2. Special Installation Instructions for SV

2.1. Install Anaconda

You can skip to the next section if you already have Anaconda (or miniconda, etc) installed.

You can use the instructions in Section 2.1 on the link below. The current GHOST software requires DRAGONS 3.0, so once you are done with the Anaconda installation, come back here to complete the DRAGONS and GHOSTDR installation.

Anaconda installation instructions (Section 2.1):

https://dragons.readthedocs.io/projects/recipe-system-users-manual/en/v3.1.0/install.html

2.2. Install DRAGONS

The current version of GHOSTDR requires DRAGONS v3.0. It is incompatible with the latest 3.1 release.

If you already have an environment set up with any version of DRAGONS 3.0, you can skip this step and go to the next section. However, note that you will modify that environment to add GHOSTDR and a new GeminiCalMgr, and an extra dependency.

It is recommended that you instead create a new GHOST-specific environment.

conda config --add channels conda-forge
conda config --add channels http://astroconda.gemini.edu/public
conda config --remove channels http://ssb.stsci.edu/astroconda

conda create -n ghost-sv python=3.7 dragons=3.0.4 ds9
conda activate ghost-sv

2.3. Install GHOSTDR and Dependencies

Here we install the development software needed to reduce GHOST data. We do not have conda packages for it yet.

# Install GHOSTDR
pip install git+https://github.com/GeminiDRSoftware/GHOSTDR.git@v1.1.0

# Install the GHOST-compatible calibration manager
curl -O https://raw.githubusercontent.com/GeminiDRSoftware/GHOSTDR/master/.jenkins/local_calibration_manager/GeminiCalMgr-1.0.2-py3-none-any.whl
pip install --force-reinstall GeminiCalMgr-1.0.2-py3-none-any.whl

# Install an extra dependency
pip install pysynphot

2.4. Additional Software

Because the data in this tutorial were obtained during commissioning, they are identified as “engineering” data. DRAGONS refuses to use such data, as a safeguard. To use the data anyway, we need to modify the program ID and make the data look non-engineering. We have a script to do that. We will use it later.

It is unclear at this time if this will be applicable to the SV data.

cd <path>/ghost_tutorial/
curl -O https://raw.githubusercontent.com/GeminiDRSoftware/GHOSTDR/master/scripts/fixprogid.py

2.5. About the state of the software

This is very much still under-development software. A lot of clean up remains to be done, to the logs, to the file naming conventions, to the code itself. It will produce scientifically correct products, as long as the input data is in the expected format.

This manual has only one example, one target in standard resolution. Depending on the data, you might need to make slight modifications, eg. if the flats and the arcs are taken with different read mode. However, when it comes to reducing two-target observations, or high resolution observations, the steps are expected be the same. The file structure of the output will be different, with an extra source or just one with no sky, etc.

The final science exposures (and the standard files) are NOT stacked. The software was delivered without stacking. So if you obtain three red exposures to increase the signal-to-noise, you will need to stack the final calibrated files yourself. The DRAGONS team will be looking into adding stacking as we do the clean-up. The generic stackFrames primitive appears to be doing a reasonable job for now.

The software has not yet been tested on very faint stars. Since having enough flux in the slit view image is critical for the extraction, very faint source hardly visible on the slit view images might be a problem. We have not yet determine how faint we can go.