search

pixelogik / NearPy

Python framework for fast (approximated) nearest neighbour search in large, high-dimensional data sets using different locality-sensitive hashes.
MIT License
763 stars 151 forks source link

Possibility to update vectors #42

Closed lboudard closed 8 years ago

lboudard commented 8 years ago

Hi,

I think it would be very useful to have possibility to update vectors index. Typically, a CF application would need to update movies/users vectors regularilly as users buy/see movies. From my undertanding of the model, it would require 1) vectorize movie before update 2) retrieve buckets containing movie vectors and remove it from set 3) vectorize movie after update and push it. While it's certainly doable, there might be some more intermediary data structures that could help in that task.

Thanks!

pixelogik commented 8 years ago

Sounds cool, but I do not have time for this currently. If you want to participate, let me know.

MaxwellRebo commented 8 years ago

Any resolution to this?

MaxwellRebo commented 8 years ago

@lboudard this has been here for a bit, any interest in taking it back up?

pixelogik commented 8 years ago

I still don't have time, sorry. Too busy with work. But if you guys wanna contribute I can add you as collaborators.

adityapatadia commented 8 years ago

I will look after PR if someone is ready to send.

On 23 June 2016 at 2:29:31 PM, Ole (notifications@github.com) wrote:

I still don't have time, sorry. Too busy with work. But if you guys wanna contribute I can add you as collaborators.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/pixelogik/NearPy/issues/42#issuecomment-227990130, or mute the thread https://github.com/notifications/unsubscribe/ABCUmQgEothfaC-MVSueGIZc5WOGLk3Jks5qOkrvgaJpZM4Gqhmu .

neerajBaji commented 8 years ago

Hi,

I am working on a similar CF problem whereby I would need to update the index. Can you sign me on as a contributor? I will give this a shot.

erramuzpe commented 8 years ago

Hi @pixelogik , @adityapatadia , @neerajBaji and @MaxwellRebo

I'm also interested in this. The way of doing it should be to look for a given index and deleting both the index and the vector corresponding to that index. Then, the update will come through a new indexing operation. Seems easy, but I'm not sure where to start from.

cheers!

erramuzpe commented 8 years ago

Related to #27 @audy

erramuzpe commented 8 years ago

I think I have it! It would be nice if someone could do this for RedisStorage too :+1: