Step 0: make a docker image compatible with the corresponding version of legacypipe
Step 1: seed_formation: get all the seed necessary to generate input randoms
Step 2: bricklist_generation: generate a list of bricks that needs to be processed
Step 3: radec_generation: generate all the radec region that will be made randoms on
Step 4: random_generation: generate the randoms in regions confined by the previous step
Step 5: random_division: divide the randoms into each brick
Step 6: obiwan_run: run obiwan with the input randoms and a list of bricks | | --- meta: name_for_run; name_for_randoms metadata | | --- run_one_brick: run one test brick to see if everything works OK | | ---terminal_debug_run: run the obiwan code on terminal, for pdb debugging | | --- mpi4py_run: run a lot of bricks with multiple nodes | | --- brick_condition: | --- list of finished, unfinished bricks for each name_for_run | --- brickstat.py: update finished, unfinished bricklist
directory needed for each production run: name_for_run=XXXX make a obiwan_out=$obiwan_run/$name_for_run/ $obiwan_out/run/name_for_run/output #for obiwan output $obiwan_out/run/name_for_run/divided_randoms #brick-[brickname].fits $obiwan_out/run/name_for_run/randoms_chunk #whole chunk of randoms $obiwan_code/brickstat/$name_for_run