The online demo returns "nurse" as the first choice, then returns the same as above. What do I need to do to adjust this behavior? It's obviously not the word list per se. Is at a configurable parameter, an additional dictionary - if so, what, where, and how do I load it? - or something else?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Well, I may have found an answer to my own conundrum. I've found that if I don't load the "phonetic transformation" file, then the list of suggestions for "nurs" also includes "nurse", which is certainly what I want.
My few other test cases still look OK. Could someone talk a bit about what the phonetic file is supposed to be good for, and forwaht kinds of misspellings it makes things better?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I agree with you. I was just playing around with Jazzy a bit, and I find that for the sentence
ths is a tst
it doesn't return 'this' as a suggestion for 'ths'. I think its a problem with not having the correct phonetic file loaded, but I can't tell for sure. I tried to create my dictionary like the javadoc for ASpellDictionary says, but I'm still having this problem. I should mention that for the same sentence aspell itself does return 'this' as a suggestion.
I'm using ubuntu Edgy Eft as my source for the dictionary (and I've also tried with a phonetic file en_phonet.dat from /usr/lib/aspell), but no luck. Ubuntu comes with a lot of aspell dictionaries installed by default, but I grepped through all the english ones and none of them contain the word 'this', so I think that to hit on this suggestion depends on the algorithm used.
Does anyone have any advice?
Thanks!
J.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I have got the Jazzy code I downloaded today working, using the dictionary provided. But one of my test cases shows the on-line demo on SF to perform much better than the downloaded code.
The specific test is "nurs", which should have "nurse" as the best suggestion, or at elast one of them. That's so on line, but my version suggests the following:
furs nuns nuts curs ours
The online demo returns "nurse" as the first choice, then returns the same as above. What do I need to do to adjust this behavior? It's obviously not the word list per se. Is at a configurable parameter, an additional dictionary - if so, what, where, and how do I load it? - or something else?
Well, I may have found an answer to my own conundrum. I've found that if I don't load the "phonetic transformation" file, then the list of suggestions for "nurs" also includes "nurse", which is certainly what I want.
My few other test cases still look OK. Could someone talk a bit about what the phonetic file is supposed to be good for, and forwaht kinds of misspellings it makes things better?
Thomas,
I agree with you. I was just playing around with Jazzy a bit, and I find that for the sentence
ths is a tst
it doesn't return 'this' as a suggestion for 'ths'. I think its a problem with not having the correct phonetic file loaded, but I can't tell for sure. I tried to create my dictionary like the javadoc for ASpellDictionary says, but I'm still having this problem. I should mention that for the same sentence aspell itself does return 'this' as a suggestion.
I'm using ubuntu Edgy Eft as my source for the dictionary (and I've also tried with a phonetic file en_phonet.dat from /usr/lib/aspell), but no luck. Ubuntu comes with a lot of aspell dictionaries installed by default, but I grepped through all the english ones and none of them contain the word 'this', so I think that to hit on this suggestion depends on the algorithm used.
Does anyone have any advice?
Thanks!
J.