2 years ago

Instalação do código Monte Python e CLASS
[Instalación de código Monte Python y CLASS/Installing Monte Python and CLASS code]

#################Instalação do Anaconda3######################
1- Download do pacote Anaconda 3 [https://www.anaconda.com/]
2- Copie o arquivo para pasta de instalação
Abrir terminal e fazer:
3- $ chmod +x Anaconda[...].sh
4- $ ./Anaconda[...].sh
5- "Enter" para ler as informações
6- yes e "Enter"
7- "Enter" (1 vez apenas)
8- yes e "Enter" (fim de instalação)
9- rm -rf anaconda3[...] (para limpar arquivo depois da instalação)
==============================================================
#PACOTES E BIBLIOTECAS PYTHON NECESSÁRIAS
conda install git
conda install -c anaconda cython
#pip install mpi4py
#pip install mpich
conda install -c conda-forge mpich
(as vezes requer)
#conda install -c anaconda setuptools
#sudo apt-get install python3-distutils --reinstall

(PACOTE NECESSÁRIOS PARA INSTALAÇÃO FORA DO AMBIENTE ANACONDA)
#install pip, git, numpy, scipy, pandas, cython, numexpr
#install libcr-dev mpich, mpi4py
###sudo apt-get install python-dev

#CLASS
git clone https://github.com/lesgourg/class_public.git
cd class_public
#modify Makefile : CCFLAG = -g -fPIC -fno-tree-vectorize (optional)
(class$ abrir makeflie e alterar)
PYTHON=python3
make
test [in python] "import classy"
make ou make all ou make class (make clean: para limpar e repeti o processo)
#./class explanatory.ini (rodar a parte do class)
$ cd python/
python3 setup.py build
python3 setup.py install --user

#As seguintes etapas tem que funcionar
--- import classy
--- import numpy
--- import scipy
--- import cython
--- import pandas
--- import numexpr

#MONTEPYTHON
#git clone https://github.com/baudren/montepython_public.git
git clone https://github.com/brinckmann/montepython_public.git
cd montepyhton_public

cp default.conf.template default.conf
in default.conf modify path root and cosmo
-Exemplo: se o class e monte python estiver na pasta Documents (linux: pwd)
----root = '/home/jaelsson/Documents'
----path['cosmo'] = root+'/class_public_jsl'
----path['clik'] = root+'/planck/code/plc_3.0/plc-3.01/'

#JLA
montepyhton_public$]cd data/JLA
wget http://supernovae.in2p3.fr/sdss_snls_jla/jla_likelihood_v6.tgz
tar -xvf jla_likelihood_v6.tgz
cp jla_likelihood_v6/data/* .
rm -rf jla_likelihood_v6*

cd [...]/montepython_public
#mkdir input chains
input$]cp jla.param lcdm_jla.param

$ mpirun -np 6 python3 montepython/MontePython.py run -p input/lcdm_jla.param -o chains/lcdm_jla -N 20000
$ python3 montepython/MontePython.py info /home/jaelsson/Documents/montepython_public/chains/lcdm_jla --no-mean

1 PASSO: Gerar matriz de covariância:

montepython_version]$ python montepython/MontePython.py info chains/lcdm_jla --no-mean --want-covmat

2 PASSO: Copiar os arquivos "lcdm_jla.bestfit" e "lcdm_jla.covmat" e Salvar os arquivos "lcdm_jla.bestfit" e "lcdm_jla.covmat" nas pastas "$bestfit/" e "$covmat/", respectivamente.

montepython_version]$ mpirun -np 6 python montepython/MontePython.py run -p input/lcdm_jla.param -o chains/lcdm_jla_cov/ -c covmat/lcdm_jla.covmat -b bestfit/lcdm_jla.bestfit -N 2000

python3 montepython/MontePython.py info /home/jaelsson/Documentos/montepython_public/chains/lcdm_jla_cov --no-mean

#PLANCK
open http://pla.esac.esa.int/pla/#cosmology
click in Likelihood
download COM_Likelihood_Data-baseline_R2.00.tar.gz
download COM_Likelihood_Code-v2.0_R2.00.tar.bz2
====
download COM_Likelihood_Data-baseline_R3.00.tar.gz
download COM_Likelihood_Code-v3.0_R3.10.tar.bz2

#COM_Likelihood_Data-extra-camspec-ext_R3.00.tar.gz
#COM_Likelihood_Data-extra-plik-ext_R3.00.tar.gz

mkdir planck
cd planck
tar -xvf [..]/COM_Likelihood_Data-baseline_R3.00.tar.gz
tar -xvf [..]/COM_Likelihood_Code-v3.0_R3.10.tar.bz2
cd plc-3.1

(conda install -c cefca pyfits) / conda install astropy
#(pip install pyfits --user)
sudo apt install gfortran
./waf configure --install_all_deps --lapack_mkl=$MKLROOT (look to the readme)
./waf install

#test planck
source bin/clik_profile.sh
cd [...]/montepython_public
cp base2015.param lcdm_planck.param
python montepython/MontePython.py run -p input/lcdm_planck.param -o chains/lcdm_planck -N 20000

===========================================================================
Links interessantes:

Toda documentação CLASS está disponível em: http://class-code.net/

Getting Started with MontePython Solution to Exercises: http://research.iac.es/congreso/cosmo2017//media/montepython.pdf

BRINCKMANN, T.; LESGOURGUES, J. MontePython 3: boosted MCMC sampler and other features. arXiv e-prints, p. arXiv:1804.07261, abr. 2018. Disponível em: https://ui.adsabs.harvard.edu/abs/2018arXiv180407261B.

Monte Python Documentation: https://readthedocs.org/projects/monte-python/downloads/pdf/latest/

Monte Python -The Monte Carlo code for class in Python: http://baudren.github.io/montepython.html

Courses: https://lesgourg.github.io/courses.html

REDES SOCIAIS
Facebook/Instagram/TikTok/Kwai/Youtube/rumble: jaelssonlima

Loading comments...