skyline.tsfresh_features package

Submodules

skyline.tsfresh_features.autobuild_features_profile_tables module

This script creates any missing Ionosphere z_fp_ and z_ts_ tables that may be missing from the database using the raw data in the IONOSPHERE_DATA_FOLDER if there is any.

Run the script with your relevant paths and Python version for example:

# BACKUP the skyline MySQL database FIRST

PYTHON_MAJOR_VERSION="2.7"
PYTHON_VIRTUALENV_DIR="/opt/python_virtualenv"
PROJECT="skyline-py2712"

cd "${PYTHON_VIRTUALENV_DIR}/projects/${PROJECT}"
source bin/activate
cd /opt/skyline/github/skyline
python${PYTHON_MAJOR_VERSION} skyline/tsfresh_features/autobuild_features_profile_tables.py
deactivate

skyline.tsfresh_features.generate_tsfresh_features module

This script determines what the known tsfresh feature names are and extracts features from a small timeseries sample of utils/data.json

Then the known TSFRESH_FEATURES are compared to the feature names that were determined from the feature extraction of the sample data set. The script reports changes and outputs a new TSFRESH_FEATURES list.

This is very important and Skyline is using the tsfresh feature names in a relational manner and any changes in tsfresh names or additions must be considered appropriately. Further some tsfresh feature name strings contain spaces and commas so they are not suitable to convert into variables without string manipulation, therefore the TSFRESH_FEATURES array is declared and once feature names are declared they should be treated as immutable objects due to the relational nature of feature profiles.

Run the script with for example:

PYTHON_MAJOR_VERSION="3.8"
PYTHON_VIRTUALENV_DIR="/opt/python_virtualenv"
PROJECT="skyline-py383"

cd "${PYTHON_VIRTUALENV_DIR}/projects/${PROJECT}"
source bin/activate
bin/python${PYTHON_MAJOR_VERSION} skyline/tsfresh_features/generate_tsfresh_features.py
deactivate

Module contents

version info