scripts to generate si_diffusion_small dataset and fix to data prepro…

crystal_diffusion/data/diffusion/data_preprocess.py

@@ -66,7 +66,7 @@ def _convert_coords_to_relative(row: pd.Series) -> List[float]:
         """
         x_lim, y_lim, z_lim = row['box']
         coord_red = [coord for triple in zip(row['x'], row['y'], row['z']) for coord in
-                     (triple[0] / x_lim, triple[1] / y_lim, triple[2] / z_lim)]
+                     ((triple[0] / x_lim) % 1, (triple[1] / y_lim) % 1, (triple[2] / z_lim) % 1)]
         return coord_red
 
     def get_x_relative(self, df: pd.DataFrame) -> pd.DataFrame:

data/si_diffusion_small/create_data.sh

@@ -0,0 +1,40 @@
+#!/bin/bash
+
+TEMPERATURE=300
+BOX_SIZE=2
+STEP=10000
+CROP=10000
+NTRAIN_RUN=10
+NVALID_RUN=5
+
+NRUN=$(($NTRAIN_RUN + $NVALID_RUN))
+
+for SEED in $(seq 2 $NRUN);
+do
+  if [ "$SEED" -le $NTRAIN_RUN ]; then
+    MODE="train"
+  else
+    MODE="valid"
+  fi
+  echo $MODE $SEED
+  mkdir -p "${MODE}_run_${SEED}"
+  cd "${MODE}_run_${SEED}"
+  lmp < ../in.si.lammps -v STEP $(($STEP + $CROP)) -v T $TEMPERATURE -v S $BOX_SIZE -v SEED $SEED
+
+  # extract the thermodynamic outputs in a yaml file
+  egrep  '^(keywords:|data:$|---$|\.\.\.$|  - \[)' log.lammps > thermo_log.yaml
+
+  mkdir -p "uncropped_outputs"
+  mv "dump.si-${TEMPERATURE}-${BOX_SIZE}.yaml" uncropped_outputs/
+  mv thermo_log.yaml uncropped_outputs/
+
+  tail -n 760076 uncropped_outputs/dump.si-300-2.yaml > lammps_dump.yaml
+
+  python ../../crop_lammps_outputs.py \
+      --lammps_yaml "uncropped_outputs/dump.si-${TEMPERATURE}-${BOX_SIZE}.yaml" \
+      --lammps_thermo "uncropped_outputs/thermo_log.yaml" \
+      --crop $CROP \
+      --output_dir ./
+
+  cd ..
+done

data/si_diffusion_small/in.si.lammps

@@ -0,0 +1,29 @@
+log log.lammps
+
+units           metal
+atom_style      atomic
+atom_modify     map array
+
+lattice         diamond 5.43
+region          simbox block 0 ${S} 0 ${S} 0 ${S}
+create_box      1 simbox
+create_atoms    1 region simbox
+
+mass 1 28.0855
+
+group Si type 1
+
+pair_style sw
+pair_coeff * * ../../si.sw Si
+
+velocity all create ${T} ${SEED}
+
+dump 1 all yaml 1 dump.si-${T}-${S}.yaml id type x y z fx fy fz
+
+thermo_style yaml
+thermo 1
+#==========================Output files========================
+
+fix 1 all nvt temp ${T} ${T} 0.01
+run ${STEP}
+unfix 1

tests/data/diffusion/test_data_preprocess.py

@@ -99,9 +99,9 @@ def sample_coordinates(box_coordinates):
     # Sample data frame
     return pd.DataFrame({
         'box': [box_coordinates],
-        'x': [[60, 6, 0.6, 0.06]],
-        'y': [[120, 12, 1.2, 0.12]],
-        'z': [[180, 18, 1.8, 0.18]]
+        'x': [[0.6, 0.06, 0.006, 0.00006]],
+        'y': [[1.2, 0.12, 0.0012, 0.00012]],
+        'z': [[1.8, 0.18, 0.018, 0.0018]]
     })