Project containing the first steps on the TDE early photometric classification to be used within FINK, for ZTF and LSST data.
Git clone this directory Create virtual environment, and activate it. For instance:
conda create --name tdes python=3.10
conda activate tdes
In order to use the Rainbow fitting, we need to install the light-curve-python package, for which we previously need to install rust, with the following command:
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
Then, we can clone and pip install the light-curve-python package:
python3 -mpip install 'light-curve[full]'
Then, simply install this project with
pip install -e .
You might need to install other packages such as matplotlib or pandas.
Firstly, you may update the config.py file to match your preference. It is important to add the non-tde data from Fink to the data input directory.
The features can be extracted using the extract_features.py script, where you can manually specify on which data you want to perform the feature extraction. The features should be merged into the features_all.csv file. The script merge_csvs.py can also be used to put all the extracted features in the same csv.
(currently unused) Then, you can use the python script snad_aad.py in the anomaly detection folder, in order to run the SNAD_AAD method on the extracted features. Currently (June 2024), we are using the Random Forest classifier, inside the RF_classifier folder.