/unagi

Search and download data from Hyper Suprime-Cam (HSC) Subaru Strategic Survey (SSP)

Primary LanguagePythonMIT LicenseMIT

Unagi (ウナギ)


Besides being the most delicious fish in the world and one of the signature Japanese dishes, unagi can also help you navigate through the public or internal data release of the Hyper Suprime-Cam (HSC) Subaru Strategic Program (SSP; also known as the HSC survey) -- an ambitious multi-band deeep photometric survey using the awesome prime-focus camera on the 8.2m Subaru telescope.

Also, unagi does not stand for anything because forced acronym is for psychopath.

If you want to learn more about the real unagi, you can check out this video on Youtube. And here is a vlog about a famous unagi place I have tried: Obana 尾花. It is pretty close to downtown Tokyo and few statioins away from Kavli-IPMU.

If you want to learn more about this amazing food, just try import unagi; unagi.unagi() in your Jupyter Notebook.

Recent Updates

  • 06/03/2019: Support for HSC PDR2 is added (need more tests)
    • The schema of all tables in the PDR2_DUD and PDR2_WIDE reruns are available in unagi/data/ folder as a .json file.
  • 06/04/2019: New features: filters.py and camera.py
    • Help you get important information about HSC camera (e.g. CCD QE, primary mirror reflectivity) and filters (e.g. transmission curves, absolute AB magnitude of Sun in each filter etc.).
  • 06/06/2019: New features: add hsc_check_coverage() method to check if a coordinate is covered by one HSC Patch. See here for example.

Applications

  • Using HSC sky object to characterize residual background

TODO List

  • Directly download data products from HSC pipeline (e.g., coadded Patch or source catalogs)

  • Access to HSC weak-lensing shape catalog

  • Access to HSC random catalog

  • Reproduce the CModel result

Installation

  • python setup.py install or python setup.py develop will do the job.
  • Right now, unagi only supports Python>=3. If you are still using Python 2, you should make the switch.
  • unagi only depends on numpy, scipy, astropy, and matplotlib. All can be installed using pip or conda.

Documents

I promisehope that documents will be available soon...but right now, please take a look at the Jupyter Notebook demos for each functionality.

Acknowledgement

Thanks the HSC collaboration for making this amazing survey happen and make these beautiful data available. Also thank the good people at NAOJ who work tirelessly to prepare the data release.

Reporting bugs

If you notice a bug in unagi (and you will~), please file an detailed issue at:

https://github.com/dr-guangtou/unagi/issues

Requesting features

If you would like to request a new feature, do the same thing.

License

Copyright 2019 Song Huang and contributors.

unagi is free software made available under the MIT License. For details see the LICENSE file.