Which libraries support incremental item additions?

[alexklibisz]

Thanks for putting this project together. Is there any way a table or small list could be included to describe which libraries support incremental item additions? I imagine for many this is an important criteria in real-time or near-real-time applications where new content is continuously added and needs to be indexed for look-ups.

[erikbern]

i don't think any of them do, unfortunately :(

[aaalgo]

Last time I checked (many years ago), FLANN's C++ API supported insertion. In theory it's also easy for all LSH-based methods to support insertion. The Python API might not have such functionality exposed though.

[alexklibisz]

@aaalgo Yes from the FLANN website:

(14 December 2012) Version 1.8.0 is out bringing incremental addition/removal of points to/from indexes

Though I'm still trying to figure out how to leverage this via Python.

[alexklibisz]

@erikbern Do you have any recommendations for "online" NN-search from your experience? One approach I'm considering is to maintain an indexed lookup (e.g. Annoy) for the majority of my items as well as a brute-force lookup (e.g. sklearn) for the newest items. Then at some threshold, rebuild the index to include the newest points.

[erikbern]

I don't have any great recommendations, but your idea makes sense though. You could even formalize it and have indexes that have size even powers of two, and any time you have two indexes of the same size, you merge them. That would give you log(n) amortized runtime, although worst case is linear.

[yurymalkov]

@alexklibisz Both hnsw and sw-graph are built incrementally, and thus naturally support incremental item addition. Hovewer, nmslib does not have a python interface for that, only c++.

@erikbern @alexklibisz I think you are talking about Bentley-Saxe method. http://www.sciencedirect.com/science/article/pii/S030643791300104X

[erikbern]

@yurymalkov yeah i've never heard of it, but think that's what i meant. basically play 2048 with the indexes :)

[alexklibisz]

For anyone interested, I built a project that supports the feature from my original question on top of Elasticsearch: https://github.com/alexklibisz/elastik-nearest-neighbors

[yurymalkov]

@alexklibisz Cool
@erikbern There are now several algorithms that support updates: flann, hnswlib, faiss (batched updates). Probably, this should be noted in readme somehow.