jthomperoo / predictive-horizontal-pod-autoscaler

Horizontal Pod Autoscaler built with predictive abilities using statistical models
Apache License 2.0
330 stars 30 forks source link

holt winters python 3-8 algorithm code fails due to outdated dependencies of statsmodel 0.12.1 #64

Closed igorshevach closed 2 years ago

igorshevach commented 2 years ago

Describe the bug running holt winters model errors out with: Cannot import name '_centered' from 'scipy.signal.signaltools'

To Reproduce Steps to reproduce the behavior:

  1. Deploy latest version of phpa
  2. Run holt-winters example
  3. See error as described above

Expected behavior holt winters should run to completion without errors

Kubernetes Details (kubectl version): Client Version: version.Info{Major:"1", Minor:"22", GitVersion:"v1.22.5", GitCommit:"5c99e2ac2ff9a3c549d9ca665e7bc05a3e18f07e", GitTreeState:"clean", BuildDate:"2021-12-16T08:38:33Z", GoVersion:"go1.16.12", Compiler:"gc", Platform:"darwin/amd64"} Server Version: version.Info{Major:"1", Minor:"22", GitVersion:"v1.22.5", GitCommit:"5c99e2ac2ff9a3c549d9ca665e7bc05a3e18f07e", GitTreeState:"clean", BuildDate:"2021-12-16T08:32:32Z", GoVersion:"go1.16.12", Compiler:"gc", Platform:"linux/amd64"} predictive-horizontal-pod-autoscaler commit a8f581d43d689c8ae439ad12b9d2425d4419949d

Additional context modifying docker image to use cpa python 3-7 instead 3-8 solves the issue, however this is just a workaround. upgrading to latest statsmodels/scipy or both should solve the issue

jthomperoo commented 2 years ago

Resolved by #67

jthomperoo commented 2 years ago

This is now available in v0.10.0.