Skip to content

Beforerr/discontinuitypy

Repository files navigation

DiscontinuityPy

PyPI Pixi Badge pdm-managed

User Guide

This package is designed to identify and analyze discontinuities in time series data.

  1. Finding the discontinuities, see this notebook
    • Corresponding to limited feature extraction / anomaly detection
  2. Calculating the properties of the discontinuities, see this notebook
    • One can use higher time resolution data

For how to use this project as a python library, please see this page.

Installation

pip install discontinuitypy

Getting started

Import the package

from discontinuitypy.utils.basic import *
from discontinuitypy.core import *

Related projects and publications

TODOs

Science part

  • Analysis
    • Contribution of discontinuities to the power spectrum
    • Check Datagap
    • Distribution of |B| over radius
    • Add error bar
    • More accurate way to obtain the properties of the discontinuities
  • Identifaction
    • Ensemble forest?
    • Smoothing is important?
    • Check change point algorithm
  • Features
    • Thickness in N direction
    • Use high resolution data for feature extraction
  • Compare with other methods of identifying IDs
    • Verify with other methods of identifying IDs

Code part

  • Optimization
    • dask for parallel computing instead of modin
    • JAX library for numpy optimization
    • shorten import time
  • Refactor
    • process_candidates to exclude sat_state logics
    • renaming feature layer candidates
  • Kedro
    • Modular pipelines

bugs

  • STEREO B less than zero (after downsampling?)