update manifest and allow building BFF with super low number of items

Manifest.toml

@@ -1,8 +1,10 @@
 # This file is machine-generated - editing it directly is not advised
 
-[[Random]]
+manifest_format = "2.0"
+
+[[deps.Random]]
 deps = ["Serialization"]
 uuid = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 
-[[Serialization]]
+[[deps.Serialization]]
 uuid = "9e88b42a-f829-5b0c-bbe9-9e923198166b"

Project.toml

@@ -1,7 +1,7 @@
 name = "FastFilter"
 uuid = "5470531f-076d-47a9-aacb-e15e1e7f44ac"
 authors = ["JokingHero <[email protected]>"]
-version = "0.1.0"
+version = "0.1.1"
 
 [deps]
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"

README.md

@@ -8,7 +8,7 @@ they are never wrong when item was actually in the initial set. What is more,
 space usage and speed of access is excellent for these data structures. 
 
 This Julia package is based of [binaryfusefilter.h](https://github.com/FastFilter/xor_singleheader/blob/master/include/binaryfusefilter.h).
-Currently only binary fuse filters are re-implemented, but we support UInt8, UInt16 and UInt32 fingerprints.
+Currently, only binary fuse filters are re-implemented, but we support UInt8, UInt16 and UInt32 fingerprints.
 
 Feel free to make a PR with other implementations.
 

src/binaryfusefilter.jl

@@ -25,9 +25,9 @@ end
 
 function calculate_size_factor(arity::UInt32, n_keys::UInt32)
 	if arity == 3
-		return Float64(max(1.125, 0.875 + 0.25 * log(1000000)/log(Float64(n_keys))))
+		return Float64(max(1.125, 0.875 + 0.25 * log(1000000)/log(Float64(n_keys + 1))))
 	elseif arity == 4
-		return Float64(max(1.075, 0.77 + 0.305 * log(600000)/log(Float64(n_keys))))
+		return Float64(max(1.075, 0.77 + 0.305 * log(600000)/log(Float64(n_keys + 1))))
 	else
 		return Float64(2.0)
 	end
@@ -143,7 +143,7 @@ function BinaryFuseFilter{T}(
 		end
 
 		startPos = zeros(UInt32, block)
-		for i in 1:length(startPos)
+		for i in eachindex(startPos)
 			# important: we do not want i * n_keys to overflow!!!
 			# in java we have (int) ((long) i * size / block)
 			startPos[i] = UInt32(((UInt64(i - 1) * UInt64(n_keys)) >> blockBits))

test/runtests.jl

@@ -3,11 +3,17 @@ using FastFilter
 using Random
 
 utypes = [UInt8, UInt16, UInt32]
-error_rates = [0.005, 1.5e-5, 2.33e-10]
+error_rates = [0.005, 1.7e-5, 2.33e-10]
 rng = MersenneTwister(42)
 
-for i in 1:length(utypes)
+for i in eachindex(utypes)
     @testset "binaryfusefilter.jl " begin
+
+        @testset "Can be build with stupid small number of items" begin
+            filter = BinaryFuseFilter{utypes[i]}([UInt64(1)])
+            @test UInt64(1) in filter
+        end
+
         n = Int(230e6)
         items = rand(rng, UInt64, n)
         filter = BinaryFuseFilter{utypes[i]}(items)