def make_run_list(ev_times, ev_runids, exclude_runs=None):
"""
Make a run list from given event times and run IDs and a list of included
runs. Run start, stop are estimated from first and last event time in each
run, which is biased but there seems to be no better way.
Parameters
----------
ev_times : array-like, shape (nevts,)
Event times in MJD days from the used data sample.
ev_runids : array-like, shape (nevts,)
Event run IDs from the used data sample.
exclude_runs : array-like or None, optional
Run IDs to exclude, for example when samples overlap.
(default: ``None``)
Returns
-------
run_list : list of dicts
Dict with keys similar to a snapshot from [1]_ in JSON format:
[
{"good_tstart": "YYYY-MM-DD HH:MM:SS",
"good_tstop": "YYYY-MM-DD HH:MM:SS",
"run": 123456},
{...}, ..., {...}
]
Times are given in iso formatted strings and run IDs as integers.
References
----------
.. [1] live.icecube.wisc.edu/snapshots
"""
# If selected runs were empty on final level, they are not considered here
used_run_ids = np.unique(ev_runids).astype(int)
ev_runids = ev_runids.astype(int)
if exclude_runs is not None:
used_run_ids = filter(lambda runid: runid not in exclude_runs,
used_run_ids)
print(" Exluded runs: {}".format(arr2str(exclude_runs, fmt="{:d}")))
run_list = []
livetimes = np.zeros(len(used_run_ids), dtype=float)
for i, runid in enumerate(used_run_ids):
ev_mask = (ev_runids == runid)
ev_t = ev_times[ev_mask]
# (Under-) Estimate livetime by difference of last and first event time
tstart = astrotime.Time(np.amin(ev_t), format="mjd").iso
tstop = astrotime.Time(np.amax(ev_t), format="mjd").iso
livetimes[i] = np.amax(ev_t) - np.amin(ev_t)
# Append valid run dict to run list
run_list.append({"run": runid, "good_tstart": tstart,
"good_tstop": tstop})
print(" Livetime: {:.3f} days".format(np.sum(livetimes)))
print(" Had {} / {} runs with non-zero livetime.".format(
np.sum(livetimes > 0), len(used_run_ids)))
return run_list
# let's just remove the few runs here)
exclude = {name: None for name in all_sample_names}
exclude["IC86, 2012-2014"] = [120028, 120029, 120030]
exclude["IC86, 2015"] = [126289, 126290, 126291]
outpath = os.path.join(PATHS.local, "runlists")
if not os.path.isdir(outpath):
os.makedirs(outpath)
for name in all_sample_names:
print("Making runlist for {}".format(name))
if name in ps_sample_names:
tracks = ps_tracks
else:
tracks = gfu_tracks
exp_files = tracks.files(name)[0]
exp = tracks.load(exp_files)
_info = arr2str(exp_files if isinstance(exp_files, list) else [exp_files],
sep="\n ")
print(" Loaded {} track sample from skylab:\n {}".format(
"PS" if name in ps_sample_names else "GFU", _info))
ev_times, ev_runids = exp["time"], exp["Run"]
run_list = make_run_list(ev_times, ev_runids, exclude[name])
fname = name.replace(", ", "_") + ".json"
with open(os.path.join(outpath, fname), "w") as outf:
json.dump(run_list, fp=outf, indent=2)
print(" Saved to:\n {}".format(os.path.join(outpath, fname)))
"\nAllow overwrites (y/n)? ")
if not res.lower() in ("y", "yes"):
print("Abort. Script has done nothing.")
sys.exit()
print(" Using output directory '{}'.".format(outpath))
else:
os.makedirs(outpath)
print("Created output directory '{}'.".format(outpath))
files = np.array(glob(os.path.join(inpath, "*.json")))
file_names = np.array(map(lambda s: os.path.basename(s), files))
dataset_nums = set(map(lambda s: s.split("_")[0], file_names))
# Combine all to a single JSON file
print("Reading files from directory:\n {}".format(inpath))
print(" Found JSON files for datasets: {}".format(arr2str(dataset_nums)))
run_ids_per_sam = {}
event_ids_per_sam = {}
for num in dataset_nums:
print("Combining IDs from set '{}':".format(num))
run_ids = []
event_ids = []
_files = files[map(lambda s: s.startswith(num), file_names)]
for fi in tqdm(_files):
with open(fi, "r") as inf:
di = json.load(inf)
run_ids += di["run_id"]
event_ids += di["event_id"]
assert len(run_ids) == len(event_ids)