Merge branch 'main' into 694-fix-corrupted-instrument-synergy-tutorial

thjsal · thjsal · commit f320ba4dd73c · 2026-03-15T23:05:47.000+02:00
diff --git a/changelog.d/687.attribution.rst b/changelog.d/687.attribution.rst
@@ -0,0 +1 @@
+Lucien Mauviard
diff --git a/changelog.d/687.fixed.rst b/changelog.d/687.fixed.rst
@@ -0,0 +1,3 @@
+Removed the old synthesise method from `synthesise.pyx`.
+
+Fixed the Signal synthesize to handle better the background
diff --git a/changelog.d/688.added.rst b/changelog.d/688.added.rst
@@ -0,0 +1 @@
+Add documentation about installing X-PSI in the Jean-Zay supercomputer.
diff --git a/changelog.d/688.attribution.rst b/changelog.d/688.attribution.rst
@@ -0,0 +1 @@
+Pierre Stammler
diff --git a/docs/source/Example_job.rst b/docs/source/Example_job.rst
@@ -3,12 +3,12 @@
 Example job
 ===========
 
-For both jobs you will need these 
+For the three jobs below you will need these 
 `auxiliary files <https://zenodo.org/record/7113931>`_ inside the ``model_data/``
 directory.
 
-Snellius
---------
+Snellius (SURF)
+---------------
 
 The following job script ``job.sh`` is an example job script for analysis on the Snellius system (see :ref:`hpcsystems`).
 
@@ -58,8 +58,8 @@ number of processes to spawn as a flag argument.
 
 Finally, note that only the root process will generate output for inspection.
 
-Helios
-------
+Helios (API)
+------------
 
 For Helios, we can use the following type of job script:
 
@@ -109,4 +109,48 @@ For Helios, we can use the following type of job script:
     #Clean the scratch automatically here.
     #But remember to remove manually in each node, if the main program ends by crashing.
     rm -rf $OUTPUT_FOLDER
-    
+  
+
+Jean-Zay (IDRIS)
+----------------
+
+For Jean-Zay, a script like the following one can be prepared. Just check your working project account first, with ``idrproj``.
+
+.. code-block:: bash
+
+   #!/bin/bash
+   #SBATCH --account=nameproject@cpu
+   #SBATCH --job-name=XPSI3_BBTest
+   #SBATCH --time=20:00:00
+   #SBATCH --partition=cpu_p1
+   #SBATCH --qos=qos_cpu-t3
+   #SBATCH --ntasks=200
+   #SBATCH --ntasks-per-node=40
+   #SBATCH --hint=nomultithread
+   #SBATCH --mail-user=myemail@mailservice.com
+   #SBATCH --mail-type=END,FAIL
+
+   module purge
+   module load miniforge/24.9.0
+   module load intel-all/19.0.4
+   module load gsl/2.5
+
+   export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$WORK/Softwares/MultiNest/MultiNest_v3.12_CMake/multinest/lib
+   export LD_PRELOAD=$MKLROOT/lib/intel64/libmkl_core.so:$MKLROOT/lib/intel64/libmkl_sequential.so
+   
+   conda activate $WORK/conda/xpsi
+   
+   cd $WORK
+   mkdir TestBBXPSIrun/
+   cd TestBBXPSIrun/
+   mkdir examples/
+   cd examples/
+   mkdir examples_modeling_tutorial/
+   cd examples_modeling_tutorial/
+   cp -r $WORK/Softwares/xpsi/examples/examples_modeling_tutorial/* ./
+   mkdir run
+   
+   srun python TestRun_BB.py > out1 2> err1
+   
+   cp -r out1 err1 run $WORK/Softwares/xpsi/examples/examples_modeling_tutorial/.
+
diff --git a/docs/source/HPC_systems.rst b/docs/source/HPC_systems.rst
@@ -224,16 +224,122 @@ If the above works, we can then continue building X-PSI:
    git clone https://github.com/xpsi-group/xpsi.git
    cd xpsi
    CC=$(which cc) pip install .
-
+   
 Batch usage
 ^^^^^^^^^^^
 
 For example job scripts, see the Helios example in :ref:`example_job`.
 
+Jean-Zay (IDRIS)
+----------------
+
+`Jean-Zay <http://www.idris.fr/eng/jean-zay/index.html>`_ is the French national supercomputer.
+
+Installation
+^^^^^^^^^^^^
+
+To install X-PSI in Jean-Zay, you need to check out if you have different projects for your IDRIS account, with the command ``idrproj``. The installation will be done only in your currently working project, so you may have to switch between your projects and redo the complete installation.
+
+For any installation, it is preferable to process in the ``$WORK`` folder of your project, because of the small storage of the ``$HOME``
+
+First prepare your modules:
+
+.. code-block:: bash
+
+   module purge
+   module load miniforge/24.9.0
+   module load cmake/3.21.3
+   module load intel-all/19.0.4
+   module gsl/2.5
+
+We intend here to install X-PSI with Intel compilers. It is preferable to avoid using the recent Cmake versions because of their dependencies with the GCC compiler. 
+
+Now, let's prepare the conda environment for X-PSI, by installing some of the required packages. However, they must be mentioned explicitly:
+
+.. code-block:: bash
+
+   cd $WORK
+   mkdir conda
+   conda create -p $WORK/conda/xpsi python'>=3.9.0' numpy'<2.0.0' cython'~=3.0.11' matplotlib'==3.9.2' scipy wrapt gsl pytest getdist tqdm nestcheck fgivenx astropy'>=5.2,<7.0.0' emcee ultranest 'h5py<3.16.0' cmap   
+
+
+Then point to the Intel compilers. If needed, mention them explicitly:
+
+.. code-block:: bash
+
+   export CC=icc
+   export CXX=icpc
+   export FC=ifort
+   #export CC=/gpfslocalsys/intel/parallel_studio_xe_2019_update4_cluster_edition/compilers_and_libraries_2019.4.243/linux/bin/intel64/icc
+   #export CXX=/gpfslocalsys/intel/parallel_studio_xe_2019_update5_cluster_edition/compilers_and_libraries_2019.4.243/linux/bin/intel64/icpc
+   #export FC=/gpfslocalsys/intel/parallel_studio_xe_2019_update5_cluster_edition/compilers_and_libraries_2019.4.243/linux/bin/intel64/ifort
+
+Then install mpi4py:
+
+.. code-block:: bash
+
+   cd $WORK
+   mkdir Softwares
+   cd Softwares
+   wget https://github.com/mpi4py/mpi4py/releases/download/4.0.3/mpi4py-4.0.3.tar.gz
+   tar zxvf mpi4py-4.0.3.tar.gz
+   cd mpi4py-4.0.3
+   python setup.py build
+   python setup.py install
+
+
+Now that the environment is set, MultiNest can be installed:
+
+.. code-block:: bash
+   
+   cd $WORK/Softwares
+   git clone https://github.com/farhanferoz/MultiNest.git  ./MultiNest
+   cd MultiNest/MultiNest_v3.12_CMake/multinest/
+   mkdir build
+   cd build
+   cmake -DCMAKE_INSTALL_PREFIX=~/pathtosoftwares/MultiNest \
+               -DCMAKE_{C,CXX}_FLAGS="-O3 -xCORE-AVX512 -mkl" \
+               -DCMAKE_Fortran_FLAGS="-O3 -xCORE-AVX512 -mkl" \
+               -DCMAKE_C_COMPILER=mpiicc    \
+               -DCMAKE_CXX_COMPILER=mpiicpc \
+               -DCMAKE_Fortran_COMPILER=mpiifort  ..
+   make
+   ls ../lib
+   
+Then its Python interface:
+
+.. code-block:: bash
+
+   cd $WORK/Softwares
+   git clone https://github.com/JohannesBuchner/PyMultiNest.git ./pymultinest
+   cd pymultinest
+   export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$WORK/Softwares/MultiNest/MultiNest_v3.12_CMake/multinest/lib
+   export LD_PRELOAD=$MKLROOT/lib/intel64/libmkl_core.so:$MKLROOT/lib/intel64/libmkl_sequential.so 
+   python setup.py install
+
+Finally, xpsi can be installed:
+
+.. code-block:: bash
+
+   cd $WORK/Softwares
+   git clone https://github.com/xpsi-group/xpsi.git
+   cd xpsi/
+   LDSHARED="icc -shared" CC=icc pip install .
+   #LDSHARED="/gpfslocalsys/intel/parallel_studio_xe_2019_update4_cluster_edition/compilers_and_libraries_2019.4.243/linux/bin/intel64/icc -shared" CC="/gpfslocalsys/intel/parallel_studio_xe_2019_update4_cluster_edition/compilers_and_libraries_2019.4.243/linux/bin/intel64/icc" pip install .
+
+   # To check your installation
+   cd ../
+   python -c "import xpsi"
+
+Batch usage
+^^^^^^^^^^^
+
+For example job scripts, see the Jean-Zay example in :ref:`example_job`. 
+
 .. _CALMIPsystem:
 
-CALMIP
-------------------------------------
+CALMIP (U. of Toulouse)
+-----------------------
 
 `CALMIP <https://www.calmip.univ-toulouse.fr>`_ is the supercomputer of `Université Fédérale de Toulouse <https://www.univ-toulouse.fr>`_
 
diff --git a/docs/source/Synthetic_data.ipynb b/docs/source/Synthetic_data.ipynb
diff --git a/docs/source/install.rst b/docs/source/install.rst
@@ -428,7 +428,7 @@ already present in the Mac OS:
 
     brew install llvm
 
-Also use homebrew to install ``cmake``, ``gfortran`` (with ``gcc``) and ``open-mpi`.
+Also use homebrew to install ``cmake``, ``gfortran`` (with ``gcc``) and ``open-mpi``.
 
 .. code-block:: bash
 
diff --git a/xpsi/Signal.py b/xpsi/Signal.py
@@ -888,6 +888,10 @@ def synthesise(self,
                     if (format == 'TXT') and (backscal_ratio != 1.0):
                         raise ValueError('TXT format not supported for backscal_ratio != 1.')
 
+                # In the other case, if phase-averaged background was provided, use it
+                elif len( data_BKG.shape ) == 1:
+                    BKG = BKG.sum( axis = 1 )
+
                 # Write expected background
                 kwargs = {'channels':self._data.channels,
                     'counts':BKG,
@@ -911,8 +915,15 @@ def _write(self, format, **kwargs):
         if format == 'TXT':
             self._write_TXT(**kwargs)
         elif format == 'FITS':
-            if len(self._data.phases) > 2:
-                self._write_EVT(**kwargs)
+            # Case of 2D counts
+            if (len(kwargs['counts'].shape) == 2):
+                # If counts is phase-resolved
+                if kwargs['counts'].shape[1] > 1:
+                    self._write_EVT(**kwargs)
+                # If counts is phase-averaged
+                else:
+                    kwargs['counts'] = kwargs['counts'][:,0]
+                    self._write_PHA(**kwargs)
             else:
                 self._write_PHA(**kwargs)
         else:
diff --git a/xpsi/tools/synthesise.pyx b/xpsi/tools/synthesise.pyx

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	+Removed the old synthesise method from `synthesise.pyx`.
	`2`	`+`
	`3`	`+Fixed the Signal synthesize to handle better the background`
Original file line number	Diff line number	Diff line change
`@@ -0,0 +1 @@`
	`1`	`+Add documentation about installing X-PSI in the Jean-Zay supercomputer.`