#!/usr/bin/env python
# coding: utf-8

# # Performance of concat operator on cmip6 decadal: DWD
# 
# **Rooki** calls climate data operations on the **rook** processing service.

# In[1]:


import os
import time
os.environ['ROOK_URL'] = 'http://rook.dkrz.de/wps'

from rooki import operators as ops


# ## Load intake catalog for cmip6 decadal

# In[2]:


import intake


# In[3]:


cat_url = "https://raw.githubusercontent.com/cp4cds/c3s_34g_manifests/master/intake/catalogs/c3s.yaml"

cat = intake.open_catalog(cat_url)
df_cmip6_decadal = cat['c3s-cmip6-decadal'].read()
df_cmip6_decadal.head()


# In[4]:


df = df_cmip6_decadal.loc[
    (df_cmip6_decadal.variable_id=="pr") 
    & (df_cmip6_decadal.experiment_id=="dcppA-hindcast")
    & (df_cmip6_decadal.table_id=="Amon")
    & (df_cmip6_decadal.source_id=="MPI-ESM1-2-LR")
    & (df_cmip6_decadal.member_year=="s2018")
]
dsets = list(df.ds_id.unique())
dsets


# ## Run concat
# 
# https://docs.xarray.dev/en/stable/generated/xarray.concat.html

# In[5]:


start = time.time()

t = "2019/2019"
#tc = "month:dec|day:01"

wf = ops.Concat(
        ops.Input('ds', dsets),
        time=t,
        #time_components=tc,
        dims="realization",
        #apply_average=1
)

# optional average
wf = ops.Average(wf, dims="realization")

# optional subset
wf = ops.Subset(
    wf, time=t, 
    # time_components=tc
)

resp = wf.orchestrate()

end = time.time()
print(f"duration: {int(end - start)} secs")
print(resp)

resp.ok


# **open with xarray**

# In[6]:


resp.size_in_mb


# In[7]:


resp.download_urls()


# In[8]:


dsets = resp.datasets()


# In[9]:


ds = dsets[0]
ds


# **provenance**

# In[10]:


prov_plot_url = resp.provenance_image()
prov_plot_url


# In[11]:


# from IPython.display import Image
# Image(prov_plot_url)


# In[12]:


resp.provenance()