Automated Search for Spectroscopic Exocomet Transits Using HARPS spectra

Run an automated search for exocomet transits from raw HARPS 2D fits files.

Data used in paper: Entire HARPS archive (from 21/03/2022)

List of dependencies

All package dependencies can be found listed in requirements.txt.

How to run

If you have a folder with fits files corresponding to different stars, this is how we created our database:
build-dataset:

1. metadata.py - saves metadata from fits files in fits folder, using GetData.py
2. spectra.py - saves spectra from fits files in fits folder for specified atomic line (CaII H & K lines), using GetData.py. +/- 2000 points from atomic line is saved
3. sanitise_reduce.py - cleans up the name of the observed targets to facilitate grouping (eg. 'reduced' name is all lower case and underscore, dashes, spaces are removed)

groups:

1. grouping.py - group observed targets by 'reduced' name and coordinates
2. create_dataset.py - create dataset folder with all the stars in subfolders with reduced name

Code that we used to generate our vetting plots use the following codes. Note these will need to be adjusted to your own database:
HR:

1. Gaia.py - get Gaia IDs and then query Gaia Catalogue to get information for HR diagram

search:
Both use param.json as parameters for the search.

1. quicksearch.py - searches through dataset for exocomet transits and returns a list of potential candidates
2. iplotter.py - different options (check iplotter.py --help):
   • --star <name> - will show interactive diagnostic plots for this specific star
   • no options - will run through results of quicksearch.py and show interactive diagnostic plots for star satisfying param.json

The interactive diagnostic plots will ask for user input to:

• classify stars - classification results will be visible in results/CandidateReport/:
  • 'candidate' - press y
  • 'not_candidate_but_real' - press n
  • 'not_candidate_but_junk' - press j
  • 'flagged' - press w
  • 'skip' - press space
• print all metadata for the spectra of the star - press d
• print metadata for all the potential detections - press o

After classification done, close the plot and the next star should pop up. At the end you will have the option to review skipped stars and flagged stars.