Synthetic photo #28
Synthetic photo #28
Conversation
djperrefort
left a comment
djperrefort
left a comment
There was a problem hiding this comment.
Overall this is great progress and looks really good. The missing transmission filter and the use of the AB offset may be the source of the problem. I also included general comments I thought you might find useful.
| import sndata | ||
| from sndata.csp import dr1, dr3 | ||
| from .. import lc_colors | ||
| from .. import spectra_chisq |
There was a problem hiding this comment.
You have multiple unused imports here that can be removed. You only need
import numpy as np
import sncosmo
import sndata
from astropy import constants as const
from astropy import units as u
from astropy.table import QTable, Table, join, unique, vstack
from sndata.csp import dr1, dr3
from .. import lc_colors
| obj_id = spec_data.meta['obj_id'] | ||
|
|
||
| # Add appropriate units | ||
| spec_data['wavelength'] *= u.angstrom |
There was a problem hiding this comment.
In general it is safer to write this as
if not spec_data['wavelength'].unit:
spec_data["wavelength"].unit = u.angstrom
| @@ -0,0 +1,196 @@ | |||
| """Create table with synthetic magnitudes and regressed magnitudes. | |||
There was a problem hiding this comment.
As a general note, I would suggest taking a quick look at PEP8 which is the formal python style guide. There are various tools and IDE's that can do this automatically for you (I find that doing it manually is not realistic for most projects.)
| """ | ||
|
|
||
| # Create quantity table to correctly handle squaring units | ||
| # http://docs.astropy.org/en/latest/table/mixin_columns.html#quantity-and-qtable |
There was a problem hiding this comment.
+1 for the inclusion of a link that describes the issue!
| gauss_process = lc_colors._lc_regression.fit_gaussian_process(data=photo_data) | ||
| # Predict photometric data at the spectra times for specific band | ||
| # Units of Jy? | ||
| pred_flux, pred_flux_err = lc_colors._lc_prediction.predict_band_flux(gp=gauss_process, band_name=band, times=spec_times) |
There was a problem hiding this comment.
lc_colors._lc_prediction.predict_band_flux -> lc_colors.predict_band_flux
| return out_table | ||
|
|
||
|
|
||
| def make_table(): |
There was a problem hiding this comment.
This function looks like a duplicate of the run_phot_comp.make_table function.
run_photo_comp.py
Outdated
| # TODO display filename at prompt? | ||
| ans = input('Overwrite joint table? (y/n)') | ||
| if ans == 'y': | ||
| tbl.write('../../../../comp.fits', overwrite=True) |
There was a problem hiding this comment.
Relative file paths like this can be tricky. They depend on the directory where you are running the code from, which is not always the same as the file's parent directory. A more robust way is to use
from pathlib import Path
base_dir = Path(__file__).parent
# Use however many parents you need
out_path = base_dir.parent.parent.parent / 'comp.fits'
analysis/spectra_chisq/__init__.py
Outdated
| 'calc_chisq', | ||
| 'tabulate_chisq' | ||
| 'tabulate_chisq', | ||
| #'photometry_to_spectra_time', |
There was a problem hiding this comment.
You don't want these commented out. The __all__ attribute defines what is imported when you run from analysis.spectra_chisq import *. However, importing * is almost always frowned upon. We use it here because the automatic documentation builder requires it to know what functions to include in the online docs.
| flux: F_nu; flux per frequency. Units: Jansky | ||
| """ | ||
|
|
||
| flux = 10**(-0.4 * (mag - zp - offset)) |
There was a problem hiding this comment.
I think this should be flux = 10**(-0.4 * (mag + offset - zp))
djperrefort
force-pushed
the
synthetic_photo
branch
from
e538f5c to
675616b
Compare
|
@spletts I talked with a friend here in Pasadena and she recently was doing something similar. I've pushed a change that borrows her code snippet. Some targets like 2005kc look pretty great, others like 2004ef and 2004dt don't. However, I think the results are very close to what we want and are mostly correct. I would look at plots of a few other targets, pick a few that look good/bad and discuss it with Michael at group meeting. |
|
Calculates synthetic photometry using the technique in Casagrande & VandenBerg 2014. |
djperrefort
force-pushed
the
synthetic_photo
branch
from
d50b483 to
5f73f69
Compare
Makes a table of synthetic magnitudes and regressed magnitudes