Welcome to the documentary chronicling Random Walk of Qifeng in the field of cosmology research.
Login to DCC
ssh [email protected]
login to PLAsTiCC
ssh [email protected]
user home depo in DCC
cd /hpc/group/cosmology/qc59/
Conda env for Scarlet2
scarlet2: installed scarlet2 by pip install
scarlet2_git: installed with github
Some useful command:
 pip show scarlet2 | grep Location
Scp:
scp [email protected]:/hpc/group/cosmology/qc59/remake_im_stamp/20148117_star.csv /Users/qifengc/Documents/2_Research.nosync/RomanScarlet2
20148117 10.1320933873459 -44.99021376377029 62675.527
• RA: 0 h 40 m 31.7 s
• Dec: −44° 59′ 24.8″
Jupyter notebooks:
Simulated_SN_example-copy3: worked for Charlotte's data
under RomanScarlet2:
Simulated_SN_exmaple_copy4-onRoman: current work on Roman images, using what works from Charlotte's data
Simulated_SN_example_copy4-onRoman-multi band: multi band working
download_fits: to download fits: use ./download_fits_file.sh in command line
remake_img_stamp/together: previous most updated code for ngmix, including retrieving files.
in dcc, under remake_img_stamp: save_roman_info.py, py code to retrieve file information.
in dcc, Together-correction-for-star.ipynb on dcc/remake_image_stamp to retrieve star catalogues and derive magnitude.
Useful paper:
Questions
[Not started] Why we need to create both scarlet1 and scarlet2 observation object
[In progress] What is the psf data from Roman, and what is the psf from scarlet2/scarlet1
[Not started] The sky will rotate for each epoch, then the psf for the same object at different time is actually different, since psf is position dependent?
[Not started] Charlotte's method does not really take into account different passbands together for one source, right? since all channels from each epoch are split and treated as separate channels.
For Dan:
Check the process of building psf
Background subtraction before and after cutout is that important?
Some left issues
Scarlet2:
- [Done]PSF not from Roman
- PSF not specifying position
- Anisotropic pixel scales: Pixel scales along x and y for Roman images are different, currently just taking an average of the two.