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

# # Dfsu - Export to netcdf
# * Read data from dfsu file
# * Convert to xarray dataset
# * Write to netcdf file

# In[1]:


# pip install mikeio xarray netcdf4


# In[2]:


import os
import mikeio
import xarray as xr


# In[3]:


ds = mikeio.read("../tests/testdata/oresund_sigma_z.dfsu")
ds


# In[4]:


nc = ds.geometry.node_coordinates
xn = nc[:,0]
yn = nc[:,1]
zn = nc[:,2]

ec = ds.geometry.element_coordinates
xe = ec[:,0]
ye = ec[:,1]
ze = ec[:,2]


# In[5]:


# Time
time = ds.time

# Node based data
node_ids = list(range(len(nc)))
z_dynamic = ds._zn
xn_da = xr.DataArray(xn, coords=[node_ids], dims=["nodes"], attrs={'units': 'meter'})
yn_da = xr.DataArray(xn, coords=[node_ids], dims=["nodes"], attrs={'units': 'meter'})
zn_da = xr.DataArray(zn, coords=[node_ids], dims=["nodes"], attrs={'units': 'meter'})
z_dyn_da = xr.DataArray(z_dynamic, coords =[time,node_ids],dims=["time", "nodes"], attrs={'units': 'meter'})

# Element based data
el_ids = list(range(len(ec)))
xe_da = xr.DataArray(xe, coords=[el_ids], dims=["elements"], attrs={'units': 'meter'})
ye_da = xr.DataArray(ye, coords=[el_ids], dims=["elements"], attrs={'units': 'meter'})
ze_da = xr.DataArray(ze, coords=[el_ids], dims=["elements"], attrs={'units': 'meter'})


# In[6]:


# Add coordinates for nodes and elements
data_dict = {'x': xn_da,
             'y' :yn_da,
             'z' : zn_da,
             'xe' : xe_da,
             'ye' : ye_da,
             'ze' : ze_da,
             'z_dynamic' : z_dyn_da}

# add rest of data
for da in ds:
        da = xr.DataArray(da.to_numpy(), 
                          coords = [time,el_ids],
                          dims=["time", "elements"],
                          attrs={'units': da.unit.name})

        data_dict[da.name] = da


# Here are some examples of global attributes, which is useful, but in most cases not required
attributes={'title:' : "Model A.2:4",
            'history': 'mikeio | xarray',
            'source': 'Mike 3 FM - Oresund',
            'instituion': 'DHI'}

# create an xarray dataset
xr_ds = xr.Dataset(data_dict, attrs=attributes)


# In[7]:


xr_ds.to_netcdf("oresund_sigma_z.nc")


# # Clean up

# In[8]:


import os
os.remove("oresund_sigma_z.nc")