From 165f88b814098c7ec960c7777b2ee8e58b9792fe Mon Sep 17 00:00:00 2001
From: Simon Belak <simon.belak@gmail.com>
Date: Fri, 10 Nov 2017 14:42:44 +0100
Subject: [PATCH] Make outliers work with empty input
---
src/metabase/feature_extraction/math.clj | 9 +++++----
test/metabase/feature_extraction/math_test.clj | 11 +++++++----
test/metabase/feature_extraction/timeseries_test.clj | 4 +++-
3 files changed, 15 insertions(+), 9 deletions(-)
diff --git a/src/metabase/feature_extraction/math.clj b/src/metabase/feature_extraction/math.clj
index a3e719a1537..92201172514 100644
--- a/src/metabase/feature_extraction/math.clj
+++ b/src/metabase/feature_extraction/math.clj
@@ -148,7 +148,8 @@
https://en.wikipedia.org/wiki/Interquartile_range"
([xs] (outliers identity xs))
([keyfn xs]
- (let [{:keys [q1 q3 iqr]} (->> xs (transduce (map keyfn) h/histogram) h/iqr)
- lower-bound (- q1 (* 1.5 iqr))
- upper-bound (+ q3 (* 1.5 iqr))]
- (remove (comp #(< lower-bound % upper-bound) keyfn) xs))))
+ (when (not-empty xs)
+ (let [{:keys [q1 q3 iqr]} (->> xs (transduce (map keyfn) h/histogram) h/iqr)
+ lower-bound (- q1 (* 1.5 iqr))
+ upper-bound (+ q3 (* 1.5 iqr))]
+ (remove (comp #(< lower-bound % upper-bound) keyfn) xs)))))
diff --git a/test/metabase/feature_extraction/math_test.clj b/test/metabase/feature_extraction/math_test.clj
index d3a53800ca0..2ff7fda7416 100644
--- a/test/metabase/feature_extraction/math_test.clj
+++ b/test/metabase/feature_extraction/math_test.clj
@@ -73,19 +73,22 @@
:lag 1}
{:autocorrelation -1.0
:lag 1}
- nil nil nil]
+ nil nil nil nil]
[(autocorrelation (range 10))
(autocorrelation [1 -1 1 -1 1 -1])
(autocorrelation [1 2 3]) ; not significant
(autocorrelation [1])
- (autocorrelation [])])
+ (autocorrelation [])
+ (autocorrelation nil)])
(expect
[nil
- #{50 100 35}]
+ #{50 100 35}
+ nil]
(let [xs (vec (repeatedly 100 rand))]
[(not-empty (outliers xs))
(set (outliers (-> xs
(assoc-in [10] 50)
(assoc-in [30] 100)
- (assoc-in [70] 35))))]))
+ (assoc-in [70] 35))))
+ (outliers nil)]))
diff --git a/test/metabase/feature_extraction/timeseries_test.clj b/test/metabase/feature_extraction/timeseries_test.clj
index fe0733fc05f..9b4300d554b 100644
--- a/test/metabase/feature_extraction/timeseries_test.clj
+++ b/test/metabase/feature_extraction/timeseries_test.clj
@@ -49,7 +49,9 @@
(expect
[[99]
+ []
[]]
[(breaks 12 (map vector (range) (concat (repeat 100 10)
(repeat 100 20))))
- (breaks 12 (map vector (range) (take 100 (cycle (range 10)))))])
+ (breaks 12 (map vector (range) (take 100 (cycle (range 10)))))
+ (breaks 4 nil)])
--
GitLab