Scripts used in clustering analysis of Realfast Candidates

• run_search.py: Runs a standard rfpipe transient search on an SDM and saves information of all the candidates found. It does not perform any clustering. Uses realfast_nocluster.yml to set search parameters. Requires rfpipe.
• inject_on_sim_data.py: Script to inject one simulated FRB on simulated data and save unclustered candidates. It generated simulated data (based on randomly chosen VLA configuration and parameters) and injects an FRB in that data (again with randomly chosen transient parameters). Then it run the standard rfpipe search on it to detect candidates from that transient and saves them without applying any clustering. Uses realfast_nocluster.yml to set search parameters. Requires rfpipe.
• random_hyperparameter_search.py: Runs hyperparameter search for a given algorithm, on a given dataset. It samples the hyperparameters randomly based on the functions in util/hs_utils.py. Requires sklean and hdbscan.
• utils: Contains functions for data processing, plotting, clustering, hyperparameter search and metric calculation.
• notebooks/ncands_vs_snr.ipynb: Notebook to estimate the number of candidates expected from a single pulse search given a DM list, boxcar list and observing configuration parameters. Realfast L-band parameters are currently implemented in the notebook. It uses Sec 2.3 of this (also implemented here) to generate the DM list.