Fluctuation Filter Module

name	Fluctuation Filter
version	v1.0.1
GitHub	weevenetwork/fluctuation-filter
authors	Mithila Ghuge, Paul Gaiduk

---

Table of Content

• Fluctuation Filter Module
  • Table of Content
  • Description
  • Module Variables
  • Module Testing
  • Dependencies
  • Input
  • Output
    • Example

---

Description

This is a Python Fluctuation Filter module and it is used to eliminate unwanted spikes from input data and provide stable output. It will delay the change of the value filtering out specific amount (instances) of different, non-persisting value.

Module Variables

There are 5 module variables that are required by each module to correctly function within weeve ecosystem. In development, these variables can overridden for testing purposes. In production, these variables are set by weeve Agent.

Environment Variables	type	Description
INPUT_LABEL	string	The input label on which anomaly is detected
WINDOW_SIZE	integer	The stable instance count which need to be consider
SEND_ON_CHANGE	boolean	Set true to Output data only when stable value changes
MODULE_NAME	string	Name of the module
MODULE_TYPE	string	Type of the module (Input, Processing, Output)
LOG_LEVEL	string	Allowed log levels: DEBUG, INFO, WARNING, ERROR, CRITICAL. Refer to logging package documentation.
INGRESS_HOST	string	Host to which data will be received
INGRESS_PORT	string	Port to which data will be received
EGRESS_URLS	string	HTTP ReST endpoint for the next module

Module Testing

To test module navigate to test directory. In test/assets edit both .json file to provide input for the module and expected output. During a test, data received from the listeners are compared against expected output data. You can run tests with make run_test.

Dependencies

The following are module dependencies:

• bottle
• requests

The following are developer dependencies:

• pytest
• flake8
• black

Input

Input to this module is JSON body single object or array of objects:

Example:

{
  "temperature": 15
}

[
  {
    "temperature": 15
  },
  {
    "temperature": 21
  },
  {
    "temperature": 25
  }
]

Output

Output of this module is JSON body:

{
    "<INPUT_LABEL>": <Processed data>,
    "<MODULE_NAME>Time": timestamp
}

Example

Input data : (23,23,24,24,24,22,21,21,21,23,24,25,25,25)
Window_Size : 3
Assume previous_stable_value = x

If Send_on_change == True then 
Expected Output : (24,21,25)

If Send_on_change == False then
Expected Output :(x,x,x,x,24,24,24,24,21,21,21,21,21,25)