Closed vaneyken closed 10 years ago
Neither is very descriptive. And maybe all our environment variables should start with "MKID_" .
Hm - agreed... Maybe this is an opportunity to define some better names. Something like:
MKID_RAWPATH MKID_PROCPATH
? (For raw data and processed data directories, respectively.)
Julian
On Jul 29, 2014, at 1:36 PM, Matthew Strader notifications@github.com wrote:
Neither is very descriptive. And maybe all our environment variables should start with "MKID_" .
— Reply to this email directly or view it on GitHub.
Okay, I've now fixed and standardised the main environment variable names used in the pipeline. That means next time you pull, your code will probably break if you're running on a local machine. To fix, just edit your .cshrc/.bashrc or whatever to use the new environment variable names. (You should fix this in any turk accounts too, although in that case it should default to the correct values when it's unable to find the appropriate environment variables anyway).
The new names, along with their values for turk, are:
MKID_RAW_PATH = "/ScienceData" (root of the raw data directory structure)
MKID_PROC_PATH = "/Scratch" (root of the processed data directory structure)
MKID_BEAMMAP_PATH = "/ScienceData/sci4alpha/sci4_beammap_palomar.h5" (default beammap file, only used by h5quicklook.py)
Any future environment variables that people set up should also use the prefix 'MKID_', for the sake of consistency, and so that it's easy to find them all when searching the environment variable list.
Not sure how this hasn't caused problems already, but seems like, for the intermediate data path the pipeline is looking in for system variable 'INTERM_PATH' in half the cases, and 'INTERM_DIR' in the others.... Def. needs fixing! Should be a quick fix, which I can do. Any preferences as to one or the other?