From 0634634857ea4cca9e9bdb3f6244eb13afcae2f9 Mon Sep 17 00:00:00 2001
From: Stijn <s.j.rotman@uvt.nl>
Date: Wed, 17 Jan 2024 12:10:24 +0100
Subject: [PATCH] Remove scipy dependency.

---
 frm/_frm_py/preprocessing.py | 8 +++-----
 pyproject.toml               | 1 -
 tests/test_data.py           | 3 +--
 3 files changed, 4 insertions(+), 8 deletions(-)

diff --git a/frm/_frm_py/preprocessing.py b/frm/_frm_py/preprocessing.py
index e5a9c85..de31c2d 100644
--- a/frm/_frm_py/preprocessing.py
+++ b/frm/_frm_py/preprocessing.py
@@ -4,9 +4,7 @@
 """
 from typing import Union
 
-from scipy.stats import zscore
-from numpy import nan_to_num
-
+import numpy as np
 # https://stackoverflow.com/a/60617044
 numeric = Union[int, float]
 
@@ -61,9 +59,9 @@ def standardise(timeseries: list[list]):
         Database of standardised time series.
     """
     try:
-        return nan_to_num(zscore(timeseries, axis=1))
+        return np.nan_to_num((timeseries - np.mean(timeseries, axis=1, keepdims=True)) / np.std(timeseries, axis=1, keepdims=True))
     except ValueError:
-        return [nan_to_num(zscore(ts)) for ts in timeseries]
+        return [np.nan_to_num((ts - np.nanmean(ts)) / np.nanstd(ts)) for ts in timeseries]
 
 
 # Defined in Lin, Keogh, Linardi, & Chiu (2003). A Symbolic Representation of Time Series,
diff --git a/pyproject.toml b/pyproject.toml
index ae9276a..d1c1820 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -23,7 +23,6 @@ classifiers = [
 ]
 dependencies = [
     "numpy",
-    "scipy",
 ]
 
 [project.optional-dependencies]
diff --git a/tests/test_data.py b/tests/test_data.py
index 3249682..b7e3193 100644
--- a/tests/test_data.py
+++ b/tests/test_data.py
@@ -1,5 +1,4 @@
 import numpy as np
-from scipy.stats import zscore
 
 ts = [
     [0, 1, 2, 2, 1, 0],
@@ -61,4 +60,4 @@
 ]
 
 np.random.seed(0)
-data = [zscore(np.random.random(np.random.randint(10, 1000))).tolist() for _ in range(100)]
+data = [np.random.random(np.random.randint(10, 1000)).tolist() for _ in range(100)]