position dependent photon field for EM interactions (optimised) by GDMarco · Pull Request #535 · CRPropa/CRPropa3

GDMarco · 2026-01-26T17:53:54Z

General description

This is a fully working version of the CRPropa code with some implementation to allow spatially dependent photon field to be initialised and employed for electromagnetic interactions. A particular example is the introduction of the galactic interstellar field (the density and interaction rate tables are at this repo: ISRF tables.)
In principle, this implementation can be used for any spatial dependent photon field properly build by the user.
To allow the introduction of such spatially variable photon fields in CRPropa, the classe and modules that have been modified arePhotonBackground.*, EMPairProduction.*, EMDoublePairProduction.*, EMInverseComptonScattering.* and EMTripletPairProduction.*. Moreover, the class InteractionRates.* has been implemented to handle the process rates.
An example on how to use the code is contained in the notebooks.

Technical details

The .cc files of PhotonBackground.*,EMPairProduction.*, EMDoublePairProduction.*, EMInverseComptonScattering.* and EMTripletPairProduction.* employ some tools of the #include <filesystem> library, only available since C++17. The proper library is specified in the CMakeLists.txt files, as well as in the header of the .cc files.

The code uses the nanoflann functionalities to look for the closest photon field point in your custom grid and evaluate the density and/or the rate correctly. In this new version, nanoflann is implemented as a header-only library among the ones in the /libs/ directory.

//InteractionRatesIsotropic::InteractionRatesIsotropic(); //InteractionRatesIsotropic::InteractionRatesPositionDependent();

Changing nomenclature (isSpatialDependent -> isPositionDependent).

Changing nomenclature

Change nomenclature.

…n field

…ckground

style & constructor of InteractionRates subclasses

JanNiklasB

Hi,

thanks for your contribution!
I have done a first review of your pull request and noticed the following things.

Please try to not include lines in your pull request that are only altered in white spaces without having any purpose, you often only change tabs to spaces or add new empty lines, this makes it difficult to spot actual changes and leads to difficult to read code.
Reduce the amount of includes you use in your files, any not needed include or already in the header existing includes should not be in the .cpp file again. This is only a style thing since otherwise the code seems a little cluttered.
For the 4 changed module files (EM*.h and EM*.cpp) I only commented in EMDoublePairProduction.* but my comments hold true for all module files:
-> The old behavior of function and classes regarding function calls can not be changed, it needs to be backwards compatible ( We need to at least notify the community first when we remove or change crucial features)
-> Everything needs to be as modular as possible, so if you use a class like the interaction rates in a function you need to only require the base class so the user can create their own versions and hand them over. Those base classes then need everything to distinguish them inside those functions
You require the function splitFilename in many of your classes but this function does not depend on any class member, so it would be better to add this function to Common.h as a standalone
Every function which is not obvious to use or where it is not obvious what it does needs to have a description
You need to add your changes to the change log CHANGELOG.md

Many comments would repeat themself so I usually mention something only 1 or 2 times.

With regards,
Jan-Niklas

src/module/EMDoublePairProduction.cpp

include/crpropa/module/EMDoublePairProduction.h

JanNiklasB · 2026-02-13T13:11:42Z

include/crpropa/module/EMDoublePairProduction.h

+  std::string getInteractionTag() const;
+
+  void initRate(std::string filename, InteractionRatesHomogeneous* intRatesHom);
+  void initRatePositionDependentPhotonField(std::string filepath, InteractionRatesPositionDependent* intRatesPosDep);


A description of initRate and initRatePositionDependentPhotonField is necessary, it is not clear for the average user how to use those functions.

Furthermore, we might need to add a initRate(std::string) overload so the old call behavior is preserverd.

Thanks for the comment! I put a brief description of the init*Rate functions, although the user do not need to handle those directly.

Could you please elaborate further the second part of this comment?

Originally initRate was callable with the signiture initRate(std::string), this still needs to be possible, otherwise existing simulations will fail.

JanNiklasB · 2026-02-13T13:50:39Z

src/module/EMDoublePairProduction.cpp

+  if (!this->photonField->hasPositionDependence()) {
+
+    this->interactionRates = new InteractionRatesHomogeneous("interactionRatesHomogeneous", false);
+    InteractionRatesHomogeneous* intRatesHom = static_cast<InteractionRatesHomogeneous*>(this->interactionRates.get());


With the requested change of initRate in the header file this pointer and the one in line 51 are not required. Furthermore initRate would be used in both cases so the if condition would is then only required inside the initRate function.

Same as before.

src/PhotonBackground.cpp

include/crpropa/PhotonBackground.h

CMakeLists.txt

JanNiklasB · 2026-02-13T17:03:29Z

CMakeLists.txt


 endif(ENABLE_PYTHON AND Python_FOUND)

+target_include_directories(crpropa


Please add nanoflann to crpropa over CRPROPA_EXTRA_INCLUDE and CRPROPA_EXTRA_SOURCES (see other libraries).

I add it to the CRPROPA_EXTRA_INCLUDE, as the other libraries (e.g. kiss). Why also to CRPROPA_EXTRA_SOURCES?

You are right, for nanoflann you only have Includes, so only to CRPROPA_EXTRA_INCLUDE

GDMarco · 2026-03-06T17:29:16Z

Thanks for the comments! I fixed and commented all the issues you raised. I will need to implement it merging the currently accepted PRs.

-> now handled internally and only casted in init*Rates functions

…r to interactions

Changes regarding the PR#535

JanNiklasB

Hey,

sorry it took me a little bit longer.
I found a very small typo in one of the example files and encountered a missing install statement in the CMakeLists.txt which I missed last time (sorry)
Other then that I think everything looks good!

JanNiklasB · 2026-03-31T19:19:58Z

doc/pages/example_notebooks/photon_propagation/galactic_propagation_gammarays_ISRF.ipynb

+    "Acon = 0.1\n",
+    "\n",
+    "# radius of the spherical observer placed at the Earth position (kpc).\n",
+    "Orad = np.tan(Scon) * D \n",


I think Scon should be Acon since Scon is nowhere defined

JanNiklasB · 2026-03-31T19:21:02Z

doc/pages/example_notebooks/photon_propagation/galactic_propagation_gammarays_ISRF.ipynb

+    "ebl = IRB_Gilmore12()\n",
+    "\n",
+    "# position dependent. (Sphere to load only the ISRF nodes close to observer and the source.)\n",
+    "isrf = ISRF_Freudenreich98(Sphere(Vector3d(x, y, z) * kpc, D * kpc + 1 * kpc))"


This is currently failing, but only because of the currently missing photon files so it is fine (fyi @Grindegreen )

JanNiklasB · 2026-03-31T19:25:41Z

CMakeLists.txt

 install(DIRECTORY ${CMAKE_BINARY_DIR}/include/ DESTINATION include FILES_MATCHING PATTERN "*.h")
 install(DIRECTORY ${CMAKE_BINARY_DIR}/data/ DESTINATION share/crpropa/ PATTERN ".git" EXCLUDE)
 install(DIRECTORY libs/kiss/include/ DESTINATION include)



please add install(DIRECTORY libs/nanoflann/include/ DESTINATION include) so the nanoflann header files are included correctly (I missed that last time, sorry)

GDMarco added 30 commits March 5, 2024 16:40

creating InteractionRates class

bd15e91

InteractionRates class in SWIG

4efbaa5

update CMakeLists.txt

1c9085e

correct InteractionRates.cpp #include

0bb1fec

avoid redefinition in InteractionRates.cpp

cfa0371

outOfLine definition in InteractionRates.cpp

dc7b60c

//InteractionRatesIsotropic::InteractionRatesIsotropic(); //InteractionRatesIsotropic::InteractionRatesPositionDependent();

update PhotonBackground.h for space dependent fields

d4cc1c8

update PhotonBackground.cpp for space dependent fields

63f5573

update EMPairProduction.h/.cpp for space dependent fields

bf17391

correct #include for PhotonBackground.h/.cpp

68a1df2

correct #include in src/PhotonBackground.cpp

28dd755

correct src/EMPairProduction.cpp

60d904c

update files for position dependent background photons & EMPP

f88fd02

correct minor errors for position dependent photon field & EMPP

abf4d14

Update PhotonBackground.h

b26d064

Changing nomenclature (isSpatialDependent -> isPositionDependent).

Update PhotonBackground.cpp

bb502eb

Changing nomenclature

Update EMPairProduction.cpp

9067fa8

Change nomenclature.

Update EMPairProduction.h

7cad8ff

Change nomenclature.

update EMInverseComptonScattering.h/.cpp for position dependent photo…

e688abf

…n field

update EMDoublePairProduction.h/.cpp for position dependent photon ba…

d0b21b0

…ckground

update EMTripletPairProduction.h/.cpp for spatial dependent photon field

9e6b95c

Update InteractionRates.h

a5d9696

Update InteractionRates.h

9fbd61a

Update InteractionRates.cpp

f573b6f

style & constructor of InteractionRates subclasses

Update PhotonBackground.h indentation

149a771

style typos & methods in InteractionRates class

715ec1f

fix PhotonBackground.h for merging

01dc99e

change exit with throw std::runtime_error() & style

f501c69

Update CMakeLists.txt C++17

615f18c

Update (back) CMakeLists.txt to C++11 version

aeba158

GDMarco added 11 commits February 2, 2026 12:29

add warnings for not finding posdepradfields v2

7d2a0dd

add --output for test

4f97506

donot test without tables

e185892

donot test without tables in testInt

4bb0728

donot test without tables in testInt v2

910461f

donot test without tables in testInt v3

00b3f41

donot test without tables in testInt v4

d5d04f0

donot test without tables in testInt v5 + EMPP

52ddb8e

donot test without tables in testInt v6

62e7085

donot test without tables in testInt v7 + EMPP v2

49d1d63

EMPP v3

8285960

GDMarco marked this pull request as ready for review February 12, 2026 14:24

JanNiklasB suggested changes Feb 13, 2026

View reviewed changes

GDMarco added 2 commits March 6, 2026 17:55

answer to PR v1

7d46929

changes reported in CHANGELOG.md

d62de4f

JanNiklasB and others added 13 commits March 9, 2026 14:09

removed nanoflann lib since its header only

028002a

removed splitFilename since it was never declared

4705a60

fixed wrong declaration of constructors

a1895df

removed unused variable

98be491

moved splitFilename to Common.cpp

a0107b4

Changed initialization parameter to constant values in class

ba8ae38

added old initRate

e726045

changed spaces to tabs for indentation (see CONTRIBUTE.md)

8407f8e

Merge branch 'master' into PhotonFieldPR

470a603

added more overloads for init*Rates and changed handling von Rates

a432e2b

-> now handled internally and only casted in init*Rates functions

Moved init*Rates to interactionRates classes and added seter and gete…

d9c95bb

…r to interactions

isPositionDependent and surface now defined in all constructors

859c3d1

Merge pull request #1 from JanNiklasB/PhotonFieldPR

9c7777b

Changes regarding the PR#535

JanNiklasB suggested changes Mar 31, 2026

View reviewed changes


		endif(ENABLE_PYTHON AND Python_FOUND)

		target_include_directories(crpropa

Conversation

GDMarco commented Jan 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

General description

Technical details

Uh oh!

JanNiklasB left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

GDMarco commented Mar 6, 2026

Uh oh!

JanNiklasB left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

GDMarco commented Jan 26, 2026 •

edited

Loading