hstaxe
is a Python package to extract spectra from all
slitless spectroscopy modes provided by the Hubble Space
Telescope Wide Field Camera
3 (WFC3)
and the Advanced Camera for Surveys
(ACS). It supports the
features provided by the previous IRAF-based aXe package.
The WFC3 team has recently tested hstaxe
installation and found
three distinct pathways to get the package installed successfully.
Each of these installations were tested with the HSTaXe WFC3
cookbook and were found
to execute all the cookbook steps successfully. NOTE: These installation
pathways were tested on macOS Catalina.
Installing the latest released version of hstaxe
with pip
involves
the following steps (conda
is used to install some non-Python
dependencies):
conda create --name hstaxe_test python=3
conda activate hstaxe_test
conda install gsl cfitsio make automake autoconf libtool pkg-config -y
conda install wcstools -c https://ssb.stsci.edu/astroconda -y
pip install hstaxe
Optionally install Jupyter:
pip install jupyter
The following steps should be followed to install
hstaxe
from using conda
using the Astroconda
channel.
This installation should start by downloading only the
conda_environment.yml file from this repository
followed by creating a conda environment using this file:
conda create --name hstaxe_test --file conda_environment.yml
Activate this environment:
conda activate hstaxe_test
Install HSTaXe from Astroconda:
conda install hstaxe -c https://ssb.stsci.edu/astroconda
These instructions describe the installation of the latest development version of the HSTaXe source code. This workflow involves the the creation of a new conda environment, retrieval of the source code from this repository (https://github.com/spacetelescope/hstaxe), and the installation of the HSTaXe software and its dependencies in that environment.
Create and activate a conda environment where hstaxe
and all its
dependencies will be installed in. We call it hstaxe_test
in this
example:
conda create --name hstaxe_test python=3
conda activate hstaxe_test
Install supporting packages before installing hstaxe
:
conda install gsl cfitsio make automake autoconf libtool pkg-config -y
conda install wcstools -c https://ssb.stsci.edu/astroconda -y
Clone the hstaxe
software from this repository:
git clone https://github.com/spacetelescope/hstaxe.git
Install hstaxe
(and its dependencies):
cd hstaxe
pip install .
The aXe WFC3 cookbook data can be downloaded by cloning this repository: https://github.com/npirzkal/aXe_WFC3_Cookbook
-
Enter the
aXe_WFC3_Cookbook
directory and work throughaXe_WFC3_Cookbook.ipynb
within the conda environment created as part of thehstaxe
installation.
The aXe ACS cookbook and associated data can be downloaded from the following area: https://stsci.box.com/s/eo98zjtyccnoq7z73akfrx94jog3pg7j
There are also a couple simple test scripts in hstaxe/tests that can be run against the example data in the aXe cookbook:
-
run_acs_cookbook.py
: runs basic aXe against ACS data -
run_cookbook.py
: runs basic aXe against WFC3 data from the cookbook -
run_cookbook_part2.py
: runs axedrizzle against WFC3 data from the cookbook
This software was originally developed by the ACS group of the Space Telescope - European Coordinating Facility (ST-ECF). The ST-ECF is a department jointly run by the European Space Agency and the European Southern Observatory. It is located at the ESO headquarters at Garching near Munich. The ST-ECF staff supports the European astronomical community in exploiting the research opportunities provided by the earth-orbiting Hubble Space Telescope.
The developers have included, roughly in order of who has worked on the software:
Norbert Pirzkal, ST-ECF/STScI
Markus Demleitner, ST-ECF
Martin Kuemmel, ST-ECF
Richard Hook, ST-ECF/STScI
Howard Bushouse, STScI
Megan Sosey, STScI
hstaxe
has the following Python requirements:
- Python 3.7 or later
- numpy
- astropy
- stwcs
- stsci.imagestats
- drizzlepac
- photutils < 1.1.0
- drizzle
In addition to the above Python packages, hstaxe
depends on the
following C-based packages to be installed:
These packages must be installed in your environment before you can install hstaxe. They can be compiled and installed locally or installed using pip and/or conda.
You will also need a C compiler suite (e.g., gcc
or clang
) to
build hstaxe
via pip from either
PyPI or from the source distribution. In addition to the above required
packages, the following packages are required to build hstaxe
:
- make
- automake
- autoconf
- libtool
- pkg-config
Note that you may also need to install the
c-blosc package as a dependency
of the tables package on certain
platforms (e.g., this is currently needed for macOS and Windows with
Python 3.9 due to missing tables
wheels on PyPI).