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

# In[1]:


from datetime import datetime
import dask.diagnostics

import rioxarray
import xarray as xr
import numpy as np
import pandas as pd
import geopandas as gpd
#import geojson
import json

from datetime import datetime

from shapely.geometry import mapping

#import zarr
import netCDF4

import os
import lxml

get_ipython().run_line_magic('matplotlib', 'inline')

import sys
import dask
import datacube

import matplotlib.pyplot as plt
from dask.distributed import Client

sys.path.append("../Scripts")
from dea_plotting import display_map
from dea_dask import create_local_dask_cluster


# In[ ]:


create_local_dask_cluster()


# In[2]:


#!conda install netCDF4 --yes
os.environ['GDAL_DISABLE_READDIR_ON_OPEN']='EMPTY_DIR' #This is KEY! otherwise we send a bunch of HTTP GET requests to test for common sidecar metadata
os.environ['AWS_NO_SIGN_REQUEST']='YES' #Since this is a public bucket, we don't need authentication
os.environ['GDAL_MAX_RAW_BLOCK_CACHE_SIZE']='200000000'  #200MB: Want this to be greater than size of uncompressed raster to overcome a 10 MB limit in the GeoTIFF driver for range request merging.
os.environ['GDAL_SWATH_SIZE']='200000000'  #also increase this if increasing MAX_RAW_BLOCK_CACHE_SIZE
os.environ['VSI_CURL_CACHE_SIZE']='200000000' #also increase this if increasing MAX_RAW_BLOCK_CACHE_SIZE
# Others?


# In[3]:


#!conda install lxml --yes
#!conda install -c conda-forge rtree=0.9.3 --yes  fixes geopandas lib spatial index dll error
filelisting_netcdf = 'https://s3-ap-southeast-2.amazonaws.com/silo-open-data/annual/index.html'


# In[4]:


oz = xr.open_rasterio(f'https://s3-ap-southeast-2.amazonaws.com/silo-open-data/annual/rh_tmax/2020.rh_tmax.nc', chunks=(1,1024,1024))


# In[5]:


oz


# In[11]:


with dask.diagnostics.ProgressBar():
    oz = oz.where(oz > -32767.0, drop=True)


# In[12]:


oz.min()


# In[13]:


oz[74].mean().compute()


# In[14]:


oz[200].mean().compute()


# In[15]:


oz[365].plot(robust=True)


# In[16]:


oz[200].shape


# In[17]:


np.unique(oz[200].values, return_counts=True)


# In[18]:


oz[200].plot()


# In[ ]:


oz.rio.to_raster(r'humid.tif')


# In[ ]: