Improve function pert_scaled

dpgen/data/gen.py

@@ -726,8 +726,7 @@ def make_scale_ABACUS(jdata):
 
 
 def pert_scaled(jdata):
- if "init_fp_style" not in jdata:
- jdata["init_fp_style"] = "VASP"
+ ### Extract data from jdata
  out_dir = jdata["out_dir"]
  scale = jdata["scale"]
  pert_box = jdata["pert_box"]
@@ -744,6 +743,7 @@ def pert_scaled(jdata):
  if "from_poscar" in jdata:
  from_poscar = jdata["from_poscar"]
 
+ ### Get the current working directory and the system path
  cwd = os.getcwd()
  path_sp = os.path.join(out_dir, global_dirname_03)
  assert os.path.isdir(path_sp)
@@ -752,35 +752,38 @@ def pert_scaled(jdata):
  sys_pe.sort()
  os.chdir(cwd)
 
- pert_cmd = os.path.dirname(__file__)
- pert_cmd = os.path.join(pert_cmd, "tools")
- pert_cmd = os.path.join(pert_cmd, "create_random_disturb.py")
- fp_style = "vasp"
- poscar_name = "POSCAR"
- if jdata["init_fp_style"] == "ABACUS":
+ ### Construct the perturbation command
+ init_fp_style = jdata.get("init_fp_style", "VASP")
+ if init_fp_style == "VASP":
+  fp_style = "vasp"
+  poscar_name = "POSCAR"
+ elif init_fp_style == "ABACUS":
  fp_style = "abacus"
  poscar_name = "STRU"
+
+ python_exec = os.path.join(
+ os.path.dirname(__file__), "tools", "create_random_disturb.py"
+ )
  pert_cmd = (
  sys.executable
- + " "
- + pert_cmd
- + " -etmax %f -ofmt %s %s %d %f > /dev/null"
- % (pert_box, fp_style, poscar_name, pert_numb, pert_atom)
+ + f" {python_exec} -etmax {pert_box} -ofmt {fp_style} {poscar_name} {pert_numb} {pert_atom} > /dev/null"
  )
+
+ ### Loop over each system and scale
  for ii in sys_pe:
  for jj in scale:
- path_work = path_sp
- path_work = os.path.join(path_work, ii)
- path_work = os.path.join(path_work, f"scale-{jj:.3f}")
+ path_work = os.path.join(path_sp, ii, f"scale-{jj:.3f}")
  assert os.path.isdir(path_work)
  os.chdir(path_work)
  sp.check_call(pert_cmd, shell=True)
+
+ ### Loop over each perturbation
  for kk in range(pert_numb):
  if fp_style == "vasp":
- pos_in = "POSCAR%d.vasp" % (kk + 1)
+ pos_in = f"POSCAR{kk}.vasp"
  elif fp_style == "abacus":
- pos_in = "STRU%d.abacus" % (kk + 1)
- dir_out = "%06d" % (kk + 1)
+ pos_in = f"STRU{kk}.abacus"
+ dir_out = f"{kk:06d}"
  create_path(dir_out)
  if fp_style == "vasp":
  pos_out = os.path.join(dir_out, "POSCAR")
@@ -805,35 +808,7 @@ def pert_scaled(jdata):
  else:
  shutil.copy2(pos_in, pos_out)
  os.remove(pos_in)
- kk = -1
- if fp_style == "vasp":
- pos_in = "POSCAR"
- elif fp_style == "abacus":
- pos_in = "STRU"
- dir_out = "%06d" % (kk + 1)
- create_path(dir_out)
- if fp_style == "vasp":
- pos_out = os.path.join(dir_out, "POSCAR")
- elif fp_style == "abacus":
- pos_out = os.path.join(dir_out, "STRU")
- if not from_poscar:
- if fp_style == "vasp":
- poscar_shuffle(pos_in, pos_out)
- elif fp_style == "abacus":
- stru_in = get_abacus_STRU(pos_in)
- stru_out = shuffle_stru_data(stru_in)
- with open(pos_out, "w") as fp:
- fp.write(
- make_abacus_scf_stru(
- stru_out,
- pp_file,
- orb_file_names,
- dpks_descriptor_name,
- type_map=jdata["elements"],
- )
- )
- else:
- shutil.copy2(pos_in, pos_out)
+
  os.chdir(cwd)
 
 
@@ -866,7 +841,7 @@ def make_vasp_md(jdata, mdata):
 
  for ii in sys_ps:
  for jj in scale:
- for kk in range(pert_numb + 1):
+ for kk in range(pert_numb):
  path_work = path_md
  path_work = os.path.join(path_work, ii)
  path_work = os.path.join(path_work, f"scale-{jj:.3f}")
@@ -987,7 +962,7 @@ def make_abacus_md(jdata, mdata):
 
  for ii in sys_ps:
  for jj in scale:
- for kk in range(pert_numb + 1):
+ for kk in range(pert_numb):
  path_work = path_md
  path_work = os.path.join(path_work, ii)
  path_work = os.path.join(path_work, f"scale-{jj:.3f}")

dpgen/data/surf.py

@@ -23,7 +23,7 @@
 import dpgen.data.tools.fcc as fcc
 import dpgen.data.tools.hcp as hcp
 import dpgen.data.tools.sc as sc
-from dpgen import ROOT_PATH, dlog
+from dpgen import dlog
 from dpgen.dispatcher.Dispatcher import make_submission_compat
 from dpgen.generator.lib.utils import symlink_user_forward_files
 from dpgen.remote.decide_machine import convert_mdata
@@ -501,26 +501,25 @@ def pert_scaled(jdata):
  sys_pe.sort()
  os.chdir(cwd)
 
+ python_exec = os.path.join(
+ os.path.dirname(__file__), "tools", "create_random_disturb.py"
+ )
  pert_cmd = (
  sys.executable
- + " "
- + os.path.join(ROOT_PATH, "data/tools/create_random_disturb.py")
- )
- pert_cmd += " -etmax %f -ofmt vasp POSCAR %d %f > /dev/null" % (
- pert_box,
- pert_numb,
- pert_atom,
+ + f" {python_exec} -etmax {pert_box} -ofmt vasp POSCAR {pert_numb} {pert_atom} > /dev/null"
  )
+
+ ### Loop over each system and scale
  for ii in sys_pe:
  for jj in scale:
- path_scale = path_sp
- path_scale = os.path.join(path_scale, ii)
- path_scale = os.path.join(path_scale, f"scale-{jj:.3f}")
+ path_scale = os.path.join(path_sp, ii, f"scale-{jj:.3f}")
  assert os.path.isdir(path_scale)
  os.chdir(path_scale)
  dlog.info(os.getcwd())
  poscar_in = os.path.join(path_scale, "POSCAR")
  assert os.path.isfile(poscar_in)
+
+ ### Loop over each perturbation
  for ll in elongs:
  path_elong = path_scale
  path_elong = os.path.join(path_elong, f"elong-{ll:3.3f}")
@@ -529,18 +528,13 @@ def pert_scaled(jdata):
  poscar_elong(poscar_in, "POSCAR", ll)
  sp.check_call(pert_cmd, shell=True)
  for kk in range(pert_numb):
- pos_in = "POSCAR%d.vasp" % (kk + 1)
- dir_out = "%06d" % (kk + 1)
+ pos_in = f"POSCAR{kk}.vasp"
+ dir_out = f"{kk:06d}"
  create_path(dir_out)
  pos_out = os.path.join(dir_out, "POSCAR")
  poscar_shuffle(pos_in, pos_out)
  os.remove(pos_in)
- kk = -1
- pos_in = "POSCAR"
- dir_out = "%06d" % (kk + 1)
- create_path(dir_out)
- pos_out = os.path.join(dir_out, "POSCAR")
- poscar_shuffle(pos_in, pos_out)
+
  os.chdir(cwd)
 
 

dpgen/data/tools/create_random_disturb.py

@@ -123,7 +123,8 @@ def create_disturbs_ase_dev(
  cell0 = atoms.get_cell()
 
  # creat nfile ofmt files.
- for fid in range(1, nfile + 1):
+ # for fid in range(1, nfile + 1): # C.Thang Nguyen: Why not start from 0?
+ for fid in range(nfile):
  # Use copy(), otherwise it will modify the input atoms every time.
  atoms_d = atoms.copy()
 
@@ -194,7 +195,7 @@ def create_disturbs_abacus_dev(
  cell0 = stru["cells"]
 
  # creat nfile ofmt files.
- for fid in range(1, nfile + 1):
+ for fid in range(nfile):
  # Use copy(), otherwise it will modify the input atoms every time.
  stru_d = stru.copy()
 

tests/data/test_gen_bulk.py

@@ -69,7 +69,7 @@ def test(self):
  self.assertEqual(len(scales), self.scale_numb)
  for scale in scales:
  perts = glob.glob(os.path.join(scale, "[0-9]*"))
- self.assertEqual(len(perts), self.pert_numb + 1)
+ self.assertEqual(len(perts), self.pert_numb)
 
 
 if __name__ == "__main__":

tests/data/test_gen_bulk_abacus.py

@@ -103,7 +103,7 @@ def testSTRU(self):
  self.assertEqual(len(scales), self.scale_numb)
  for scale in scales:
  perts = glob.glob(os.path.join(scale, "[0-9]*"))
- self.assertEqual(len(perts), self.pert_numb + 1)
+ self.assertEqual(len(perts), self.pert_numb)
 
 
 if __name__ == "__main__":