These are some tools for 'study7'. This will install self-contained versions of python3, casa (V6 optional), and admit. Only tested on Linux.
The master branch of ADMIT is still the classic python2 version that should work with casa 5.x. The python3 branch of ADMIT will work with casa 5.x, and is working with a pip installed version of CASA6.
-
get this source code:
git clone https://github.com/teuben/study7 cd study7 -
bootstrap your OS (ubuntu for now) with system tools we need down the line (useful for AWS or other virgin Ubuntu systems)
./bootstrap_ubuntu -
miniconda3 (or anaconda3) for our controlled python envirment
./install_miniconda3 casa6=1 ./test_alma1 ./test_casa6
This test will download large amounts of data from the archive. control-C it once it starts, as it can take hours on a slow connection. These data are downloaded in the data subdirectory. Remove these, since it can go fast on a good connection.
If you are brave and want to try out CASA6, add the "casa6=1" argument to the install_miniconda3 script. This has not been tested recently.
2b. Development of astroquery (optional)
git clone -b admit https://github.com/teuben/astroquery
pip3 install -e astroquery
-
casa (via QAC)
./install_qac source casa_start.sh ./test_qac1
This test should take a few mins. Note that this casa is (should be) orthogonal to the casa installed in the classic way.
-
admit
./install_admit source admit_start.sh ./test_admit1
This test should take a few mins.
See also https://casa.nrao.edu/casadocs/casa-5.6.0/introduction/casa6-installation-and-usage
An experimental CASA 6 is now available, which you can install via our miniconda3 install via the "casa6=1" optional command line argument.
Within study7 we have experimented with several python3 environments:
- miniconda3
- anaconda3
- raw python3 build
- virtual environment (the CASA team recommended way)
The important constraint is that Python 3.6 is needed, as well as libgfortran3.
The first official release of CASA 6, including an all inclusive environment with python in a tar file, is planned for July 2019. This will be based on the same science and functionality as the upcoming CASA 5.6 to be used in ALMA Cycle 7 processing. CASA 6.1 is a more full traditional environment, and will be available June/July 2020.
This install worked for me in Ubuntu 18, because they happen to overlap version of the shared libraries with the CASA6 developers. In Ubuntu 20 this failed for me because they got updated (now libgfortran3 became libgfortran5 and libtinfo5 became libtinfo6). I happen to have both mounted, so this did the trick for me (YMMV):
sudo cp -a /a6/usr/lib/x86_64-linux-gnu/libgfortran.so.3* /usr/lib/x86_64-linux-gnu/
sudo cp -a /a6/lib/x86_64-linux-gnu/libtinfo.so.5* /lib/x86_64-linux-gnu
The command
ipython profile create casa6
will create a casa6 profile, so you can simplify calling casa using
ipython --profile=casa6
where files in ~/.ipython/profile_casa6/startup are executed before the command prompt appears. This is where you can place some conventient import commands, to make casa6 look like casa.
Mine are in ~/.ipython/profile_casa6/startup/10-import.py
import numpy as np
import os
import sys
and ~/.ipython/profile_casa6/startup/60-casa.py
Best is to look at the redo_* scripts, and pick your poison. Each, for sanity, will work within a fresh admitNNN directory. The wgetc script referenced is useful to keep a cache, otherwise replace it with wgetc
- redo_2 - classic admit w/ casa5
- redo_35 - new admit3 w/ casa6
- redo_36 - new admit3 w/ anaconda3
- redo_360 - new admit3 w/ python 3.6.10 and casa60
- redo_361 - new admit3 w/ python 3.6.10 and casa61
- study7 (this repo): https://github.com/teuben/study7
- casa: https://casa.nrao.edu/index.shtml
- admit: https://github.com/astroumd/admit
- astroquery: https://astroquery.readthedocs.io/en/latest/
- astroquery: https://github.com/teuben/astroquery (branch: admit)
- pyvo: https://pyvo.readthedocs.io/en/latest/
- alminer: https://alminer.readthedocs.io