consimilo.minhash — consimilo 0.1.1

build-minhash clj

(build-minhash feature-coll)

(build-minhash [feature & features] hashvalues)

iterates through a document feature collection: ['token-1' token-2' ... 'token-n], updating the minhash with each feature. Complete minhash is returned.

large-prime clj

merge-minhash clj

(merge-minhash minhash1 minhash2)

merges two minhashes together by taking the elementwise minimum between the two minhash vectors

permutations clj

update-minhash clj

(update-minhash hashvalues feature)

updates minhash with each document feature (token, shingle, n-gram, etc...) Tokens are hashed using sha1 hash and truncated at max-hash to allow hashing of documents with varying feature sizes. One minhash should be created for each document