The pylibamazed library for Subaru-PFS project.
pylibamazed is a Python package wrapping numerical algorithms for the analysis of 1D spectroscopic data of astrophysical sources.
- Estimate source category (galaxy, star or QSO)
- Classify source type (sub-classification into the source category, e.g. the spectral type of a star for the star category)
- Estimate redshift
- Provide redshift reliability
- Estimate radial velocity
- Measure fluxes of emission lines
The pylibamazed algorithms are mainly developped in C++ and wrapped in Python. These algorithms are dependant on the thirdparties listed below. It is recommended to install third parties on your system using your own package manager. However, pylibamazed provides a python script to install these thirdparties. To install third parties using pylibamazed internal script, refer to the related third parties section.
drp_1d has the following strict requirements:
Required third parties:
lbfgspp needs to be installed from source Required python packages:
- numpy >=1.25.0
- astropy >=3.1.1
- cython >=0.17.0
- pandas >=1.0.0
- h5py >=2.9
- pytest-mock >=3.11.1
- jsonschema >=4.17.3
- ninja >=1.11
Prerequisites to install pylibamazed from source code:
To install the pylibamazed python module, run :
pip install .
All the building stages are managed by the python package manager. Several options is used to drive the python package deployement or the C++ part building.
The build process uses the pip package manager.
If the thirdparties are not installed in a regular directory, you can specify the path to find thirdparties. If some thirdparties have been installed with the internal pylibamazed script, you must specify the corresponding directory as follows
pip install -v . -C cmake.define.CMAKE_PREFIX_PATH=/your/thirdparties/directory
You can build the library either in Release, Debug or Coverage mode (default to Release). For instance to build pylibamazed in Debug mode, run:
pip install -v . -C cmake.build-type=Debug
To speed up the building time, you can disable test :
pip install -v . -C cmake.define.BUILD_TESTING=OFF
You can specify a directory for C++ building files. Tests are also available in this directory.
pip install -v . -C build-dir=build
For developers. In order to speed up the compilation of the library and the installation of the python module, you must use the following options build-dir=build and --no-build-isolation
These options, requires to mandatory install the following packages in your virtural environement.
pip install scikit-build-core numpy setuptools_scm ninja
Then, you can run the following command to build the library and install the python module
pip install -v -C build-dir=build --no-build-isolation
To test the C++ part, in pylibamazed root directory, run:
pip install -v . -C build-dir=build
cd build
ninja test
To test the python part, in pylibamazed root directory, run:
pip install -v . -C build-dir=build
pytest
The python API documentation could be generated as follows:
Build the documentation:
cd $ROOT_DIR/pylibamazed/doc
make html
Then open in your web browser:
$ROOT_DIR/pylibamazed/build/html/index.html
As stated earlier pylibamazed depends on several third parties (refer to this section for the complete list). It is recommended to install third parties on your system using your own package manager. However, pylibamazed provides a python script to install theses thirdparties.
On ubuntu :
sudo apt install -y \
libboost-dev libboost-filesystem-dev libboost-thread-dev libboost-timer-dev libboost-program-options-dev libboost-test-dev \
libgsl-dev \
libfftw3-dev \
libopenblas-dev \
libeigen3-dev \lbfgspp needs to be installed from source.
buildandinstallthirdparty.py [-h] [--workdir WORKDIR] [--prefix PREFIX] [-j PARALLEL] [--extra_flags EXTRA_FLAGS] [--force] [name1 ...]
Name argument corresponds to the third party name and could take the following values: [boost | gsl | fftw | openblas | eigen | lbfgspp].
For instance, to install the fftw third partie into the thirdparty directory, execute:
python tools/buildandinstallthirdparty.py fftw
Other command line options:
--workdir: specifies the working directory for the third party building (absolute path)
python tools/buildandinstallthirdparty.py fftw --workdir=/tmp
--prefix: specifies the installation directory for third parties (absolute path)
python tools/buildandinstallthirdparty.py fftw --prefix=/usr/local
-j: specifies the number of make jobs to run simultaneously
python tools/buildandinstallthirdparty.py fftw -j 4
--extra_flags: specifies extra_flag to give to the build stage of third party
python tools/buildandinstallthirdparty.py fftw --extra_flags=
--force: forces the library building and overwrites existing built library
python tools/buildandinstallthirdparty.py fftw --force
In order to launch tests and see python code coverage, in pylibamazed folder
coverage run --source=pylibamazed -m pytest
coverage report
coverage html
Drag and drop the created index.js in your web navigator
To create a wheel, you need to previously install build package.
pip install build
Then to create a wheel, run the command :
python -m build -C build-dir=build -C cmake.define.BUILD_TESTING=OFF -C cmake.define.CMAKE_PREFIX_PATH=/your/thirdparties/directory
The wheel is available in the dist directory.
Please send your bug reports or questions to amazed-support@lam.fr
Copyright Aix Marseille Univ, CNRS, CNES, LAM/CeSAM
This software is a computer program whose purpose is to estimate the spectrocopic redshift of astronomical sources (galaxy/quasar/star) from there 1D spectrum.
This software is governed by the CeCILL-C license under French law and abiding by the rules of distribution of free software. You can use, modify and/ or redistribute the software under the terms of the CeCILL-C license as circulated by CEA, CNRS and INRIA at the following URL "http://www.cecill.info".
As a counterpart to the access to the source code and rights to copy, modify and redistribute granted by the license, users are provided only with a limited warranty and the software's author, the holder of the economic rights, and the successive licensors have only limited liability.
In this respect, the user's attention is drawn to the risks associated with loading, using, modifying and/or developing or reproducing the software by the user in light of its specific status of free software, that may mean that it is complicated to manipulate, and that also therefore means that it is reserved for developers and experienced professionals having in-depth computer knowledge. Users are therefore encouraged to load and test the software's suitability as regards their requirements in conditions enabling the security of their systems and/or data to be ensured and, more generally, to use and operate it in the same conditions as regards security.
The fact that you are presently reading this means that you have had knowledge of the CeCILL-C license and that you accept its terms.