weevenetwork/outlier-detector repository overview

⁠Outlier Detector

• Outlier Detector⁠
  • Description⁠
  • Environment Variables⁠
    • Module Specific⁠
    • Set by the weeve Agent on the edge-node⁠
  • Module Testing⁠
  • Dependencies⁠
  • Input⁠
  • Output⁠

⁠Description

Outlier Detector is a processing module responsible for data sanitization and anomaly detection of data passing through weeve data services. Outlier Detector checks if received data are within constraints associated with some maximum and minimum acceptance value or change rate. If the data are outside of the constraints, the module can either remove them, smooth them or keep them. The filter parameters are optional. If they are not provided, the module will not perform any filtering.

⁠Environment Variables

⁠Module Specific

The following module configurations can be provided in a data service designer section on weeve platform:

⁠Set by the weeve Agent on the edge-node

Other features required for establishing the inter-container communication between modules in a data service are set by weeve agent.

⁠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

bottle
requests

⁠Input

Input to this module is:

• JSON body single object, example:

{
    "temperature": 15,
    "timestamp": 1234567890
}

⁠Output

Output of this module is as follows and depending on cleaning settings.

• JSON body single object, example:

{
    "temperature": 15,
    "timestamp": 1234567890
}