Dependencies:
- numpy (tested on version 1.20.1, but no known compatibility issues)
- multiprocess (tested on version 0.70.12.2 but no known compatibility issues). Needed for parallelizing the likelihood calculation on Macs running Apple M1 processors.
Installation:
- Download edmcmc.py and save it somewhere on your drive temporarily. In this example, I will assume it was saved to the "Downloads" folder on a unix machine (located at
~/Downloads/ - Open a python terminal and run the following commands:
import sys-
print(sys.path)
On my machine, that returns the following output:
['', '/anaconda3/lib/python37.zip', '/anaconda3/lib/python3.7', '/anaconda3/lib/python3.7/lib-dynload', '/anaconda3/lib/python3.7/site-packages', '/anaconda3/lib/python3.7/site-packages/aeosa', '/anaconda3/lib/python3.7/site-packages/mpyfit-0.9.0-py3.7-macosx-10.7-x86_64.egg']This is a list of directories where Python will search for files to import when asked. - Move edmcmc.py into one of the directories listed in the previous step. Since
'/anaconda3/lib/python3.7/site-packages'is where other third-party software is usually kept, I chose to put it in there. This can be done with a GUI file system viewer, or using the following command:mv ~/Downloads/edmcmc.py /anaconda3/lib/python3.7/site-packages/
At this point, edmcmc.py should be installed. Run python, and test it using import edmcmc . If that runs successfully, you can use the online test notebook to fully test the functionality. Note that multi-processing does not work on all machines yet.