Skip to content

Commit

Permalink
Revert "Connected to CLINT data repo & added plotting"
Browse files Browse the repository at this point in the history 
This reverts commit d3c9cef.
  • Loading branch information
@mmaiergerber
mmaiergerber committed Jun 16, 2024
1 parent d3c9cef commit 9168c8e
Showing 1 changed file with 51 additions and 96 deletions.
147 changes: 51 additions & 96 deletions shearwater/processes/wps_cyclone.py
View file
Original file line number Diff line number Diff line change
@@ -1,11 +1,10 @@
from pywps import Process, LiteralInput, ComplexOutput, LiteralOutput
from pywps import Process, LiteralInput, ComplexOutput
from pywps.app.Common import Metadata
# from tensorflow.keras import models
# import pickle
import numpy as np
import numpy
import pandas as pd
import metview as mv
# from datetime import datetime
import os
from pywps import FORMATS
Expand All @@ -19,22 +18,22 @@


class Cyclone(Process):
"""A process to forecast tropical cyclone activity."""
"""A process to forecast tropical cyclone activities."""
def __init__(self):
inputs = [
LiteralInput(
"start_day",
"Start Day",
data_type="string",
abstract="Enter the start date, like 2024-05-01",
default="2024-05-01"
abstract="Enter the start date, like 2021-01-01",
default="2022-01-01"
),
LiteralInput(
"end_day",
"End Day",
data_type="string",
abstract="Enter the end date, like 2024-05-28",
default="2024-05-28"
abstract="Enter the end date, like 2023-10-12",
default="2022-01-31"
),
LiteralInput(
"area",
Expand All @@ -53,21 +52,18 @@ def __init__(self):
# abstract='netcdf',
# # keywords=['output', 'result', 'response'],
# data_type='string'),
LiteralOutput('output_png', 'Cyclone activity forecast',
abstract='png file',
data_type='string'),
#ComplexOutput('output_csv', 'Cyclone activity forecast',
# abstract='csv file',
# as_reference=True,
# # keywords=['output', 'result', 'response'],
# supported_formats=[FORMATS.CSV],)
ComplexOutput('output_csv', 'Cyclone activity forecast',
abstract='csv file',
as_reference=True,
# keywords=['output', 'result', 'response'],
supported_formats=[FORMATS.CSV],)
]

super(Cyclone, self).__init__(
self._handler,
identifier='cyclone',
title='Cyclone',
abstract='A process to forecast tropical cyclone activity.',
abstract='A process to forecast tropical cyclone activities.',
# keywords=['hello', 'demo'],
metadata=[
Metadata('PyWPS', 'https://pywps.org/'),
Expand All @@ -84,34 +80,45 @@ def _handler(self, request, response):
# TODO: lazy load tensorflow ... issues with sphinx doc build
from tensorflow.keras import models

# master_catalog = intake.open_catalog(["https://gitlab.dkrz.de/data-infrastructure-services/intake-esm/-/raw/master/esm-collections/cloud-access/dkrz_catalog.yaml"]) # noqa
# # master_catalog = intake.open_catalog('/pool/data/Catalogs/dkrz_catalog.yaml')
# era5_catalog = master_catalog['dkrz_era5_disk']

# query = {
# 'era_id': 'ET',
# 'level_type': 'surface',
# 'table_id': 128,
# # 'frequency':'hourly',
# 'code': 34,
# 'dataType': 'an',
# 'validation_date': '2023-06-27',
# }

# my_catalog = era5_catalog.search(**query)
# # my_catalog.df

# era_path = my_catalog.df['uri'].iloc[0]
# response.outputs['output'].data = f'netcdf {era_path}'
# response.outputs['output_csv'].data = 'csv ' + request.inputs['model'][0].data
# return response

start_date = request.inputs['start_day'][0].data
end_date = request.inputs['end_day'][0].data
# area = request.inputs['area'][0].data

parameters = [[138, 'vo', [850]],
[157, 'r', [700]],
[131, 'u', [200,850]],
[132, 'v', [200,850]],
[34, 'sst', [0]],
[137, 'tcwv', [0]]]
reso=2.5
area_bbox = [0,20,-30,90]

data=pd.DataFrame()
for param1 in parameters:
path = f'/pool/data/ERA5/ET/{"sf" if param1[2]==[0] else "pl"}/an/1D/{str(param1[0]).zfill(3)}'
fs1_param = mv.read(f"{path}/ET{'sf' if param1[2]==[0] else 'pl'}00_1D_{start_date[:7]}_{str(param1[0]).zfill(3)}.grb")
fs1_param = fs1_param.select(date=start_date.replace("-",""),level=param1[2])
fs1_param_interp = mv.read(data=fs1_param,grid=[reso,reso],area=area_bbox,interpolation = '"--interpolation=grid-box-average"')
for level in param1[2]:
data.loc[:,f"{param1[1]}{'_'+str(level) if (param1[1]=='u' or param1[1]=='v') else ''}"] = fs1_param_interp.select(level=level).to_dataset().to_dataframe().reset_index(drop=True)[param1[1]]

data.loc[:,['latitude','longitude']]= fs1_param_interp.select(level=level).to_dataset().to_dataframe().reset_index()[['latitude','longitude']]
data.loc[:,'time'] = start_date
data.loc[:,'shear'] = ((data.u_200-data.u_850)**2+(data.v_200-data.v_850)**2)**0.5
data.loc[:,'sst'] = data.sst.fillna(0)

variables = ['vo','r','u_200','u_850','v_200','v_850','tcwv','sst','shear']
# to be updated with data repository
data1 = pd.read_csv(
"https://github.com/climateintelligence/shearwater/raw/main/data/test_dailymeans_Sindian_1.zip")
# ("../shearwater/data/test_dailymeans_Sindian_1.zip")
data2 = pd.read_csv(
"https://github.com/climateintelligence/shearwater/raw/main/data/test_dailymeans_Sindian_2.zip")
# ("../shearwater/data/test_dailymeans_Sindian_2.zip")
data = pd.concat((data1, data2), ignore_index=True)
data = data.loc[(data.time >= start_date) & (data.time <= end_date)]

variables = ['vo', 'r', 'u_200', 'u_850', 'v_200', 'v_850', 'tcwv', 'sst', 'shear']
# with open("https://github.com/climateintelligence/shearwater/raw/main/data/full_statistics.pkl", 'rb') as f:
# means, stds = pickle.load(f)
means, stds = pd.read_pickle(
"https://github.com/climateintelligence/shearwater/raw/main/data/full_statistics.zip")

Expand Down Expand Up @@ -155,60 +162,8 @@ def _handler(self, request, response):
data = data[["latitude", "longitude", "time"]]
data['predictions_lag0'] = prediction.reshape(-1, 1)

lag=0
workdir = Path(self.workdir)
outfilename = os.path.join(workdir, f'tcactivity_48_17_{start_date.replace("-","")}_lag{lag}_Sindian')

if True:
predscol=f"predictions_lag{lag}"
gpt = mv.create_geo(latitudes=data["latitude"].values, longitudes=data["longitude"].values, values=data[predscol].values).set_dates([pd.Timestamp(start_date)]*data.shape[0])
fs = mv.geo_to_grib(geopoints = gpt, grid = [2.5,2.5], tolerance = 1.5)*1e2

cont_gen = mv.mcont(
legend = "on",
contour_line_colour = "avocado",
contour_shade = "on",
contour_shade_technique = "grid_shading",
contour_shade_max_level_colour = "red",
contour_shade_min_level_colour = "blue",
contour_shade_colour_direction = "clockwise"
)
cont_tc = mv.mcont(
legend = "on",
contour_line_colour = "avocado",
contour_shade = "on",
contour_max_level = 105,
contour_min_level = 0,
contour_shade_technique = "grid_shading",
contour_shade_max_level_colour = "red",
contour_shade_min_level_colour = "blue",
contour_shade_colour_direction = "clockwise"
)

cont_oper = mv.mcont(
contour_automatic_setting = "style_name",
contour_style_name = "prob_green2yellow",
legend = "on"
)
coastlines = mv.mcoast(
map_coastline_land_shade = "on",
map_coastline_land_shade_colour = "grey"
)

gview = mv.geoview(
map_area_definition = "corners",
area = area_bbox,
coastlines = coastlines
)
legend = mv.mlegend(
legend_text_font_size = 0.5,
)

mv.setoutput(mv.png_output(output_name = outfilename+".png"))
mv.plot(gview,fs,cont_oper,legend)
response.outputs['output_png'].data = outfilename+".png"
else:
data.to_csv(outfilename+".csv")
response.outputs['output_csv'].data = outfilename+".csv"

prediction_path = os.path.join(workdir, "prediction_Sindian.csv")
data.to_csv(prediction_path)

response.outputs['output_csv'].file = prediction_path
return response

0 comments on commit 9168c8e

Please sign in to comment.