%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/talandel/TOOLS/monitor-sedna/notebook/core/monitor.py'>
If you submit the job with job scheduler; below are list of enviroment variable one can pass
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 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. For DELTA experiment, year corresponds to really 'year'
%env month=
for monitoring this corresponds to file path path-XIOS.{month}/
For DELTA experiment, year corresponds to really 'month'
proceed saving? True or False , Default is setted as True
proceed plotting? True or False , Default is setted as True
proceed computation? or just load computed result? True or False , Default is setted as True
save output file used for plotting
using kerchunked file -> False, not using kerhcunk -> True
name of control file to be used for computation/plots/save/ We have number of M_xxx.csv
Monitor.sh calls M_MLD_2D
and AWTD.sh, Fluxnet.sh, Siconc.sh, IceClim.sh, FWC_SSH.sh, Integrals.sh , Sections.sh
M_AWTMD
M_Fluxnet
M_Ice_quantities
M_IceClim M_IceConce M_IceThick
M_FWC_2D M_FWC_integrals M_FWC_SSH M_SSH_anomaly
M_Mean_temp_velo M_Mooring
M_Sectionx M_Sectiony
%%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= irene4256.c-irene.mg1.tgcc.ccc.cea.fr starting dask cluster on local= True workers 16 10000000000 rome local cluster starting This code is running on irene4256.c-irene.mg1.tgcc.ccc.cea.fr using SEDNA_DELTA_MONITOR file experiment, read from ../lib/SEDNA_DELTA_MONITOR.yaml on year= 2015 on month= 12 outputpath= ../results/SEDNA_DELTA_MONITOR/ daskreport= ../results/dask/7449030irene4256.c-irene.mg1.tgcc.ccc.cea.fr_SEDNA_DELTA_MONITOR_12M_FWC_integrals/ CPU times: user 406 ms, sys: 71.3 ms, total: 478 ms Wall time: 10.4 s
Client-2e2a76b2-6f4a-11ed-92ee-080038b9324b
Connection method: Cluster object | Cluster type: distributed.LocalCluster |
Dashboard: http://127.0.0.1:8787/status |
11a0842f
Dashboard: http://127.0.0.1:8787/status | Workers: 16 |
Total threads: 128 | Total memory: 221.88 GiB |
Status: running | Using processes: True |
Scheduler-463bb4c2-5896-4e29-b4c0-a8636602b88e
Comm: tcp://127.0.0.1:44585 | Workers: 16 |
Dashboard: http://127.0.0.1:8787/status | Total threads: 128 |
Started: Just now | Total memory: 221.88 GiB |
Comm: tcp://127.0.0.1:34427 | Total threads: 8 |
Dashboard: http://127.0.0.1:44835/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:44962 | |
Local directory: /tmp/dask-worker-space/worker-epdregry |
Comm: tcp://127.0.0.1:37242 | Total threads: 8 |
Dashboard: http://127.0.0.1:32802/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:33865 | |
Local directory: /tmp/dask-worker-space/worker-nzze09zh |
Comm: tcp://127.0.0.1:36038 | Total threads: 8 |
Dashboard: http://127.0.0.1:39027/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:35384 | |
Local directory: /tmp/dask-worker-space/worker-l7namusr |
Comm: tcp://127.0.0.1:38539 | Total threads: 8 |
Dashboard: http://127.0.0.1:37685/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:40847 | |
Local directory: /tmp/dask-worker-space/worker-afemiq1o |
Comm: tcp://127.0.0.1:41357 | Total threads: 8 |
Dashboard: http://127.0.0.1:45023/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:35302 | |
Local directory: /tmp/dask-worker-space/worker-ahhotafe |
Comm: tcp://127.0.0.1:37427 | Total threads: 8 |
Dashboard: http://127.0.0.1:40330/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:43824 | |
Local directory: /tmp/dask-worker-space/worker-3xet84e5 |
Comm: tcp://127.0.0.1:34294 | Total threads: 8 |
Dashboard: http://127.0.0.1:38211/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:36530 | |
Local directory: /tmp/dask-worker-space/worker-hm808o6b |
Comm: tcp://127.0.0.1:46821 | Total threads: 8 |
Dashboard: http://127.0.0.1:33171/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:40515 | |
Local directory: /tmp/dask-worker-space/worker-ma0qrvwe |
Comm: tcp://127.0.0.1:35362 | Total threads: 8 |
Dashboard: http://127.0.0.1:38146/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:44171 | |
Local directory: /tmp/dask-worker-space/worker-310bfm1v |
Comm: tcp://127.0.0.1:45616 | Total threads: 8 |
Dashboard: http://127.0.0.1:45926/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:35447 | |
Local directory: /tmp/dask-worker-space/worker-iepuct82 |
Comm: tcp://127.0.0.1:42782 | Total threads: 8 |
Dashboard: http://127.0.0.1:36822/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:36484 | |
Local directory: /tmp/dask-worker-space/worker-vpnwzip1 |
Comm: tcp://127.0.0.1:38452 | Total threads: 8 |
Dashboard: http://127.0.0.1:41040/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:32810 | |
Local directory: /tmp/dask-worker-space/worker-2362uq1u |
Comm: tcp://127.0.0.1:42574 | Total threads: 8 |
Dashboard: http://127.0.0.1:38534/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:36557 | |
Local directory: /tmp/dask-worker-space/worker-lfchvr9l |
Comm: tcp://127.0.0.1:44124 | Total threads: 8 |
Dashboard: http://127.0.0.1:42701/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:44357 | |
Local directory: /tmp/dask-worker-space/worker-d5dl6sdg |
Comm: tcp://127.0.0.1:43888 | Total threads: 8 |
Dashboard: http://127.0.0.1:38108/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:34064 | |
Local directory: /tmp/dask-worker-space/worker-8f807gxr |
Comm: tcp://127.0.0.1:33856 | Total threads: 8 |
Dashboard: http://127.0.0.1:45829/status | Memory: 13.87 GiB |
Nanny: tcp://127.0.0.1:39472 | |
Local directory: /tmp/dask-worker-space/worker-_ywpyy13 |
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_integrals | calc.FWC_load_integrals(data,nc_outputpath) | BBFG | FWC_integrals | (12000,24000) | Km^3 | I-1 |
Each computation consists of
%%time
import os
calcswitch=os.environ.get('calc', 'True')
lazy=os.environ.get('lazy','False' )
loaddata=((df.Inputs != '').any())
print('calcswitch=',calcswitch,'df.Inputs != nothing',loaddata, 'lazy=',lazy)
data = load.datas(catalog_url,df.Inputs,month,year,daskreport,lazy=lazy) if ((calcswitch=='True' )*loaddata) else 0
data
calcswitch= True df.Inputs != nothing False lazy= False CPU times: user 344 µs, sys: 0 ns, total: 344 µs Wall time: 343 µs
0
%%time
monitor.auto(df,data,savefig,daskreport,outputpath,file_exp='SEDNA'
)
#calc= True #save= True #plot= False Value='FWC_integrals' Zone='BBFG' Plot='FWC_integrals' cmap='' clabel='Km^3' clim= (12000, 24000) outputpath='../results/SEDNA_DELTA_MONITOR/' nc_outputpath='../nc_results/SEDNA_DELTA_MONITOR/' filename='SEDNA_FWC_integrals_BBFG_FWC_integrals' #3 Start computing data= calc.FWC_load_integrals(data,nc_outputpath) monitor.optimize_dataset(data) start loading data filename= ../nc_results/SEDNA_DELTA_MONITOR/SEDNA_maps_BBFG_FWC_2D/t_*/y_*/x_*.nc dim ('x', 'y', 't') load computed data completed add optimise here once otimise can recognise
<xarray.Dataset> Dimensions: (t: 365) Coordinates: * t (t) object 2015-01-01 12:00:00 ... 2015-12-31 12:00:00 Data variables: FWC_Arctic (t) float64 dask.array<chunksize=(1,), meta=np.ndarray> FWC_CRF (t) float64 dask.array<chunksize=(1,), meta=np.ndarray>
#4 Saving SEDNA_FWC_integrals_BBFG_FWC_integrals data=save.datas(data,plot=Plot,path=nc_outputpath,filename=filename) start saving data saving data in a csv file ../nc_results/SEDNA_DELTA_MONITOR/SEDNA_FWC_integrals_BBFG_FWC_integrals2015-01-01_2015-12-31.nc save computed data at ../nc_results/SEDNA_DELTA_MONITOR/SEDNA_FWC_integrals_BBFG_FWC_integrals2015-01-01_2015-12-31.nc completed CPU times: user 3min 28s, sys: 13.7 s, total: 3min 42s Wall time: 4min 40s