search

LSSTDESC / DC2-production

Configuration, production, validation specifications and tools for the DC2 Data Set.
BSD 3-Clause "New" or "Revised" License
11 stars 7 forks source link

Sprinkler Examples #68

Closed dannygoldstein closed 6 years ago

dannygoldstein commented 6 years ago

Subtask of #56.

On a sprinkler-focused telecon Friday 12/15/17 I promised @rbiswas4, @jbkalmbach, and @linan7788626 small examples of glSN+host+lens catalogs that could be used for debugging purposes.

I am diving into catalog production now and should have examples for you soon.

dannygoldstein commented 6 years ago

Hi all (esp. @linan7788626, @rbiswas4, @jbkalmbach),

I'm happy to announce the release of the first set of DC2 sprinkler examples. Please use these to develop your downstream codes.

What I have produced is a catalog of lensed images, lenses, and host galaxies of six strongly gravitationally lensed Type Ia supernovae (glSNe; 3 doubles, 3 quads). Each row in the catalog corresponds to one image of a glSN. I used the new protoDC2 cosmological simulation to supply host galaxies for these objects. protoDC2 uses a bulge+disk galaxy model via galacticus, and I have provided information about both components in the catalog. I used the Hsiao template to model the time-evolving SN Ia SED. The host galaxy luminosities obey M_AB = -2.5 log L, where M_AB is the absolute magnitude in the AB system.

All coordinates are given with respect to the centroid of the lens (i.e., the lens is at x=0, y=0).

Here are the columns in this test catalog and their definitions:

The example catalog is available here as a pickled pandas dataframe:

http://w.astro.berkeley.edu/~dgold/static/glsn_protodc2.pkl

To read it into python do:

import pickle
df = pickle.load(open('glsn_protodc2.pkl','rb'))

Please let me know if you have any questions.

Thanks, Danny

dannygoldstein commented 6 years ago

Also many thanks to @dkorytov and @yymao for helping me do this over the holidays.

Danny

rbiswas4 commented 6 years ago

Thanks @dannygoldstein . I have tried to do anything with it yet, but will try as soon as I am back.

linan7788626 commented 6 years ago

Hi @dannygoldstein , thanks for the catalog, I played with it in past days, it looks good. One thing I want to mention is that the names of SEDs of the components of the hosts would be needed for the further work as well. On the other hand, could you please provide a catalog of the sources only? For example in current data, you have six systems, what I need are the properties of these six unlensed sources and the corresponding lenses (6 rows) instead of (17 rows), the shrinked catalog is like the list below. To match the future results from others, we would better set up an identical SystemID for each lensing system at the starting point (i.e., this catalog). Thanks a lot. By the way, my flight is tomorrow, but I will work on this next week. Please let me know if you have any questions. Cheers, Nan.

Sources (Host Galaxies)

Lens galaxies

dannygoldstein commented 6 years ago

@linan7788626, here is the toy catalog with the sources only, containing all the columns you requested except the galaxy component SEDs:

http://w.astro.berkeley.edu/~dgold/static/system_catalog_protodc2.pkl

I am not sure exactly how to obtain the protoDC2 galaxy component SEDs. I know the protoDC2 galaxies have been generated by Galacticus - perhaps @abensonca, @dkorytov, or @yymao could comment on how we could access the protoDC2 galaxy SEDs.

yymao commented 6 years ago

@dannygoldstein @linan7788626 Here's an example of getting all SED info from protoDC2. You may want to further filter sed_labels to only get the relevant info you needed. 

gc = GCRCatalogs.load_catalog('protoDC2')
sed_labels = [q for q in gc.list_all_quantities(True) if q.startswith('SEDs/')]
print('\n'.join(sorted(sed_labels)))
seds = gc.get_quantities(sed_labels) # loads all SEDs
dannygoldstein commented 6 years ago

@yymao

In the SEDs quantities I see many items of the form

I am interpreting this to mean: luminosity of the disk component of the galaxy between 1933 and 1933+474=2407 A in the galaxy's rest frame.

What are the units on these quantities? Are they specific or integrated luminosities?

Thanks, Danny

yymao commented 6 years ago

@dannygoldstein Your interpretation is correct. The unit is 4.4659e13 W/Hz (AB luminosity).

I think they are integrated but I'm not sure on this one --- pinging @evevkovacs @abensonca @dkorytov

abensonca commented 6 years ago

As @yymao said, the luminosities are in AB luminosity units (what sometimes gets called "maggies"). The reason for this is that usually these luminosities are being computed for broad-band filters which we want to compare to data that is given in magnitudes. Outputting in maggies then means we can simply do M = -2.5 log L (where M is magnitude, and L is the output luminosity) - i.e the zero point is already taken care of.

Because of this, the luminosities are essentially integrated averages over the filter. Specifically, the luminosity is found by integrating the SED multiplied by the filter response curve over wavelength, and then this is zero-pointed by dividing by the integral of the AB reference luminosity multiplied by the filter response curve integrated over wavelength.

In the case of these SED outputs the filter response curve is just a constant value between the two wavelengths, and zero outside (i.e. a top-hat filter).

jbkalmbach commented 6 years ago

Thanks @dannygoldstein. I'm back from holiday and looking through the catalog. One question. For a given image you mention the following:

rbiswas4 commented 6 years ago

@dannygoldstein Now that we are all back, can we (atleast you @jbkalmbach @danielsf to find a time more easily, but others can join if interested) briefly meet to discuss the sprinkler? There are a few details we would like to understand :

linan7788626 commented 6 years ago

@dannygoldstein @yymao Thanks for the catalog and the example to extract SEDs. I was just back yesterday and am working on the project now. Cheers.

jbkalmbach commented 6 years ago

Ok, I think @dannygoldstein answered our questions in our conversation this morning and we can close this issue. What do you think @rbiswas4 ?

jbkalmbach commented 6 years ago

This was closed during our discussion last week.