bernddoser / taco

Tools for Automated Characterisation of Oscillations

Tools for Automated Characterisation of Oscillations (TACO)

The TACO modules will be restructured to be fully pythonic. Please find the heritage bash modules here.

Git usage

It is recommended to use git for downloading the TACO source code

git clone --recurse-submodules https://github.com/HITS-TOS/TACO.git

The dependency sloscillations is integrated as a git submodule and will be available using --recurse-submodules during git clone. If the flag was not used, it can be done afterwards with

git submodule update --init --recursive

Note: As long as the repository is private a personal access token is needed for the authentication.

Tests (mostly not functioning, ignore)

Tests are implemented using pytest and can be executed with

python3 -m pytest

Jupyterlab (functional, but not recommended)

The Jupyterlab docker container provides a comfortable way to perform TACO modules and can by started with

docker build -t taco-jupyterlab -f .devcontainer/Dockerfile-jupyterlab .
docker run -it --rm -p 8888:8888 taco-jupyterlab

Open the printed URL in your browser to access Jupyterlab. The jupyter notebook work/pipeline.ipynb is a good starting point.

Install TACO with conda

Basing on Miniconda, TACO can be installed with

conda env create
conda activate taco

Install TACO with conda

Download and install the packages as per the requirements.txt file.

Running high-throughput pipeline

For processing a long list of stars the high-throughput pipeline is available. Before running the pipline, please execute

export PATH=$PWD/src:$PATH
export PYTHONPATH=$PWD/src:$PWD/libs/sloscillations:$PYTHONPATH

once from the TACO root directory. Then the high-troughput pipline can be started with

pipeline.py -i <input directory> -s <settings file>

taking every <name>.dat file in the input directory and write the results in a directory <name>. A settings file with all entries is available at pipeline/pipeline_settings_full.yaml.

Tip

Copy the settings-file into a result directory and executing the pipline from there, leaves the run parameters documented.

Tested operation system architectures

TACO docker-jupyterlab was tested on:

• Linux (Ubuntu and CentOS)
• MacOS (M1-Chip) (please consider HITS-TOS#25)
• Windows 11, Docker engine 4.15.0 using WSL 2

TACO conda high-throughput pipeline was tested on:

• Linux (CentOS)