agnpy - v0.4.0: introduced proton synchrotron

This is the first agnpy version including hadronic radiative processes. @IlariaViale and @dimaniad6 have implemented a ProtonSynchrotron class.

- Jupyter Notebook
Published by cosimoNigro over 3 years ago

agnpy - v0.3.0: introduced proton particles distribution

v0.3.0 introduces proton particles distributions.

In the previous versions of the package, only an electron distribution was available, the latter had to be defined through a dictionary ```python

set the spectrum normalisation (total energy in electrons in this case)

spectrum_norm = 1e48 * u.Unit("erg")

define the spectral function parametrisation through a dictionary

spectrumdict = { "type": "PowerLaw", "parameters": {"p": 2.8, "gammamin": 1e2, "gamma_max": 1e7}, }

set the remaining quantities defining the blob

Rb = 1e16 * u.cm B = 1 * u.G z = Distance(1e27, unit=u.cm).z deltaD = 10 Gamma = 10

blob = Blob(Rb, z, deltaD, Gamma, B, spectrumnorm, spectrumdict) ```

This was not very handy: I have simplified - and broke - the previous blob API. From version 0.3.0 this is how electron distribution are defined and passed to the blob

```python

electron distribution

ne = BrokenPowerLaw( k=1e-8 * u.Unit("cm-3"), p1=1.9, p2=2.6, gammab=1e4, gammamin=10, gammamax=1e6, mass=m_e, )

set the quantities defining the blob

Rb = 1e16 * u.cm z = Distance(1e27, unit=u.cm).z deltaD = 10 Gamma = 10 B = 1 * u.G

blob = Blob(Rb, z, deltaD, Gamma, B, ne=ne) ```

The new way of defining particle distributions also allows to define proton distributions python n_p = PowerLaw(k=0.1 * u.Unit("cm-3"), p=2.3, gamma_min=10, gamma_max=1e6, mass=m_p) blob = Blob(R_b, z, delta_D, Gamma, B, n_e=n_e, n_p=n_p) The only difference between particle distributions is the mass argument. This will eventually allow to describe other particle distributions.

To achieve this:

ElectronDistribution has been changed to ParticleDistribution. This is the base class from which all the different distributions (PowerLaw, BrokenPowerLaw etc..) inherit;
@dimaniad6 additionally implemented an InterpolatedDistribution to read an arbitrary input particle distribution (two arrays: one for of Lorentz factors and the other for densities);
the Blob class has been improved:
- the different initialisation method available under the spectrum_norm_type argument of the Blob are now implemented in the base ParticleDistribution class. I.e. one can initialise from_total_density, from_total_energy_density, from_density_at_gamma_1, from_total_energy;
- several quantities are now provided as properties and are evaluated on the fly from the base parameters ( $Rb$, $z$, $\deltaD$, $\Gamma$, $B$);
the other classes remain largely unchanged. They have just been modified internally to deal with the new definition of particle distributions;
note that the prefactor of the particle distribution has been renamed from k_e to k, as we are not describing exclusively electrons distributions now. This means that the corresponding parameter in the fit classes is log10_k instead of log10_k_e.

- Jupyter Notebook
Published by cosimoNigro over 3 years ago

agnpy - v0.2.0: incorporated sherpa and Gammapy wrappers

In this release, the sherpa and Gammapy wrapper - that before were simple examples in the documentation - have been added to the source code.

A agnpy.fit module has been introduced with two physical scenarios:

SycnhrotronSelfComptonModel, representing the sum of synchrotron and synchrotron self-Compton (SSC) radiation. This scenario is commonly considered to model BL Lac sources;
ExternalComptonModel, representing the sum of synchrotron and synchrotron self-Compton radiation along with an external Compton (EC) component. EC scattering can be computed considering a list of target photon fields. This scenario is commonly considered to model flat spectrum radio quasars (FSRQs).

For each scenario, a sherpa or gammapy backend can be selected, such that the fit can be performed with both packages. Some helper functions are added to load directly a file representing a MWL SED with a proper format into a data object of the two packages.

- Jupyter Notebook
Published by cosimoNigro almost 4 years ago

agnpy - v0.1.8 minor fix in the usage of the default matplotlibrc

In the previous release I forgot to include the agnpy/utils/matplotlibrc in the MANIFEST.in. To load it I am now using improtlib.resources rather than the deprecated Path(__file__).parent, see https://github.com/wimglenn/resources-example.