Memory modelling for insect path integration.
It is recommended to use a virtual Python environment,
which can be done by sourcing the env.sh
script:
source env.sh
The quickest way to run an experiment setup is using the command-line interface.
The following will run the unmodified Stone model with parameters specified in
setups/stone-as-is.json
, and output a report:
./cli.py setups/stone-as-is.json --report
Results will also be saved in the results
directory.
For documentation of the CLI options, run:
./cli.py --help
The env.sh
script also creates an IPython kernel for use with Jupyter;
choose the pim
kernel under Kernel -> Change Kernel...
Jupyter's saving of cell outputs can be a nice way to share results, but it easily wreaks havoc on repositories if one does not pay attention. It is a good idea to Edit -> Clear All Outputs before committing a notebook, and a Git hook can be used to help with remembering to do so. Install it like so:
cp git-hooks/pre-commit .git/hooks/pre-commit
If the hook stops the commit, but you explicitly want to commit cell output,
ignore the hook using --no-verify
when committing.