Data preprocessing & generation
Browse filesdata generation to derive HLS chips and MERRA, flux csv.
- fluxconfig.yaml +78 -0
- make_chips.py +283 -0
- prep_input.py +169 -0
fluxconfig.yaml
ADDED
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
model:
|
2 |
+
name: "Base_Flux"
|
3 |
+
n_channel: 6
|
4 |
+
n_class: 1
|
5 |
+
embed_dim: 1024
|
6 |
+
dropout_rate: 0.5
|
7 |
+
|
8 |
+
device_name: "cuda"
|
9 |
+
|
10 |
+
n_iteration: 50
|
11 |
+
|
12 |
+
prithvi_model_new_weight: "/vol/cephfs/impact/srija/Prithvi-Global-downstream_v0/new_flood/checkpoint.pt"
|
13 |
+
|
14 |
+
training:
|
15 |
+
train_batch_size: 16
|
16 |
+
shuffle: True
|
17 |
+
|
18 |
+
optimizer:
|
19 |
+
name: "AdamW"
|
20 |
+
params:
|
21 |
+
lr: 5e-5
|
22 |
+
|
23 |
+
|
24 |
+
scheduler:
|
25 |
+
use: 1
|
26 |
+
name: "ReduceLROnPlateau"
|
27 |
+
|
28 |
+
dropout:
|
29 |
+
use: 1
|
30 |
+
val: 0.2
|
31 |
+
|
32 |
+
bn: 1
|
33 |
+
|
34 |
+
|
35 |
+
|
36 |
+
testing:
|
37 |
+
test_batch_size: 16
|
38 |
+
shuffle: False
|
39 |
+
|
40 |
+
normalization: "z-score-std"
|
41 |
+
test_year: 2021
|
42 |
+
|
43 |
+
|
44 |
+
data:
|
45 |
+
n_frame: 1
|
46 |
+
chips: "/vol/cephfs/impact/srija/Prithvi-Global-downstream_v0/new_flood_v2/chips/"
|
47 |
+
input_size: [6,50, 50]
|
48 |
+
means_for2018test: [0.07286696773903256, 0.10036772476940378, 0.11363777043869523, 0.2720510638470194, 0.2201167122609674, 0.1484162876040495]
|
49 |
+
stds_for2018test: [0.13271414936598172, 0.13268933338964875, 0.1384673725283858, 0.12089142598551804, 0.10977084890500641, 0.0978705241034744]
|
50 |
+
merra_means_for2018test: [282.011721, 295.823746,288.291530, 278.243071,0.552373,55.363476, 48.984387, 202.461732, 22.907336,0.000004]
|
51 |
+
merra_stds_for2018test: [9.141752,11.374619,10.224494,7.912334,0.178115,50.069111,48.238661,74.897672,9.277971,0.000014]
|
52 |
+
gpp_means_for2018test: [3.455948]
|
53 |
+
gpp_stds_for2018test: [3.754123]
|
54 |
+
means_for2019test: [0.07287311832611834,0.10025904848484847,0.1122947444733045,0.27563822551226563,0.21583184092352084,0.14331408109668098]
|
55 |
+
stds_for2019test: [0.13511944688809177,0.1349403534769768,0.14037014996437144,0.12365673294486092,0.10852189245620811,0.09485890083382985]
|
56 |
+
merra_means_for2019test: [281.960274,295.974675, 288.330014,278.306133,0.548831,55.167287,48.381169,202.003449,23.097742,0.000004]
|
57 |
+
merra_stds_for2019test: [9.077508,11.436697,10.178588,7.750465, 0.175302,50.656796,49.182061,73.949519,9.422290,0.000016]
|
58 |
+
gpp_means_for2019test: [3.581604]
|
59 |
+
gpp_stds_for2019test: [3.889343]
|
60 |
+
means_for2020test: [0.07372144372093026,0.10117611215116282,0.11269885680232558,0.2775572554069766,0.21387001372093037, 0.14144541145348838]
|
61 |
+
stds_for2020test: [0.13324302628303733, 0.13308921403475235, 0.13829909331863693,0.12039809083338567,0.1088096350639653,0.09366368859284444]
|
62 |
+
merra_means_for2020test: [282.373169, 296.706468, 288.852922, 278.612209, 0.540145, 53.830276, 53.827718, 206.817980, 23.077581, 0.000003]
|
63 |
+
merra_stds_for2020test: [9.296960, 11.402008, 10.311107, 8.064209, 0.171909, 49.945953, 48.907351, 74.591578, 8.746668, 0.000014 ]
|
64 |
+
gpp_means_for2020test: [3.668982]
|
65 |
+
gpp_stds_for2020test: [3.804261]
|
66 |
+
means_for2021test: [0.06743080268702287,0.09420638137404584,0.10626692164885497,0.2692502415877864,0.21780909367938925,0.1468194037862596]
|
67 |
+
stds_for2021test: [0.12261400510468322,0.12276593355350174,0.12836180894665594,0.11639597942158948,0.10570861595781685,0.09646322486302224]
|
68 |
+
merra_means_for2021test: [281.762443,294.883832,287.753053, 279.168366,0.569313,61.064687,45.930611,200.842519,23.072735,0.000003]
|
69 |
+
merra_stds_for2021test: [9.040586,11.143439,10.063070,8.121612,0.172953,52.172274,47.056911,76.875468,9.553304,0.000010]
|
70 |
+
gpp_means_for2021test: [3.787582]
|
71 |
+
gpp_stds_for2021test: [3.862494]
|
72 |
+
|
73 |
+
|
74 |
+
logging:
|
75 |
+
checkpoint_dir: "/vol/cephfs/impact/srija/Prithvi-Global-downstream_v0/Prithvi-global-v1/flux_base_pred_logs/"
|
76 |
+
metrics_dir: "/vol/cephfs/impact/srija/Prithvi-Global-downstream_v0/Prithvi-global-v1/flux_base_pred_logs/metrics/"
|
77 |
+
plots_dir: "/vol/cephfs/impact/srija/Prithvi-Global-downstream_v0/Prithvi-global-v1/flux_base_pred_logs/plots/"
|
78 |
+
|
make_chips.py
ADDED
@@ -0,0 +1,283 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
"""
|
2 |
+
make_chips.py
|
3 |
+
|
4 |
+
This script reads in HLS S30/L30 data and extracts
|
5 |
+
band information around a chip_size x chip_size subset
|
6 |
+
of the original raster grid. Snowy and cloudy chips beyond a
|
7 |
+
threshold are discarded.
|
8 |
+
|
9 |
+
Author: Besart Mujeci, Srija Chakraborty, Christopher Phillips
|
10 |
+
|
11 |
+
Usage:
|
12 |
+
python make_chips.py
|
13 |
+
"""
|
14 |
+
import rclone
|
15 |
+
from pathlib import Path
|
16 |
+
import shutil
|
17 |
+
import pandas as pd
|
18 |
+
from collections import Counter
|
19 |
+
import cartopy.crs as ccrs
|
20 |
+
import numpy as np
|
21 |
+
import rasterio
|
22 |
+
from rasterio.transform import from_gcps
|
23 |
+
from rasterio.warp import transform
|
24 |
+
from rasterio.windows import Window
|
25 |
+
import os
|
26 |
+
|
27 |
+
|
28 |
+
# --- --- ---
|
29 |
+
def point_to_index(dataset, long, lat):
|
30 |
+
"""
|
31 |
+
Converts long/lat point to row, col position on rasterio grid.
|
32 |
+
|
33 |
+
Args:
|
34 |
+
dataset (Rasterio Object): rasterio object
|
35 |
+
long (float): longitude float
|
36 |
+
lat (float): latitude float
|
37 |
+
|
38 |
+
|
39 |
+
Returns:
|
40 |
+
tuple: tuple representing point mapping on grid
|
41 |
+
"""
|
42 |
+
from_crs = rasterio.crs.CRS.from_epsg(4326)
|
43 |
+
to_crs = dataset.crs
|
44 |
+
new_x,new_y = transform(from_crs,to_crs, [long], [lat])
|
45 |
+
new_x = new_x[0]
|
46 |
+
new_y = new_y[0]
|
47 |
+
|
48 |
+
# get row and col
|
49 |
+
row, col = dataset.index(new_x,new_y)
|
50 |
+
return(row, col)
|
51 |
+
# --- --- ---
|
52 |
+
|
53 |
+
|
54 |
+
# --- --- --- Citation for this function: Christopher Phillips
|
55 |
+
def check_qc_bit(data, bit):
|
56 |
+
"""
|
57 |
+
Function to check QC flags
|
58 |
+
|
59 |
+
Args:
|
60 |
+
data (numpy array): rasterio numpy grid
|
61 |
+
bit (int): 1 or 4 representing cloud or snow
|
62 |
+
|
63 |
+
Returns:
|
64 |
+
numpy array: numpy array with flagged indices marking cloud/snow
|
65 |
+
"""
|
66 |
+
qc = np.array(data//(10**bit), dtype='int')
|
67 |
+
qc = qc-((qc//2)*2)
|
68 |
+
|
69 |
+
return np.sum(qc)/qc.size
|
70 |
+
# --- --- ---
|
71 |
+
|
72 |
+
|
73 |
+
# --- --- --- rclone configuration, file collection
|
74 |
+
cfg = ""
|
75 |
+
result = rclone.with_config(cfg).run_cmd("ls", extra_args=[f"{idir}/"])
|
76 |
+
output_lines = result['out'].decode('utf-8').splitlines()
|
77 |
+
file_list = [line.split(maxsplit=1)[1] for line in output_lines if line]
|
78 |
+
# --- --- ---
|
79 |
+
|
80 |
+
|
81 |
+
# --- --- --- Options
|
82 |
+
hls_type = 'L30' # Switch between 'L30' and 'S30' manually.
|
83 |
+
idir = "" # Raw Images Dir
|
84 |
+
odir = "" # Output Chips Dir
|
85 |
+
chip_size = 50 # Chip dimensions
|
86 |
+
scale = 0.0001 # Scale value for HLS bandssqm
|
87 |
+
cthresh = 0.05 # Cloud threshold
|
88 |
+
sthresh = 0.02 # Snow/ice threshold
|
89 |
+
# --- --- ---
|
90 |
+
|
91 |
+
|
92 |
+
# --- --- --- Read station site data
|
93 |
+
df = pd.read_csv("./TILED_filtered_flux_sites_2018_2021.csv")
|
94 |
+
stations = df['SITE_ID'].tolist()
|
95 |
+
tiles = [tile.split(";")[0] for tile in df['tiles'].tolist()]
|
96 |
+
sYear = df['start_year'].tolist()
|
97 |
+
eYear = df['end_year'].tolist()
|
98 |
+
longs = df['LOCATION_LONG'].tolist()
|
99 |
+
lats = df['LOCATION_LAT'].tolist()
|
100 |
+
all_years = [str(sYear[i]) + "-" + str(eYear[i]) for i in range(len(df))]
|
101 |
+
coords = [str(lat) + ";" + str(long) for lat, long in zip(lats, longs)]
|
102 |
+
# --- --- ---
|
103 |
+
|
104 |
+
|
105 |
+
for i, line in enumerate(tiles):
|
106 |
+
station_data = [stations[i], coords[i].split(";")[0], coords[i].split(";")[1], all_years[i].split("-")[0], all_years[i].split("-")[1], "filler", tiles[i]]
|
107 |
+
tile = station_data[-1].strip()
|
108 |
+
print(f"Working on {tile}")
|
109 |
+
|
110 |
+
# Determine years for this station
|
111 |
+
years = range(int(station_data[3]), int(station_data[4])+1)
|
112 |
+
for year in years:
|
113 |
+
print(year)
|
114 |
+
|
115 |
+
# Build path to this tile and locate all tifs
|
116 |
+
tifs1 = sorted([filepath for filepath in file_list if tile in filepath and "B01" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
117 |
+
tifs2 = sorted([filepath for filepath in file_list if tile in filepath and "B02" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
118 |
+
tifs3 = sorted([filepath for filepath in file_list if tile in filepath and "B03" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
119 |
+
tifs4 = sorted([filepath for filepath in file_list if tile in filepath and "B04" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
120 |
+
tifs5 = sorted([filepath for filepath in file_list if tile in filepath and "B05" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
121 |
+
tifs6 = sorted([filepath for filepath in file_list if tile in filepath and "B06" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
122 |
+
tifs7 = sorted([filepath for filepath in file_list if tile in filepath and "B07" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
123 |
+
tifs8 = sorted([filepath for filepath in file_list if tile in filepath and "B08" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
124 |
+
tifs8A = sorted([filepath for filepath in file_list if tile in filepath and "B8A" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
125 |
+
tifs9 = sorted([filepath for filepath in file_list if tile in filepath and "B09" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
126 |
+
tifs10 = sorted([filepath for filepath in file_list if tile in filepath and "B10" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
127 |
+
tifs11 = sorted([filepath for filepath in file_list if tile in filepath and "B11" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
128 |
+
tifs12 = sorted([filepath for filepath in file_list if tile in filepath and "B12" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
129 |
+
tifsF = sorted([filepath for filepath in file_list if tile in filepath and "Fmask" in filepath and hls_type in filepath and str(year) == filepath.split(".")[3][:4]]) # Numbered by band
|
130 |
+
|
131 |
+
# Loop over each tif
|
132 |
+
first = True
|
133 |
+
chip_flag = False # Flag for detecting chip size errors
|
134 |
+
for i in range(len(tifs2)):
|
135 |
+
|
136 |
+
# Open tifs based on HLS product
|
137 |
+
skip_file_iteration = False
|
138 |
+
if (hls_type == 'L30'):
|
139 |
+
# Ensure the sorted files are aligned correctly.
|
140 |
+
# If a band is missing then things can go out of order.
|
141 |
+
# Push 'filler' if layer is missing a band to maintain sorting.
|
142 |
+
checkListMain = [tifs2, tifs3, tifs4, tifs5, tifs6, tifs7, tifsF]
|
143 |
+
checkList = [tifs2[i], tifs3[i], tifs4[i], tifs5[i], tifs6[i], tifs7[i], tifsF[i]]
|
144 |
+
checkList = ['.'.join(ele.split(".")[2:4]) for ele in checkList]
|
145 |
+
counts = Counter(checkList)
|
146 |
+
common_value, _ = counts.most_common(1)[0]
|
147 |
+
for z, value in enumerate(checkList):
|
148 |
+
if value != common_value:
|
149 |
+
checkListMain[z].insert(i, "filler") # Push
|
150 |
+
skip_file_iteration=True
|
151 |
+
print(f"Misaligned - {checkList}")
|
152 |
+
break
|
153 |
+
if skip_file_iteration:
|
154 |
+
continue
|
155 |
+
try:
|
156 |
+
if not os.path.exists(f"./{tile}"):
|
157 |
+
os.makedirs(f"./{tile}")
|
158 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs2[i]}", f"./{tile}")
|
159 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs3[i]}", f"./{tile}")
|
160 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs4[i]}", f"./{tile}")
|
161 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs5[i]}", f"./{tile}")
|
162 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs6[i]}", f"./{tile}")
|
163 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs7[i]}", f"./{tile}")
|
164 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifsF[i]}", f"./{tile}")
|
165 |
+
except:
|
166 |
+
print(f"MISALIGNED FOR - {tifs2[i]} check if all bands exist")
|
167 |
+
continue
|
168 |
+
|
169 |
+
src2 = rasterio.open(tifs2[i])
|
170 |
+
src3 = rasterio.open(tifs3[i])
|
171 |
+
src4 = rasterio.open(tifs4[i])
|
172 |
+
src5 = rasterio.open(tifs5[i])
|
173 |
+
src6 = rasterio.open(tifs6[i])
|
174 |
+
src7 = rasterio.open(tifs7[i])
|
175 |
+
srcF = rasterio.open(tifsF[i])
|
176 |
+
|
177 |
+
elif (hls_type == 'S30'):
|
178 |
+
# Ensure the sorted files are aligned correctly.
|
179 |
+
# If a band is missing then order is compromised.
|
180 |
+
# Push 'filler' if layer is missing a band to maintain sorting.
|
181 |
+
checkListMain = [tifs2, tifs3, tifs4, tifs8A, tifs11, tifs12, tifsF]
|
182 |
+
checkList = [tifs2[i], tifs3[i], tifs4[i], tifs8A[i], tifs11[i], tifs12[i], tifsF[i]]
|
183 |
+
checkList = ['.'.join(ele.split(".")[2:4]) for ele in checkList]
|
184 |
+
counts = Counter(checkList)
|
185 |
+
common_value, _ = counts.most_common(1)[0]
|
186 |
+
for z, value in enumerate(checkList):
|
187 |
+
if value != common_value:
|
188 |
+
checkListMain[z].insert(i, "filler")
|
189 |
+
skip_file_iteration=True
|
190 |
+
break
|
191 |
+
if skip_file_iteration:
|
192 |
+
continue
|
193 |
+
try:
|
194 |
+
if not os.path.exists(f"./{tile}"):
|
195 |
+
os.makedirs(f"./{tile}")
|
196 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs2[i]}", f"./{tile}")
|
197 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs3[i]}", f"./{tile}")
|
198 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs4[i]}", f"./{tile}")
|
199 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs8A[i]}", f"./{tile}")
|
200 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs11[i]}", f"./{tile}")
|
201 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifs12[i]}", f"./{tile}")
|
202 |
+
rclone.with_config(cfg).copy(f"{idir}/{tifsF[i]}", f"./{tile}")
|
203 |
+
except:
|
204 |
+
print(f"MISALIGNED FOR - {tifs2[i]} check if all bands exist")
|
205 |
+
continue
|
206 |
+
|
207 |
+
|
208 |
+
src2 = rasterio.open(f"./{tifs2[i]}")
|
209 |
+
src3 = rasterio.open(f"./{tifs3[i]}")
|
210 |
+
src4 = rasterio.open(f"./{tifs4[i]}")
|
211 |
+
src5 = rasterio.open(f"./{tifs8A[i]}")
|
212 |
+
src6 = rasterio.open(f"./{tifs11[i]}")
|
213 |
+
src7 = rasterio.open(f"./{tifs12[i]}")
|
214 |
+
srcF = rasterio.open(f"./{tifsF[i]}")
|
215 |
+
|
216 |
+
else:
|
217 |
+
raise ValueError(f'HLS product type must be \"L30\" or \"S30\" not \"{hls_type}\".')
|
218 |
+
|
219 |
+
# Station remains in the same spot/tile so only gather information once.
|
220 |
+
if first:
|
221 |
+
row, col = point_to_index(src2, float(station_data[2]), float(station_data[1]))
|
222 |
+
|
223 |
+
y_offset = row - (chip_size // 2)
|
224 |
+
x_offset = col - (chip_size // 2)
|
225 |
+
|
226 |
+
window = Window(y_offset, x_offset, chip_size, chip_size)
|
227 |
+
window_data = src2.read(window=window, boundless=True)
|
228 |
+
window_transform = src2.window_transform(window)
|
229 |
+
|
230 |
+
first = False
|
231 |
+
|
232 |
+
# Subset tif
|
233 |
+
bands = []
|
234 |
+
for src in (src2,src3,src4,src5,src6,src7): # Set the tuple to match desired bands
|
235 |
+
|
236 |
+
# Scale and clip reflectances
|
237 |
+
band = np.clip(src.read(1)[y_offset:y_offset + chip_size, x_offset:x_offset + chip_size]*scale, 0, 1)
|
238 |
+
bands.append(band)
|
239 |
+
bands = np.array(bands)
|
240 |
+
|
241 |
+
# Check chip size and break out if wrong shape
|
242 |
+
if (bands.shape[1] != chip_size) or (bands.shape[2] != chip_size):
|
243 |
+
print(f'ERROR: Chip for tile {tile} is wronge size!\n Size is {band.shape[1:]} and not ({chip_size},{chip_size}).\nSkipping to next tile.')
|
244 |
+
chip_flag = True
|
245 |
+
break
|
246 |
+
|
247 |
+
# Subset Fmask to get imperfections
|
248 |
+
cbands = np.array(srcF.read(1)[y_offset:y_offset + 50, x_offset:x_offset + 50], dtype='int')
|
249 |
+
cloud_frac = check_qc_bit(cbands, 1)
|
250 |
+
snow_frac = check_qc_bit(cbands, 4)
|
251 |
+
|
252 |
+
# Check cloud fraction
|
253 |
+
if (cloud_frac > cthresh):
|
254 |
+
print("CLOUDY")
|
255 |
+
continue
|
256 |
+
|
257 |
+
# Check snow/ice fraction
|
258 |
+
if (snow_frac > sthresh):
|
259 |
+
print("SNOWY")
|
260 |
+
continue
|
261 |
+
|
262 |
+
# Save chip with new metadata
|
263 |
+
out_meta = src2.meta
|
264 |
+
out_meta.update({'driver':'GTiff', 'height':bands.shape[1],
|
265 |
+
'width':bands.shape[2], 'count':bands.shape[0], 'dtype':bands.dtype,
|
266 |
+
'transform':window_transform})
|
267 |
+
save_name = f'./chips/{tifs2[i].replace("B02", f"{station_data[0]}_merged.{chip_size}x{chip_size}pixels")}'
|
268 |
+
if not os.path.exists(save_name):
|
269 |
+
os.makedirs(f"./chips/{tile}")
|
270 |
+
with rasterio.open(save_name, 'w', **out_meta) as dest:
|
271 |
+
dest.write(bands)
|
272 |
+
|
273 |
+
rclone.with_config(cfg).copy(f"./chips/{tile}", f"{odir}/{tile}/")
|
274 |
+
shutil.rmtree(Path(f"./chips/"))
|
275 |
+
|
276 |
+
# If chip is the wrong size break to next station
|
277 |
+
if chip_flag:
|
278 |
+
print("Breaking to tile -- wrong size ")
|
279 |
+
break
|
280 |
+
shutil.rmtree(Path(f"./{tile}"))
|
281 |
+
break
|
282 |
+
|
283 |
+
print('Done chipping.')
|
prep_input.py
ADDED
@@ -0,0 +1,169 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
"""
|
2 |
+
prep_input.py
|
3 |
+
|
4 |
+
This script reads in MERRA2 SLV,LND data in combination with
|
5 |
+
flux station data together with HLS chips to produce a CSV with
|
6 |
+
aggregate data values that can be used to train for GPP flux prediction.
|
7 |
+
|
8 |
+
Author: Besart Mujeci, Srija Chakraborty, Christopher Phillips
|
9 |
+
|
10 |
+
Usage:
|
11 |
+
python prep_input.py
|
12 |
+
"""
|
13 |
+
import rclone
|
14 |
+
from pathlib import Path
|
15 |
+
import shutil
|
16 |
+
import os
|
17 |
+
import pandas as pd
|
18 |
+
import netCDF4 as nc
|
19 |
+
import numpy as np
|
20 |
+
import pandas as pd
|
21 |
+
|
22 |
+
|
23 |
+
# --- --- ---
|
24 |
+
def convert_HLS_date(chip_name):
|
25 |
+
"""
|
26 |
+
Extracts date string from HLS tile name and returns date object
|
27 |
+
|
28 |
+
Args:
|
29 |
+
chip_name (string): name of hls file
|
30 |
+
|
31 |
+
Returns:
|
32 |
+
datetime: datetime object of time string
|
33 |
+
"""
|
34 |
+
hls_date = chip_name.split('.')[3][:7]
|
35 |
+
year = int(hls_date[:4])
|
36 |
+
day = int(hls_date[4:])
|
37 |
+
date = datetime(year, 1, 1)+timedelta(days=day-1)
|
38 |
+
|
39 |
+
return date
|
40 |
+
# --- --- ---
|
41 |
+
|
42 |
+
|
43 |
+
# --- --- --- Set up rclone and get chips and merra files
|
44 |
+
rawdir = ''
|
45 |
+
merradir = ''
|
46 |
+
cfg = ""
|
47 |
+
result = rclone.with_config(cfg).run_cmd("ls", extra_args=[f"{rawdir}/"])
|
48 |
+
output_lines = result['out'].decode('utf-8').splitlines()
|
49 |
+
file_list = [line.split(maxsplit=1)[1] for line in output_lines if line]
|
50 |
+
result = rclone.with_config(cfg).run_cmd("ls", extra_args=[f"{merradir}/"])
|
51 |
+
output_lines = result['out'].decode('utf-8').splitlines()
|
52 |
+
merras = [line.split(maxsplit=1)[1] for line in output_lines if line]
|
53 |
+
# --- --- ---
|
54 |
+
|
55 |
+
|
56 |
+
# --- --- --- Set up paths
|
57 |
+
# Location of station tile list
|
58 |
+
station_file = './TILED_filtered_flux_sites_2018_2021.csv'
|
59 |
+
# Location to save the input file
|
60 |
+
spath = './all_inputs.csv'
|
61 |
+
odir = ''
|
62 |
+
# --- --- ---
|
63 |
+
|
64 |
+
|
65 |
+
# --- --- --- Get station information
|
66 |
+
stations = {}
|
67 |
+
fn = open(station_file, 'r')
|
68 |
+
for line in list(fn)[1:]:
|
69 |
+
dummy = line.split(',')
|
70 |
+
stations[dummy[1].strip()] = (dummy[1], float(dummy[9]), float(dummy[8]), dummy[3], dummy[4])
|
71 |
+
fn.close()
|
72 |
+
flux_nets = os.listdir("./fluxnets/flux_sites_2018_2021/")
|
73 |
+
# --- --- ---
|
74 |
+
|
75 |
+
# Locate all HLS chips
|
76 |
+
chips = sorted(file_list)
|
77 |
+
skipped = []
|
78 |
+
|
79 |
+
# Make the input file to which to save the data
|
80 |
+
out_fn = open(spath, 'w')
|
81 |
+
out_fn.write(f'Chip,Station,T2MIN,T2MAX,T2MEAN,TSMDEWMEAN,GWETROOT,LHLAND,SHLAND,SWLAND,PARDFLAND,PRECTOTLAND,GPP')
|
82 |
+
|
83 |
+
# And loop over them
|
84 |
+
for chip in chips:
|
85 |
+
rclone.with_config(cfg).copy(f"{rawdir}/{chip}", f"./{chip}")
|
86 |
+
|
87 |
+
# Match to an Ameriflux station
|
88 |
+
chip_name = chip.split('/')[-1]
|
89 |
+
tile = chip_name.split('.')[2][1:]
|
90 |
+
station_name = chip_name.split('.')[6].split("_")[0]
|
91 |
+
try: # Skip tiles for which no station exists
|
92 |
+
station = stations[station_name]
|
93 |
+
except:
|
94 |
+
print(f"exception - {('station dict indexing', station_name, tile)}")
|
95 |
+
continue
|
96 |
+
date = helpers.convert_HLS_date(chip_name)
|
97 |
+
|
98 |
+
# Locate station from tile and pull in the daily reference value
|
99 |
+
try: # Skip tiles for which no station data is available
|
100 |
+
station_file = [fluxnet for fluxnet in flux_nets if station[0] in fluxnet][0]
|
101 |
+
flux_df = pd.read_csv(f"genai-usra-east/impact/fluxnets/flux_sites_2018_2021/{station_file}")
|
102 |
+
except:
|
103 |
+
print(f"exception - {('station exception', station_name, tile)}")
|
104 |
+
continue
|
105 |
+
|
106 |
+
flux_times = np.array(flux_df.TIMESTAMP, dtype='str')
|
107 |
+
flux_gpp = np.array(flux_df.GPP_NT_VUT_REF)
|
108 |
+
try: # Skip if cannot find CO2 data
|
109 |
+
quality_flag = np.array(flux_df.NEE_VUT_REF_QC)
|
110 |
+
if quality_flag[flux_times==date.strftime("%Y%m%d")][0] >= 0.6:
|
111 |
+
co2 = flux_gpp[flux_times==date.strftime("%Y%m%d")][0]
|
112 |
+
else: # Quality not met, skip
|
113 |
+
print(f"co2 quality not met for - {('co2', station_name, tile)}")
|
114 |
+
continue
|
115 |
+
except:
|
116 |
+
print(f"co2 quality not met for - {('co2 exception', station_name, tile)}")
|
117 |
+
skipped.append(('co2 exception', station_name, tile))
|
118 |
+
continue
|
119 |
+
|
120 |
+
|
121 |
+
# Pull MERRA-2 data for temperature and dew
|
122 |
+
merra_file = [file for file in merras if "slv" in file and str(date.strftime("%Y%m%d")) in file][0]
|
123 |
+
rclone.with_config(cfg).copy(f"{merradir}/{merra_file}", f"./merra/")
|
124 |
+
merra_fn = nc.Dataset(f'./merra/{merra_file}')
|
125 |
+
|
126 |
+
# Pull in the MERRA-2 grid and find closest point
|
127 |
+
mlons = merra_fn.variables['lon'][:]
|
128 |
+
mlats = merra_fn.variables['lat'][:]
|
129 |
+
xind = np.argmin((mlons-station[1])**2)
|
130 |
+
yind = np.argmin((mlats-station[2])**2)
|
131 |
+
|
132 |
+
# Read the variables and collect stats based on time dimension
|
133 |
+
tmax = np.max(merra_fn.variables['T2M'], keepdims=True, axis=0)
|
134 |
+
tmin = np.min(merra_fn.variables['T2M'], keepdims=True, axis=0)
|
135 |
+
tmean = np.nanmean(merra_fn.variables['T2M'][:,yind, xind])
|
136 |
+
tmax = tmax[0,yind,xind]
|
137 |
+
tmin = tmin[0,yind,xind]
|
138 |
+
tdewmean = np.nanmean(merra_fn.variables['T2MDEW'][:,yind, xind])
|
139 |
+
|
140 |
+
|
141 |
+
shutil.rmtree(Path(f"./merra"))
|
142 |
+
|
143 |
+
# Pull MERRA-2 data for surface data
|
144 |
+
merra_file = [file for file in merras if "lnd" in file and str(date.strftime("%Y%m%d")) in file][0]
|
145 |
+
rclone.with_config(cfg).copy(f"{merradir}/{merra_file}", f"./merra/")
|
146 |
+
merra_fn = nc.Dataset(f'./merra/{merra_file}')
|
147 |
+
|
148 |
+
# Pull in the MERRA-2 grid and find closest point
|
149 |
+
mlons = merra_fn.variables['lon'][:]
|
150 |
+
mlats = merra_fn.variables['lat'][:]
|
151 |
+
xind = np.argmin((mlons-station[1])**2)
|
152 |
+
yind = np.argmin((mlats-station[2])**2)
|
153 |
+
|
154 |
+
# Read the variables and collect stats based on time dimension
|
155 |
+
GWETROOT = np.nanmean(merra_fn.variables['GWETROOT'][:,yind,xind])
|
156 |
+
LHLAND = np.nanmean(merra_fn.variables['LHLAND'][:,yind,xind])
|
157 |
+
SHLAND = np.nanmean(merra_fn.variables['SHLAND'][:,yind,xind])
|
158 |
+
PARDFLAND = np.nanmean(merra_fn.variables['PARDFLAND'][:,yind,xind])
|
159 |
+
PRECTOTLAND = np.nanmean(merra_fn.variables['PRECTOTLAND'][:,yind,xind])
|
160 |
+
SWLAND = np.nanmean(merra_fn.variables['SWLAND'][:,yind,xind])
|
161 |
+
|
162 |
+
shutil.rmtree(Path(f"./merra"))
|
163 |
+
shutil.rmtree(Path(f"./{tile}"))
|
164 |
+
# Save chip name, MERRA-2 values, and Ameriflux measurement to data file
|
165 |
+
out_fn.write(f'\n{chip_name},{station[0]},{tmin:.2f},{tmax:.2f},{tmean:.2f},{tdewmean:.2f},{GWETROOT:.2f},{LHLAND:.2f},{SHLAND:.2f},{SWLAND:.2f},{PARDFLAND:2f},{PRECTOTLAND:2f},{co2}')
|
166 |
+
# Close the file
|
167 |
+
out_fn.close()
|
168 |
+
rclone.with_config(cfg).copy(f"{spath}", f"{odir}")
|
169 |
+
print("DONE")
|