%matplotlib inline
import pandas as pd
import socket
host = socket.getfqdn()
from core import load, zoom, calc, save,plots,monitor
#reload funcs after updating ./core/*.py
import importlib
importlib.reload(load)
importlib.reload(zoom)
importlib.reload(calc)
importlib.reload(save)
importlib.reload(plots)
importlib.reload(monitor)
<module 'core.monitor' from '/ccc/work/cont003/gen7420/odakatin/monitor-sedna/notebook/core/monitor.py'>
# 'month': = 'JOBID' almost month but not really,
# If you submit the job with job scheduler, above
#below are list of enviroment variable one can pass
#%env local='2"
# local : if True run dask local cluster, if not true, put number of workers
# setted in the 'local'
# if no 'local ' given, local will be setted automatically to 'True'
#%env ychunk='2'
#%env tchunk='2'
# controls chunk. 'False' sets no modification from original netcdf file's chunk.
# ychunk=10 will group the original netcdf file to 10 by 10
# tchunk=1 will chunk the time coordinate one by one
#%env control=FWC_SSH
# name of control file to be used for computation/plots/save/
#%env file_exp=
# 'file_exp': Which 'experiment' name is it?
#. this corresopnds to intake catalog name without path and .yaml
#%env year=
# for Validation, this correspoinds to path/year/month 's year
# for monitoring, this corresponids to 'date' having * means do all files in the monitoring directory
# setting it as *0[0-9] &*1[0-9]& *[2-3][0-9], the job can be separated in three lots.
#%env month=
# for monitoring this corresponds to file path path-XIOS.{month}/
#
#%env save= proceed saving? True or False , Default is setted as True
#%env plot= proceed plotting? True or False , Default is setted as True
#%env calc= proceed computation? or just load computed result? True or False , Default is setted as True
#%env save=False
%%time
# 'savefig': Do we save output in html? or not. keep it true.
savefig=True
client,cluster,control,catalog_url,month,year,daskreport,outputpath = load.set_control(host)
!mkdir -p $outputpath
!mkdir -p $daskreport
client
local True using host= irene4756.c-irene.mg1.tgcc.ccc.cea.fr starting dask cluster on local= True workers 16 10000000000 False not local in tgcc rome local cluster starting This code is running on irene4756.c-irene.mg1.tgcc.ccc.cea.fr using SEDNA_ALPHA_MONITOR file experiment, read from ../lib/SEDNA_ALPHA_MONITOR.yaml on year= * on month= 28 outputpath= ../results/SEDNA_ALPHA_MONITOR/28/ daskreport= ../results/dask/2673105irene4756.c-irene.mg1.tgcc.ccc.cea.fr_SEDNA_ALPHA_MONITOR_28FWC_SSH/ CPU times: user 390 ms, sys: 233 ms, total: 623 ms Wall time: 12.2 s
Client
|
Cluster
|
df=load.controlfile(control)
#Take out 'later' tagged computations
#df=df[~df['Value'].str.contains('later')]
df
Value | Inputs | Equation | Zone | Plot | Colourmap | MinMax | Unit | Oldname | Unnamed: 10 | |
---|---|---|---|---|---|---|---|---|---|---|
FWC_SSH | calc.FWC_SSH_load(data,nc_outputpath) | BBFG | FWC_SSH | None | None | m | S-1 |
Each computation consists of
%%time
import os
calcswitch=os.environ.get('calc', 'True')
loaddata=((df.Inputs != '').any())
print('calcswitch=',calcswitch,'df.Inputs != nothing',loaddata)
data = load.datas(catalog_url,df.Inputs,month,year,daskreport) if ((calcswitch=='True' )*loaddata) else 0
data
calcswitch= True df.Inputs != nothing False CPU times: user 1.2 ms, sys: 41 µs, total: 1.24 ms Wall time: 1.06 ms
0
%%time
monitor.auto(df,data,savefig,daskreport,outputpath,file_exp='SEDNA'
)
#calc= True #save= False #plot= True Zone='BBFG' Value='FWC_SSH' cmap='None' clabel='m' clim= None outputpath='../results/SEDNA_ALPHA_MONITOR/28/' nc_outputpath='../nc_results/SEDNA_ALPHA_MONITOR/28/' filename='SEDNA_FWC_SSH_BBFG_FWC_SSH' #3 Start computing dtaa= calc.FWC_SSH_load(data,nc_outputpath) start saving data filename= ../nc_results/SEDNA_ALPHA_MONITOR/28/SEDNA_maps_ALL_SSH_anomaly/t_*/x_*/y_*.nc load computed data completed start saving data filename= ../nc_results/SEDNA_ALPHA_MONITOR/28/SEDNA_maps_BBFG_FWC_2D/t_*/x_*/y_*.nc load computed data completed
<xarray.Dataset> Dimensions: (t: 15, x: 6560, y: 5264) Coordinates: * x (x) int64 1 2 3 4 5 6 7 ... 6554 6555 6556 6557 6558 6559 6560 * y (y) int64 1277 1278 1279 1280 1281 ... 6536 6537 6538 6539 6540 * t (t) object 2004-11-01 12:00:00 ... 2004-11-15 12:00:00 nav_lat (y, x) float32 dask.array<chunksize=(56, 6560), meta=np.ndarray> nav_lon (y, x) float32 dask.array<chunksize=(56, 6560), meta=np.ndarray> mask2d (y, x) bool dask.array<chunksize=(56, 6560), meta=np.ndarray> e1te2t (y, x) float64 dask.array<chunksize=(56, 6560), meta=np.ndarray> Data variables: SSH_anomaly (t, y, x) float32 dask.array<chunksize=(15, 56, 6560), meta=np.ndarray> FWC2D (t, y, x) float32 dask.array<chunksize=(1, 56, 6560), meta=np.ndarray>
array([ 1, 2, 3, ..., 6558, 6559, 6560])
array([1277, 1278, 1279, ..., 6538, 6539, 6540])
array([cftime.DatetimeNoLeap(2004, 11, 1, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 2, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 3, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 4, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 5, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 6, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 7, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 8, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 9, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 10, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 11, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 12, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 13, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 14, 12, 0, 0, 0), cftime.DatetimeNoLeap(2004, 11, 15, 12, 0, 0, 0)], dtype=object)
|
|
|
|
|
|
#5 Plotting filename= plots.FWC_SSH(data,path=outputpath,filename=filename,save=savefig,cmap=cmap,clim=clim,clabel=clabel) ../results/SEDNA_ALPHA_MONITOR/28/SEDNA_FWC_SSH_BBFG_FWC_SSH_20041101-20041115.html starts plotting ../results/SEDNA_ALPHA_MONITOR/28/SEDNA_FWC_SSH_BBFG_FWC_SSH_20041101-20041115.html created
distributed.nanny - WARNING - Restarting worker
CPU times: user 7min 11s, sys: 1min 37s, total: 8min 48s Wall time: 8min 48s