fermiPy/extpipe

Extension Pipeline

The extension pipeline consists of two analysis scripts:

• run-region-analysis : Tests a given source for extension while iteratively adding up to 5 point sources to the inner ROI. The outputs of this step are the extension analysis results and a sequence of up to 5 PS models.
• run-halo-analysis : Using the PS models derived in run-region-analysis test the target source for an extended component that is superimposed on the source.

After running these you will need to aggregate the analysis results using the extpipe-aggregate script.

Creating Analysis Directories

Before running the analysis you will need to create the directory structure. The fermipy-clone-configs script will generate a set of analysis directories for each entry in a source list file. To create directories manually for the two analysis scripts:

$ fermipy-clone-configs --source_list=source_list.yaml --script=run-region-analysis --basedir=v0 config.yaml
$ fermipy-clone-configs --source_list=source_list.yaml --script=run-halo-analysis --basedir=v0 config.yaml

where config.yaml is the baseline configuration and source_list.yaml is a hash of analysis configurations keyed to the analysis subdirectory name.

After generating the analysis directories to send things to the batch farm run fermipy-dispatch with the root analysis directory as its argument:

$ fermipy-dispatch --runscript=run-region-analysis.sh v0

The setup_config.py script can be used to automatically the analysis directories for a matrix of different analysis configurations. To create directories for standard analysis:

$ python scripts/setup_config.py --config=std_psf0123_joint2a --script=run-region-analysis \
haloanalysis/sourcelists/3fgl_srcs_list_glat050.yaml haloanalysis/sourcelists/3fhl_srcs_list_glat050.yaml
$ python scripts/setup_config.py --config=std_psf0123_joint2a --script=run-halo-analysis \
haloanalysis/sourcelists/3fgl_srcs_list_glat050.yaml haloanalysis/sourcelists/3fhl_srcs_list_glat050.yaml

To Create directories for alternative IEMs:

$ python scripts/setup_config.py --config=std_psf0123_joint2a --script=run-region-analysis \
--model=all --script=run-region-analysis haloanalysis/sourcelists/fhes_list.yaml
$ python scripts/setup_config.py --config=std_psf0123_joint2a --script=run-halo-analysis \
--model=all --script=run-region-analysis haloanalysis/sourcelists/fhes_list.yaml

Running the Analysis

First create a conda environment with the releases of fermipy and extpipe that you plan to use:

$ conda create --name extanalysis-vXX python=2.7 numpy scipy matplotlib==2.0.0 astropy pytest pyyaml ipython healpy
$ source activate extanalysis-vXX
$ cd <path to fermipy>
$ python setup.py install
$ cd <path to extpipe>
$ python setup.py install

For testing purposes you can also run the following instead of install:

$ python setup.py develop

Aggregating Analysis Results

After the analysis has completed you need to run the aggregation script to merge the results for each target into a single catalog file. This step of the analysis also selects the best-fit model for each source.

$ extpipe-aggregate <basedir> --batch

where <basedir> is the root analysis directory. After this step completes you can build a single catalog file by running extpipe-build-catalog:

$ extpipe-build-catalog std_psf0123_joint2a_stdmodel/*_cat.fits --output=catalog.fits --ts_threshold=16
$ extpipe-build-catalog std_psf0123_joint2a_stdmodel/*_lnl.fits --output=catalog_lnl.fits --ts_threshold=16