Pbh

[Kerianne28]

This branch implements everything needed to inject primordial black holes (PBHs) into the .h5 file that will be used by calc_events.

Synthetic.py Changes

_add_pbh:_ This function takes the .h5 file that is output by perform_pop_syn, creates everything needed for the injected PBHs, then combines the PBH data with the data from the previous .h5 file into a new .h5 file.

This function does the following:

• Reads in the .h5 file without PBHs. Maximum and minimum l and b values are determined for masking the data later.

• Uses an NFW profile to determine how many PBHs there are within 8.3kpc of the galactic center.

• Populates entire sphere with PBHs depending on the masses and fraction dark matter given.

• Determines radii values for each PBH.

• Converts radii from spherical galactocentric to galactic coordinates (and assigns longitude and latitude values).

• Masks the data, giving you only the PBHs in the field of view you need.

• Calculates the Maxwellian velocity distribution for each radius.

• Randomly samples one value, and assigns it as the rms velocity (then converted to vx, vy, and vz).

• All other values needed are assigned to each PBH: magnitudes, pop_id, etc.

• A new .h5 file is created that consists of all of the data from the original .h5 file (without PBHs. From perform_pop_syn), combined with all of the new PBH information for the current field.

_generate_pbh_config_file:_

• Generates a config file with the parameters that are needed for the add_pbh function.

Note: I am not actually sure if this config file is needed. I followed the layout of other PopSyCLE functions in which some of the parameters that are set when you call the function, are the same as the parameters in the config.yaml file. It may be easier to only have a config file and not put the parameters in the call to the function, or vice versa. For now, I followed the style and created the .yaml file and put the parameters in the call to the function as well.

Run.py changes

• Created an optional add_argument statement for the PBH config file.

• Added a segment that instructs what to do if a pbh_config file is given.

Other files added

_pbhconfig.yaml:

• Has all of the default parameters needed for the add_pbh function.

_radial_velocity_dispersiondigitized.csv:_

• Stores data points that are needed in the PBH velocity determination stage. It is a list of radii and dispersion points that are interpolated for all of the PBH radii, to assign each their own dispersion value. This dispersion value then is used to get f(v), which is used to generate the Maxwellian distribution in which the PBH velocity is drawn from.

[MichaelMedford]

Additional comments that we just addressed via video chat:

• Master has to be pulled into the pbh branch before proceeding
• pbh_config_filename needs to be added to arguments of generate_slurm_scripts, and included in the optional parameters section to be passed to the slurm_template

[jluastro]

Actually, the way it is currently implemented is slightly faster. But I would recommend documenting that step a little more.

I also just thought of this: do all hdf5 implementations preserve key order? If not, then Michael’s suggested implementation is preferred or perhaps instead just do:

key_list.remove(‘lat_bin_edges’) key_list.remove(‘lon_bin_edges’)

This will throw an error if they don’t exist; but you can catch it if needed.

On Feb 14, 2020, at 2:12 PM, Michael Medford notifications@github.com wrote:

@MichaelMedford commented on this pull request.

In popsycle/synthetic.py https://github.com/jluastro/PopSyCLE/pull/15#discussion_r379668253:

• #########

• Set random seed

• np.random.seed(seed)
• t0 = time.time()

• Define parameters for NFW profile calculations

• r_s = r_vir/c #kpc, scale radius
• g = 4.3*(10*-3) #(pckm^2)/(Msun*s^2)
• h = 70 #km/(s*Mpc)

• Read in the hdf5 file that doesn't have PBHs. Product of perform_pop_syn.

• no_pbh_hdf5_file = h5py.File(hdf5_file)
• key_list = list(no_pbh_hdf5_file)
• key_list = key_list[:len(key_list)-2] Is this meant to remove the 'lat_bin_edges' and 'lon_bin_edges' keys? If so, then I would rather see something more explicit, like: key_list = [key for key in key_list if 'bin_edges' not in key]

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/jluastro/PopSyCLE/pull/15?email_source=notifications&email_token=AALEJQFZ35BL7QCASXHYC4TRC4JMDA5CNFSM4KQDGB22YY3PNVWWK3TUL52HS4DFWFIHK3DMKJSXC5LFON2FEZLWNFSXPKTDN5WW2ZLOORPWSZGOCVU3JCI#pullrequestreview-359249033, or unsubscribe https://github.com/notifications/unsubscribe-auth/AALEJQGSKC5IGR2MTTX4Q3TRC4JMDANCNFSM4KQDGB2Q.

[MichaelMedford]

I would be surprised if the key order weren't preserved given that the h5py is wrapped around hdf5, that is allocating memory in an order. But I couldn't find documentation on it. My suggestion was motivated by wanting to (1) guarantee correct behavior and (2) improve readability.

On Tue, Feb 18, 2020, 8:39 AM Jessica Lu notifications@github.com wrote:

Actually, the way it is currently implemented is slightly faster. But I would recommend documenting that step a little more.

I also just thought of this: do all hdf5 implementations preserve key order? If not, then Michael’s suggested implementation is preferred or perhaps instead just do:

key_list.remove(‘lat_bin_edges’) key_list.remove(‘lon_bin_edges’)

This will throw an error if they don’t exist; but you can catch it if needed.

On Feb 14, 2020, at 2:12 PM, Michael Medford notifications@github.com wrote:

@MichaelMedford commented on this pull request.

In popsycle/synthetic.py < https://github.com/jluastro/PopSyCLE/pull/15#discussion_r379668253>:

• #########

• Set random seed

• np.random.seed(seed)
• t0 = time.time()

• Define parameters for NFW profile calculations

• r_s = r_vir/c #kpc, scale radius
• g = 4.3*(10*-3) #(pckm^2)/(Msun*s^2)
• h = 70 #km/(s*Mpc)

• Read in the hdf5 file that doesn't have PBHs. Product of

  perform_pop_syn.

• no_pbh_hdf5_file = h5py.File(hdf5_file)
• key_list = list(no_pbh_hdf5_file)
• key_list = key_list[:len(key_list)-2] Is this meant to remove the 'lat_bin_edges' and 'lon_bin_edges' keys? If so, then I would rather see something more explicit, like: key_list = [key for key in key_list if 'bin_edges' not in key]

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub < https://github.com/jluastro/PopSyCLE/pull/15?email_source=notifications&email_token=AALEJQFZ35BL7QCASXHYC4TRC4JMDA5CNFSM4KQDGB22YY3PNVWWK3TUL52HS4DFWFIHK3DMKJSXC5LFON2FEZLWNFSXPKTDN5WW2ZLOORPWSZGOCVU3JCI#pullrequestreview-359249033>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AALEJQGSKC5IGR2MTTX4Q3TRC4JMDANCNFSM4KQDGB2Q .

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/jluastro/PopSyCLE/pull/15?email_source=notifications&email_token=AC3R7JEG7MKI6FKWSN7ET5LRDQFMBA5CNFSM4KQDGB22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEMCV4WY#issuecomment-587554395, or unsubscribe https://github.com/notifications/unsubscribe-auth/AC3R7JG5SLG6MSYEU54A2I3RDQFMBANCNFSM4KQDGB2Q .