from tempfile import tempdir
import firebase_admin
from firebase_admin import credentials
from firebase_admin import db
from find_s1_image import find_img_value
import pandas as pd
import statsmodels.api as sm
import seaborn as sns
sns.set()
import time
import json
import numpy as np
from sklearn.cluster import KMeans
from find_study_area_values import find_study_area_values
from make_area_estimate_image import make_area_estimate_image
from sentinelhub import WebFeatureService, BBox, CRS, DataSource, MimeType, CRS, BBox, WmsRequest,DataCollection
cred = credentials.Certificate('servicekey.json')
x = 0.0003
y = 0.0003
fromdate = '20220701'
todate = '20220715'

s1_images = ['IW-VH-DB']
s2_images = []
# s1_images = ["B02", "B03", "B04", "B05"]
#s2_images = ["RVI-NEW"]
try:
    firebase_admin.initialize_app(cred, {'databaseURL': 'https://farmbase-b2f7e-31c0c.firebaseio.com/'})
except:
    print('fire running')
study_uid = "MBFshQgimkS6AwSDtqAyOpRXZW12"
gt_uid = 'hxA3VYEFpaOluurIbHxdzd4gPlw1'
different_gt_uid = 'joCrKyhfobbzgCxCQPHJfOOKKPF3'
study_field_id = '1599307314955'
gt_fields_dir = 'Downloads/gt_egypt.json'
#try:
with open(gt_fields_dir) as f:
    gt_fields = json.load(f)

new_gt_json = {}
x = 0.0004
y = 0.0004
gt_fields = gt_fields["Sheet1"]
for single_gt_field in gt_fields:
  gt_lat = single_gt_field["lat"]
  gt_lng = single_gt_field["lng"]
  gt_crop = single_gt_field["Crop"]
  try:
    new_gt_json[gt_crop].append([gt_lat, gt_lng])
  except:
    new_gt_json[gt_crop] = []
    new_gt_json[gt_crop].append([gt_lat, gt_lng])
print(new_gt_json)

sentinelSettings = db.reference('SentinelSettings').get()
clientID = sentinelSettings["ClientID"]
clientSecret = sentinelSettings["ClientSecret"]
wmsID = sentinelSettings["WMSID"]
rviID = sentinelSettings["RVIID"]
demID = sentinelSettings["DEMID"]
field_num = 0
from_date = '2022-06-25T00:00:00'
to_date = '2022-09-10T23:59:59'
max_num = 16

def append_img_vals(temp_arr, img_values):
    if temp_arr is not None:
        for k in temp_arr:
            img_values.append(k)
    return img_values

crop_wise_img_values = {}
all_fields_img_values = []
i_num = 0
for (crop, all_points) in new_gt_json.items():
  for single_point in all_points:
    midlat = float(single_point[0])
    midlng = float(single_point[1])
    bounds = [(midlng-y),(midlat-x),(midlng+y),(midlat+x)]
    img_values = []
    for image in s1_images:
            temp_arr = find_img_value(crop,DataCollection.SENTINEL1_IW, image, bounds, from_date, to_date, clientID, clientSecret, rviID, max_num)
            img_values = append_img_vals(temp_arr, img_values)
            #print(img_values)
            if len(img_values) > 0:
              i_num = i_num + 1
              print('i_num: ' + str(i_num))
              all_fields_img_values.append(img_values)
              try:
                crop_wise_img_values[crop].append(img_values)
              except:
                crop_wise_img_values[crop] = []
                crop_wise_img_values[crop].append(img_values)
print(i_num)
print(all_fields_img_values)
print(crop_wise_img_values)

np.savetxt("all_fields_img_values.csv", all_fields_img_values, delimiter = ",")
np.savetxt("crop_wise_img_values.csv", crop_wise_img_values, delimiter = ",")
k_groups = 14
kmeans = KMeans(k_groups)
#kmeans = KMeans(n_clusters =2, random_state = 0)
kmeans.fit(all_fields_img_values)
for (crop, crop_values) in crop_wise_img_values.items():
  cluster_array = kmeans.predict(crop_values)
  unique, count_array = np.unique(cluster_array, return_counts = True)
  result = np.column_stack((unique, count_array))
  print(result)

new_count_array = sorted(result, key = lambda x:x[1], reverse=True)
print(new_count_array)
pred_vals = []
sum_count_array = np.sum(count_array)
j = 0
for i in new_count_array:
    if j/sum_count_array < 0.7:
        pred_vals.append(i[0])
        j = j + i[1]
    else:
        break

print(pred_vals)

study_pixel_values_arr = []
midlat = (float(study_field["FieldMinLat"]) + float(study_field["FieldMaxLat"]))/2
midlng = (float(study_field["FieldMinLong"]) + float(study_field["FieldMaxLong"]))/2
bounds = [float(study_field["FieldMinLong"]),float(study_field["FieldMinLat"]),float(study_field["FieldMaxLong"]),float(study_field["FieldMaxLat"])]
print(bounds)
image_num = 0
for image in s2_images:
    temp_arr,w,h = find_study_area_values(DataCollection.SENTINEL2_L1C,image, bounds, from_date, to_date, clientID, clientSecret, wmsID, max_num)
for image in s1_images:
    temp_arr,w,h = find_study_area_values(DataCollection.SENTINEL1_IW, image, bounds, from_date, to_date, clientID, clientSecret, rviID, max_num)

Identified_clusters = kmeans.fit_predict(temp_arr)
# np.savetxt('identified_arr.csv',Identified_clusters, delimiter=",")
make_area_estimate_image(Identified_clusters, w, h, study_uid, study_field_id, pred_vals)