Well THAT'S messy

for file in ../connections_2014-10-7-1*; do conn="-c ../connections_"`echo $file | cut -d "_" -f 2`; types="-t ../neuron_types_"`echo $file | cut -d "_" -f 2`; locs="-l ../locations_"`echo $file | cut -d "_" -f 2`; ./pickle_to_json.py $conn $types $locs; done

For all the connection files that were generated today, create three variables called “conn”, “types”, and “locs” that have a command line switch path in them generated from a fixed prefix and a cut from the connection file name. Then invoke the script “pickle_to_json.py” with those variables as arguments.

Effectively, the connection, neuron type, and location files are all related by their date, so this makes a single JSON file out of the multiple files. I just didn’t want to run pickle_to_json.py a bunch of times by hand, as that seemed error-prone.