This Fortran library provides basic data structures, algorithms and some handy utilities, which are applicable for almost all of intrinsic types.
# fortran_lib should be cloned rather than downloaded as a tar file
# because `git ls-files` is used in the build process
git clone https://github.com/kshramt/fortran_lib
cd fortran_lib
make
Compilation may fail with internal compiler error: Killed (program f951)
if your computer does not have enoguh memory (about 3 GB for GFortran).
*.f90.erb
files are preprocessed by eRuby and cpp
during the compilation.
Preprocessed pure Fortran code is also available at https://github.com/kshramt/fortran_lib_expanded and https://bitbucket.org/kshramt/fortran_lib_expanded.
- Git
- GNU Make 3.82 or newer
- GFortran 4.9 or newer
- Ruby 1.9 or newer
- BLAS and Lapack
GNU version of basic commands (sed
, awk
, getopt
and so on) are assumed to be installed.
For example, script/make_include_make.sh
does not work with BSD-sed.
Please install GNU Coreutils and GNU version of other necessary commands.
These programs estimate the ETAS parameters. Please run
bin/etas_solve.sh --help
for the details.
example/etas_solve/README.md describes output format of etas_solve.exe
.
bin/etas_solve.sh \
--t_normalize_len=1 \
--m_for_K=6 \
--t_pre=0 \
--t_begin=0 \
--t_end=1000 \
--data_file=example/etas_solve/catalog.tsv |
release/bin/etas_solve.exe
These programs calculate the intensity function given data and parameters. Please run
bin/etas_intensity.sh --help
for the details.
These programs calculate log likelihood, gradient and Hessian for given data and successive parameter sets. Please run
bin/etas_log_likelihood.sh --help
for the details.
This script generates a synthetic catalog for the temporal ETAS model.
This module contains types and routines for automatic differentiation up to second-order derivatives with 1, 2 and 5 independent variables.
This module contains almost_equal
.
This program is distributed under the terms of the GNU General Public License version 3.