VASP Bugfixes: Volumetric data and XDATCAR parsing for monatomic stru…

…ctures (#4104)

* fix parsing of monatomic xdatcars / clean up parsing code

* precommit

* make mypy happy

* fix volumetric bug with newer POTCAR hashing

* ruffn

* fix ubuntu-latest test to run on 3.12 as gh workflow indicates, stop trying to run 3.13

* bump down ubuntu-latest to 3.11, mypy too aggressive

* change lint version to 3.11 explicitly

* pin mypy < 1.12 for linting wf

.github/workflows/lint.yml

@@ -18,11 +18,11 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v5
         with:
-          python-version: '3.x'
+          python-version: '3.11'
 
       - name: Install dependencies
         run: |
-          pip install --upgrade ruff mypy pyright
+          pip install --upgrade ruff 'mypy<1.12' pyright
 
       - name: ruff
         run: |

.github/workflows/test.yml

@@ -39,7 +39,7 @@ jobs:
             resolution: highest
             extras: ci,optional,numpy-v1 # Test NP1 on Windows (quite buggy ATM)
           - os: ubuntu-latest
-            python: ">3.10"
+            python: "3.12"
             resolution: lowest-direct
             extras: ci,optional
           - os: macos-latest

src/pymatgen/io/vasp/inputs.py

@@ -350,7 +350,18 @@ def from_str(
 
         except ValueError:
             vasp5_symbols = True
-            symbols: list[str] = [symbol.split("/")[0] for symbol in lines[5].split()]
+
+            # In VASP 6.x.x, part of the POTCAR hash is written to POSCAR-style strings
+            # In VASP 6.4.2 and up, the POTCAR symbol is also written, ex.:
+            # ```MgSi
+            # 1.0
+            # -0.000011    4.138704    0.000002
+            # -2.981238    2.069353    3.675251
+            # 2.942054    2.069351    4.865237
+            # Mg_pv/f474ac0d  Si/79d9987ad87```
+            # whereas older VASP 5.x.x POSCAR strings would just have `Mg Si` on the last line
+
+            symbols: list[str] = [symbol.split("/")[0].split("_")[0] for symbol in lines[5].split()]
 
             # Atoms and number of atoms in POSCAR written with VASP appear on
             # multiple lines when atoms of the same type are not grouped together

src/pymatgen/io/vasp/outputs.py

@@ -4358,68 +4358,64 @@ def __init__(
         coords_str: list = []
         structures: list = []
         preamble_done: bool = False
+        parse_poscar: bool = False
+        num_sites: int | None = None
+        restart_preamble: bool = False
         if ionicstep_start < 1:
             raise ValueError("Start ionic step cannot be less than 1")
         if ionicstep_end is not None and ionicstep_end < 1:
             raise ValueError("End ionic step cannot be less than 1")
 
+        file_len = sum(1 for _ in zopen(filename, mode="rt"))
         ionicstep_cnt = 1
+        ionicstep_start = ionicstep_start or 0
         with zopen(filename, mode="rt") as file:
             title = None
-            for line in file:
+            for iline, line in enumerate(file):
                 line = line.strip()
                 if preamble is None:
                     preamble = [line]
                     title = line
-                elif title == line:
+
+                elif title == line and len(coords_str) > 0:
+                    # sometimes the title is the same as the only chemical species in the structure
+                    # only enter this block if the coords have been read
+                    parse_poscar = True
+                    restart_preamble = True
                     preamble_done = False
-                    poscar = Poscar.from_str("\n".join([*preamble, "Direct", *coords_str]))
-                    if ionicstep_end is None:
-                        if ionicstep_cnt >= ionicstep_start:
-                            structures.append(poscar.structure)
-                    else:
-                        if ionicstep_start <= ionicstep_cnt < ionicstep_end:
-                            structures.append(poscar.structure)
-                        if ionicstep_cnt >= ionicstep_end:
-                            break
-                    ionicstep_cnt += 1
-                    coords_str = []
-                    preamble = [line]
+
                 elif not preamble_done:
                     if line == "" or "Direct configuration=" in line:
                         preamble_done = True
-                        tmp_preamble = [preamble[0]]
-                        for i in range(1, len(preamble)):
-                            if preamble[0] != preamble[i]:
-                                tmp_preamble.append(preamble[i])
-                            else:
-                                break
-                        preamble = tmp_preamble
                     else:
                         preamble.append(line)
-                elif line == "" or "Direct configuration=" in line:
+
+                elif line == "" or "Direct configuration=" in line and len(coords_str) > 0:
+                    parse_poscar = True
+                    restart_preamble = False
+                else:
+                    coords_str.append(line)
+
+                if (parse_poscar and (num_sites is None or len(coords_str) == num_sites)) or iline == file_len - 1:
+                    if num_sites is None:
+                        num_sites = len(coords_str)
+
                     poscar = Poscar.from_str("\n".join([*preamble, "Direct", *coords_str]))
-                    if ionicstep_end is None:
-                        if ionicstep_cnt >= ionicstep_start:
-                            structures.append(poscar.structure)
-                    else:
-                        if ionicstep_start <= ionicstep_cnt < ionicstep_end:
-                            structures.append(poscar.structure)
-                        if ionicstep_cnt >= ionicstep_end:
-                            break
+                    if (ionicstep_end is None and ionicstep_cnt >= ionicstep_start) or (
+                        ionicstep_end is not None and ionicstep_start <= ionicstep_cnt < ionicstep_end
+                    ):
+                        structures.append(poscar.structure)
+                    elif (ionicstep_end is not None) and ionicstep_cnt >= ionicstep_end:
+                        break
+
                     ionicstep_cnt += 1
                     coords_str = []
-                else:
-                    coords_str.append(line)
+                    parse_poscar = False
+                    if restart_preamble:
+                        preamble = [line]
 
             if preamble is None:
                 raise ValueError("preamble is None")
-            poscar = Poscar.from_str("\n".join([*preamble, "Direct", *coords_str]))
-            if (
-                (ionicstep_end is None and ionicstep_cnt >= ionicstep_start)
-                or ionicstep_start <= ionicstep_cnt < ionicstep_end  # type: ignore[operator]
-            ):
-                structures.append(poscar.structure)
 
         self.structures = structures
         self.comment = comment or self.structures[0].formula

tests/io/vasp/test_outputs.py

@@ -1452,6 +1452,12 @@ def test_init(self):
         l2 = Locpot(poscar=poscar, data=data, data_aug=None)
         assert l2.data_aug == {}
 
+    def test_vasp_6x_style(self):
+        filepath = f"{VASP_OUT_DIR}/LOCPOT.vasp642.gz"
+        locpot = Locpot.from_file(filepath)
+        assert locpot.dim == (2, 2, 5)
+        assert {str(ele) for ele in locpot.structure.composition} == {"Mg", "Si"}
+
 
 class TestChgcar(PymatgenTest):
     @classmethod
@@ -1718,6 +1724,10 @@ def test_init(self):
 
         assert structures[0].lattice != structures[-1].lattice
 
+        xdatcar = Xdatcar(f"{VASP_OUT_DIR}/XDATCAR_monatomic.gz")
+        assert len(xdatcar.structures) == 10
+        assert all(len(structure.composition) == 1 for structure in xdatcar.structures)
+
 
 class TestDynmat:
     def test_init(self):