search

itkach / aard2-android

Aard2 for Android, a simple dictionary app
GNU General Public License v3.0
464 stars 98 forks source link

Needs to have proper stemming #58

Open thdoan opened 8 years ago

thdoan commented 8 years ago

In it's current form the word lookup is not very useful if used from another app because most words that you read in English have tenses or gerunds, so the lookup would fail or lead to hilarious results. For example, I just looked up "peeped" from Moon+ Reader and Aard 2 brought me to the definition for "pee-pee" LOL.

The WordNet app has proper stemming, so maybe you can borrow some of their code?

Frenzie commented 8 years ago

Isn't that more something to take care of while you're building a slob file? That being said, maybe there should be tools or tips to help with that.

On Fri, Dec 2, 2016 at 11:13 AM, Tom Doan notifications@github.com wrote:

In it's current form the word lookup is not very useful if used from another app because most words that you read in English have tenses or gerunds, so the lookup would fail or lead to hilarious results. For example, I just looked up "peeped" from Moon+ Reader and Aard 2 brought me to the definition for "pee-pee" LOL.

The WordNet app has proper stemming, so maybe you can borrow some of their code?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/itkach/aard2-android/issues/58, or mute the thread https://github.com/notifications/unsubscribe-auth/AAMYBfgRtyhejN4vIQwjs-Hpqc-Sd6TWks5rD-85gaJpZM4LCZIF .

nikita-moor commented 5 years ago

It's called morphology search and usually is based on Hunspell (Nuspell, MySpell) library. Some Android dictionary shells, such as GoldenDict Mobile and MDict, use it quite effective. Hope Aard would support morphology too.

@Frenzie Just stemming is not enough for flective languages. Some words could have hundreds of forms in agglutinative languages, so including them all would bloat the dictionary.