-
Notifications
You must be signed in to change notification settings - Fork 3
/
dvc.yaml
304 lines (304 loc) · 13.6 KB
/
dvc.yaml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
stages:
csv-cif-no-spin-500-data:
foreach:
- WSe2
- MoS2
do:
cmd: echo "Generate ${item}_500 dataset in csv+cif format from VASP output..." && python scripts/vasp_to_csv_cif.py --input-vasp "datasets/raw_vasp/high_density_defects/${item}_500"
--input-structures-list "datasets/POSCARs/${item}" --POSCARs-in-input-list
--output-csv-cif "${paths.datasets_root}datasets/csv_cif/high_density_defects/${item}_500"
--pristine-folder datasets/others/pristine_high_density
deps:
- datasets/POSCARs/${item}
- datasets/others/pristine_high_density
- datasets/raw_vasp/high_density_defects/${item}_500
- scripts/vasp_to_csv_cif.py
outs:
- ${paths.datasets_root}datasets/csv_cif/high_density_defects/${item}_500
csv-cif-spin-500-data:
foreach:
- BP
- GaSe
- hBN
- InSe
do:
cmd: echo "Generate ${item}_500 dataset in csv+cif format from VASP output..." && python scripts/vasp_to_csv_cif.py --input-vasp "datasets/raw_vasp/high_density_defects/${item}_spin_500"
--input-structures-list "datasets/POSCARs/${item}" --POSCARs-in-input-list
--output-csv-cif "${paths.datasets_root}datasets/csv_cif/high_density_defects/${item}_spin_500"
--pristine-folder datasets/others/pristine_high_density --separate-spins
deps:
- datasets/POSCARs/${item}
- datasets/others/pristine_high_density
- datasets/raw_vasp/high_density_defects/${item}_spin_500
- scripts/vasp_to_csv_cif.py
outs:
- ${paths.datasets_root}datasets/csv_cif/high_density_defects/${item}_spin_500
csv-cif-vacancy-pairs-no-spin:
foreach:
- GaSe
- InSe
- hBN
do:
cmd: echo "Generate ${item}_500 dataset in csv+cif format from VASP output..." && python scripts/vasp_to_csv_cif.py --input-vasp datasets/raw_vasp/vacancy_pairs/${item}
--input-structures-list "datasets/POSCARs/vacancy_pairs/${item}" --POSCARs-in-input-list
--output-csv-cif "${paths.datasets_root}datasets/csv_cif/vacancy_pairs/${item}"
--pristine-folder datasets/others/pristine_high_density
deps:
- datasets/POSCARs/vacancy_pairs/${item}
- datasets/others/pristine_high_density
- datasets/raw_vasp/vacancy_pairs/${item}
- scripts/vasp_to_csv_cif.py
outs:
- ${paths.datasets_root}datasets/csv_cif/vacancy_pairs/${item}
csv-cif-spin-500-data-occu-tol-investigation:
foreach:
- material: BP
occu_tol: 1e-8
- material: GaSe
occu_tol: 1e-8
- material: hBN
occu_tol: 1e-8
- material: InSe
occu_tol: 1e-8
- material: BP
occu_tol: 0.05
- material: GaSe
occu_tol: 0.05
- material: hBN
occu_tol: 0.05
- material: InSe
occu_tol: 0.05
- material: BP
occu_tol: 0.1
- material: GaSe
occu_tol: 0.1
- material: hBN
occu_tol: 0.1
- material: InSe
occu_tol: 0.1
do:
cmd: echo "Generate ${item}_500 dataset in csv+cif format from VASP output..." && python scripts/vasp_to_csv_cif.py --input-vasp "datasets/raw_vasp/high_density_defects/${item.material}_spin_500"
--input-structures "datasets/POSCARs/${item.material}" --output-csv-cif "${paths.datasets_root}datasets/csv_cif/high_density_defects/${item.material}_spin_500_occu_tol_${item.occu_tol}"
--pristine-folder datasets/others/pristine_high_density --separate-spins --band-occupancy-tolerence ${item.occu_tol}
deps:
- datasets/POSCARs/${item.material}
- datasets/others/pristine_high_density
- datasets/raw_vasp/high_density_defects/${item.material}_spin_500
- scripts/vasp_to_csv_cif.py
outs:
- ${paths.datasets_root}datasets/csv_cif/high_density_defects/${item.material}_spin_500_occu_tol_${item.occu_tol}
csv-cif-low-density-8x8-Innopolis-v1:
foreach:
- WSe2
- MoS2
do:
cmd: echo "Creating technical files needed to preserve the historical structure indexing for low density defects ${item} dataset..." && python scripts/separate_csv_cif.py --input-folder datasets/csv_cif/dichalcogenides_x1s6_202109
--output-folder ${paths.datasets_root}datasets/csv_cif/low_density_defects_Innopolis-v1/${item} --base-material ${item}
--supercell-size 8
deps:
- datasets/csv_cif/dichalcogenides_x1s6_202109
- scripts/separate_csv_cif.py
outs:
- datasets/csv_cif/low_density_defects_Innopolis-v1/${item}
tar-index-low-density-8x8:
foreach:
- material: WSe2
tar_folder: wse2
- material: MoS2
tar_folder: mos2
do:
cmd:
ratarmount -o modules=subdir,subdir=${item.tar_folder}_8x8_5933
datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.tar_folder}_8x8_5933.tar.gz
datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.material}_8x8_5933/ &&
sleep 5 &&
ratarmount -u datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.material}_8x8_5933/
deps:
- datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.tar_folder}_8x8_5933.tar.gz
outs:
- datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.tar_folder}_8x8_5933.tar.gz.index.sqlite
csv-cif-low-density-8x8:
foreach:
- material: WSe2
tar_folder: wse2
- material: MoS2
tar_folder: mos2
do:
cmd:
echo "Generate low_density_defects/${item.material} dataset in csv+cif format from VASP output..." && ratarmount -o modules=subdir,subdir=${item.tar_folder}_8x8_5933
datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.tar_folder}_8x8_5933.tar.gz
${paths.datasets_root}datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.material}_8x8_5933/ &&
python scripts/vasp_to_csv_cif.py --input-vasp "${paths.datasets_root}datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.material}_8x8_5933"
--input-structures-list datasets/csv_cif/low_density_defects_Innopolis-v1/${item.material}
--input-structures-csv-cif datasets/csv_cif/low_density_defects_Innopolis-v1/${item.material}
--output-csv-cif "${paths.datasets_root}datasets/csv_cif/low_density_defects/${item.material}"
--pristine-folder datasets/others/pristine_high_density
--vasprun-glob-prefix poscar_??- &&
sleep 5 &&
ratarmount -u datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.material}_8x8_5933/
deps:
- scripts/vasp_to_csv_cif.py
- datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.tar_folder}_8x8_5933.tar.gz
- datasets/csv_cif/low_density_defects_Innopolis-v1/${item.material}
- datasets/others/pristine_high_density
- datasets/raw_vasp/dichalcogenides8x8_vasp_nus_202110/${item.tar_folder}_8x8_5933.tar.gz.index.sqlite
outs:
- ${paths.datasets_root}datasets/csv_cif/low_density_defects/${item.material}
processed-low-density:
foreach:
- WSe2
- MoS2
do:
cmd: echo "Converting the structures from csv+cif format to the platform-specific pickle format for low density ${item} dataset..." && python scripts/parse_csv_cif.py --input-name=low_density_defects/${item} --fill-missing-band-properties --normalize-homo-lumo ${paths.output_folder_arg}
deps:
- datasets/csv_cif/low_density_defects/${item}
- scripts/parse_csv_cif.py
outs:
- ${paths.datasets_root}datasets/processed/low_density_defects/${item}/data.pickle.gz
- ${paths.datasets_root}datasets/processed/low_density_defects/${item}/targets.csv.gz
processed-high-density:
foreach:
- BP_spin
- GaSe_spin
- hBN_spin
- InSe_spin
- MoS2
- WSe2
do:
cmd: echo "Converting the structures from csv+cif format to the platform-specific pickle format for high density ${item} dataset..." && python scripts/parse_csv_cif.py --input-name=high_density_defects/${item}_500 --normalize-homo-lumo --fill-missing-band-properties ${paths.output_folder_arg}
deps:
- scripts/parse_csv_cif.py
- datasets/csv_cif/high_density_defects/${item}_500
outs:
- ${paths.datasets_root}datasets/processed/high_density_defects/${item}_500/targets.csv.gz
- ${paths.datasets_root}datasets/processed/high_density_defects/${item}_500/data.pickle.gz
processed-pilot:
cmd: python scripts/parse_csv_cif.py --input-name=pilot --fill-missing-band-properties ${paths.output_folder_arg}
deps:
- scripts/parse_csv_cif.py
- datasets/csv_cif/pilot
outs:
- ${paths.datasets_root}datasets/processed/pilot/targets.csv.gz
- ${paths.datasets_root}datasets/processed/pilot/data.pickle.gz
processed-vacancy-pairs:
foreach:
- hBN
- GaSe
- InSe
do:
cmd: python scripts/parse_csv_cif.py --input-name=vacancy_pairs/${item} --normalize-homo-lumo --fill-missing-band-properties ${paths.output_folder_arg}
deps:
- scripts/parse_csv_cif.py
- datasets/csv_cif/vacancy_pairs/${item}
outs:
- ${paths.datasets_root}datasets/processed/vacancy_pairs/${item}/targets.csv.gz
- ${paths.datasets_root}datasets/processed/vacancy_pairs/${item}/data.pickle.gz
rolos-2d-materials-point-defects:
cmd:
./scripts/pack_for_rolos.sh
deps:
- scripts/pack_for_rolos.sh
#- datasets/processed/high_density_defects/{BP_spin,GaSe_spin,hBN_spin,InSe_spin,MoS2,WSe2}_500/targets.csv.gz
#- datasets/csv_cif/high_density_defects/{BP_spin,GaSe_spin,hBN_spin,InSe_spin,MoS2,WSe2}_500
- datasets/processed/low_density_defects
- datasets/csv_cif/low_density_defects
outs:
- datasets/others/rolos/2d-materials-point-defects
matminer:
foreach:
- high_density_defects/BP_spin_500
- high_density_defects/GaSe_spin_500
- high_density_defects/hBN_spin_500
- high_density_defects/InSe_spin_500
- high_density_defects/MoS2_500
- high_density_defects/WSe2_500
- low_density_defects/MoS2
- low_density_defects/WSe2
- vacancy_pairs/hBN
- vacancy_pairs/GaSe
- vacancy_pairs/InSe
- pilot
do:
cmd: echo "Creating matminer descriptors for ${item}, to be used with CatBoost" && python scripts/compute_matminer_features.py --input-name ${item} --n-proc=${matminer_cpus} ${paths.output_folder_arg}
deps:
- datasets/csv_cif/${item}
- scripts/compute_matminer_features.py
outs:
- ${paths.datasets_root}datasets/processed/${item}/matminer.csv.gz
matminer-debug:
cmd: python scripts/compute_matminer_features.py --input-name pilot --n-proc=2 --debug ${paths.output_folder_arg}
deps:
- datasets/csv_cif/pilot
- scripts/compute_matminer_features.py
outs:
- ${paths.datasets_root}datasets/processed/pilot/matminer_dbg.csv.gz
MoS2-2Mo-structures:
foreach:
- 1_V_1W_Mo
- 2V_Mo
- 2W_Mo
do:
# No clean output as the UUIDs will be regenerated
cmd: python scripts/defects_generation.py --output-path datasets/POSCARs/MoS2-2Mo/${item}
--config-path datasets/generation_configs/MoS2-2Mo/${item}.yaml
deps:
- datasets/generation_configs/MoS2-2Mo/${item}.yaml
- defects_generation/molecules/MoS2.cif
- scripts/defects_generation.py
outs:
- datasets/POSCARs/MoS2-2Mo/${item}
vacancy-pairs-configs:
foreach:
- GaSe
- InSe
- MoS2
- P
do:
# No clean output as the UUIDs will be regenerated
cmd: python scripts/defects_generation.py --output-path datasets/POSCARs/vacancy_pairs/${item}
--config-path datasets/generation_configs/vacancy_pairs/${item}.yaml
deps:
- datasets/generation_configs/vacancy_pairs/${item}.yaml
- defects_generation/molecules/${item}.cif
- scripts/defects_generation.py
outs:
- datasets/POSCARs/vacancy_pairs/${item}
vacancy-pairs-configs-hBN:
cmd: python scripts/defects_generation.py --output-path datasets/POSCARs/vacancy_pairs/hBN
--config-path datasets/generation_configs/vacancy_pairs/hBN.yaml
deps:
- datasets/generation_configs/vacancy_pairs/hBN.yaml
- scripts/defects_generation.py
# And this BN/hBN difference is why we have a separate entity
- defects_generation/molecules/BN.cif
outs:
- datasets/POSCARs/vacancy_pairs/hBN
vacancy-pairs-experiments:
foreach:
- vacancy_pairs
- vacancy_pairs_GaSe
do:
cmd: python scripts/folds_from_config.py --experiment-name ${item}
deps:
- scripts/folds_from_config.py
- datasets/experiments/${item}/config.yaml
outs:
- datasets/experiments/${item}/folds.csv.gz
train-only-split:
cmd: python scripts/prepare_data_split.py
--datasets high_density_defects/{BP_spin_500,GaSe_spin_500,hBN_spin_500,InSe_spin_500,MoS2_500,WSe2_500}
low_density_defects/{MoS2,WSe2} --test-size 0 --experiment-name combined_mixed_all_train --n-folds 2
--targets homo_lumo_gap_min formation_energy_per_site --strategy train_test
deps:
- scripts/prepare_data_split.py
- datasets/csv_cif/high_density_defects/BP_spin_500
- datasets/csv_cif/high_density_defects/GaSe_spin_500
- datasets/csv_cif/high_density_defects/hBN_spin_500
- datasets/csv_cif/high_density_defects/InSe_spin_500
- datasets/csv_cif/high_density_defects/MoS2_500
- datasets/csv_cif/high_density_defects/WSe2_500
- datasets/csv_cif/low_density_defects/MoS2
- datasets/csv_cif/low_density_defects/WSe2
outs:
- datasets/experiments/combined_mixed_all_train/folds.csv.gz
- datasets/experiments/combined_mixed_all_train/config.yaml