no issue at all; update the readme

Author: folkvir

README.md

@@ -313,18 +313,19 @@ The HyperLogLog algorithm is able to estimate cardinalities greather than `10e9`
 const {HyperLogLog} = require('bloom-filters')
 
 // create a new HyperLogLog with 100 registers
-const sketch = new HyperLogLog(100)
-
-// push some occurrences in the sketch
-sketch.update('alice')
-sketch.update('alice')
-sketch.update('bob')
-
-// count occurrences
-console.log(sketch.count())
+const sketch = new HyperLogLog(128)
+// push 10000 distinct elements
+const n = 2 ** 14
+for (let i = 0; i<n; i++) {
+    sketch.update(i.toString())
+}
 
-// print accuracy
-console.log(sketch.accuracy())
+// print the relative error
+console.log(sketch.relative_error()) // 1.04 / Math.sqrt(128)
+// print the approximated number of distinct elements
+console.log(sket.count())
+// print the real error; should always be less than n * sketch.relative_error() * 3
+console.log(n - sketch.count())
 ```
 
 ### MinHash

tests/hyperloglog.test.ts

@@ -109,3 +109,16 @@ test('should create an HyperLogLog from a JSON export', () => {
     expect(newFilter._correctionBias).toEqual(sketch._correctionBias)
     expect(newFilter._registers).toEqual(sketch._registers)
 })
+
+test('issue#(https://github.com/Callidon/bloom-filters/issues/69)', () => {
+    // create a new HyperLogLog with 100 registers
+    const sketch = new HyperLogLog(128)
+    // push 10000 distinct elements
+    const n = 2 ** 14
+    for (let i = 0; i<n; i++) {
+        sketch.update(i.toString())
+    }
+    // count occurrences
+    expect(sketch.relative_error()).toEqual(1.04 / Math.sqrt(128))
+    expect(n - sketch.count()).toBeLessThan(n * sketch.relative_error() * 3)
+})