rsspolymlp 0.4.3__tar.gz → 0.4.4__tar.gz

{rsspolymlp-0.4.3/src/rsspolymlp.egg-info → rsspolymlp-0.4.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: rsspolymlp
-Version: 0.4.3
+Version: 0.4.4
 Summary: A framework for random structure search using polynomial MLPs
 Author-email: Hayato Wakai <wakai@cms.mtl.kyoto-u.ac.jp>
 Maintainer-email: Hayato Wakai <wakai@cms.mtl.kyoto-u.ac.jp>
@@ -79,11 +79,6 @@ pip install rsspolymlp
    - Global RSS with polynomial MLPs
    - Unique structure identification and RSS result summarization
    - Phase stability analysis
- - [Development kit for polynomial MLPs](docs/rsspolymlp_dev.md)
-   - MLP dataset generation
-   - DFT dataset division
-   - Polynomial MLP development
-   - Pareto-optimal MLP selection
  - Python API
    - [RSS workflow](docs/api_rsspolymlp.md)
    - [VASP calculation utilities](src/rsspolymlp/utils/vasp_util/readme.md)

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/README.md

@@ -48,11 +48,6 @@ pip install rsspolymlp
    - Global RSS with polynomial MLPs
    - Unique structure identification and RSS result summarization
    - Phase stability analysis
- - [Development kit for polynomial MLPs](docs/rsspolymlp_dev.md)
-   - MLP dataset generation
-   - DFT dataset division
-   - Polynomial MLP development
-   - Pareto-optimal MLP selection
  - Python API
    - [RSS workflow](docs/api_rsspolymlp.md)
    - [VASP calculation utilities](src/rsspolymlp/utils/vasp_util/readme.md)

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "rsspolymlp"
-version = "0.4.3"
+version = "0.4.4"
 description = "A framework for random structure search using polynomial MLPs"
 license = "MIT"
 authors = [

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/analysis/rss_summarize.py

@@ -6,17 +6,18 @@ import xml.etree.ElementTree as ET
 from collections import defaultdict
 from pathlib import Path
 from time import time
-from typing import Optional
+from typing import Literal, Optional
 
 import numpy as np
-from pypolymlp.utils.vasp_utils import write_poscar_file
 
+from pypolymlp.utils.vasp_utils import write_poscar_file
 from rsspolymlp.analysis.ghost_minima import detect_ghost_minima
 from rsspolymlp.analysis.struct_matcher.gen_struct_rep import (
     StructRepGenerator,
     StructReps,
 )
 from rsspolymlp.analysis.struct_matcher.identify_uniqstr import (
+    UniqueStructIdentifier,
     identify_unique_structure,
 )
 from rsspolymlp.common.composition import compute_composition
@@ -37,6 +38,7 @@ class RSSResultSummarizer:
         element_order: list = None,
         num_process: int = -1,
         backend: str = "loky",
+        assign_method: Literal["redrep", "greedy", "HK", "pymatgen"] = "redrep",
         primitive_symprecs: list[float] = [1e-5, 1e-4, 1e-3, 1e-2],
         output_poscar: bool = False,
         thresholds: list[float] = None,
@@ -50,6 +52,7 @@ class RSSResultSummarizer:
         self.element_order = element_order
         self.num_process = num_process
         self.backend = backend
+        self.assign_method = assign_method
         self.primitive_symprecs = primitive_symprecs
         self.output_poscar = output_poscar
         self.thresholds = thresholds
@@ -89,6 +92,11 @@ class RSSResultSummarizer:
             axis_tol = 0.1
             pos_tol = 0.1
 
+        self.analyzer = UniqueStructIdentifier(
+            axis_tol=axis_tol,
+            pos_tol=pos_tol,
+            assign_method=self.assign_method,
+        )
         for comp_tag, res_paths in json_paths_c.items():
             print(f"Composition {comp_tag}: summarizing...")
             time_start = time()
@@ -116,6 +124,7 @@ class RSSResultSummarizer:
                             rss_results_c[comp_tag][res_path],
                             axis_tol=axis_tol,
                             pos_tol=pos_tol,
+                            assign_method=self.assign_method,
                             pre_analyzer=self.analyzer,
                         )
 
@@ -124,11 +133,11 @@ class RSSResultSummarizer:
             unique_structs = self.analyzer.unique_str
             unique_str_prop = self.analyzer.unique_str_prop
             enthalpies = np.array([s["energy"] for s in unique_str_prop])
-            distances = np.array([s["least_distance"] for s in unique_str_prop])
             sort_idx = np.argsort(enthalpies)
 
             if not self.parse_vasp:
                 os.makedirs("ghost_minima", exist_ok=True)
+                distances = np.array([s["least_distance"] for s in unique_str_prop])
                 is_ghost_minima, ghost_minima_info = detect_ghost_minima(
                     enthalpies[sort_idx], distances[sort_idx]
                 )
@@ -209,6 +218,7 @@ class RSSResultSummarizer:
                     keep_unique=True,
                     axis_tol=axis_tol,
                     pos_tol=pos_tol,
+                    assign_method=self.assign_method,
                     pre_analyzer=self.analyzer,
                 )
             time_finish = time() - time_start
@@ -220,7 +230,7 @@ class RSSResultSummarizer:
             sort_idx = np.argsort(enthalpies)
             sort_idx = sorted(
                 sort_idx,
-                key=lambda i: len(unique_str_sorted[i]),
+                key=lambda i: len(unique_structs[i]),
                 reverse=True,
             )
             unique_str_sorted = [unique_structs[i] for i in sort_idx]
@@ -260,12 +270,12 @@ class RSSResultSummarizer:
 
         print("   - Converting reduced crystal structure representation...")
         generator = StructRepGenerator(
-            num_process=self.num_process,
-            backend=self.backend,
             primitive_symprecs=self.primitive_symprecs,
         )
         struct_reps_list = generator.generate_struct_reps_parallel(
             rss_results,
+            num_process=self.num_process,
+            backend=self.backend,
         )
         if self.update_parent:
             print("   - Eliminating duplicate structures...")
@@ -586,6 +596,14 @@ def log_all_unique_structures(
     with open(file_name, "a") as f:
         print("unique_structures:", file=f)
         for idx1, _str in enumerate(unique_structs):
+            if unique_str_prop is not None:
+                sorted_data = sorted(
+                    zip(unique_structs[idx1], unique_str_prop[idx1]),
+                    key=lambda x: x[1]["energy"],
+                )
+                unique_structs[idx1], unique_str_prop[idx1] = map(
+                    list, zip(*sorted_data)
+                )
             print(f"  - struct_No: {idx1 + 1}", file=f)
             print("    structures:", file=f)
             for idx2, _str in enumerate(unique_structs[idx1]):
@@ -597,7 +615,9 @@ def log_all_unique_structures(
                 objprop = PropUtil(polymlp_st.axis.T, polymlp_st.positions.T)
                 print(f"    - sub_struct_No: '{idx1 + 1}_{idx2 + 1}'", file=f)
                 print(f"      struct_path: {_str.struct_path}", file=f)
-                print(f"      pressure: {_str.pressure}", file=f)
+                print(
+                    f"      pressure: {unique_str_prop[idx1][idx2]['pressure']}", file=f
+                )
                 print(f"      enthalpy: {e}", file=f)
                 print(f"      axis: {objprop.abc}", file=f)
                 print(
@@ -627,7 +647,7 @@ def log_all_unique_structures(
             if unique_other_prop is not None:
                 print("    properties:", unique_other_prop[idx1], file=f)
 
-    comp_res = compute_composition(unique_structs[0][0].original_structure.elements)
+    comp_res = compute_composition(unique_structs[0][0].original_polymlp_st.elements)
 
     rss_result_all = {
         "elements": comp_res.unique_elements.tolist(),

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/analysis/struct_matcher/identify_uniqstr.py

@@ -18,7 +18,7 @@ class UniqueStructIdentifier:
         self,
         axis_tol: float = 0.01,
         pos_tol: float = 0.01,
-        assign_method: Literal["redrep", "greedy", "HK"] = "redrep",
+        assign_method: Literal["redrep", "greedy", "HK", "pymatgen"] = "redrep",
         keep_unique: bool = False,
         verbose: bool = False,
     ):
@@ -43,6 +43,10 @@ class UniqueStructIdentifier:
         self.assign_method = assign_method
         self.keep_unique = keep_unique
         self.verbose = verbose
+        if self.assign_method == "pymatgen":
+            from rsspolymlp.utils.pymatgen_utils import PymatUtil
+
+            self.pymat = PymatUtil()
 
         self.unique_str: list[StructReps] = []  # Store unique structures
         self.unique_str_prop: list[dict] = []  # Store unique structure properties
@@ -94,6 +98,12 @@ class UniqueStructIdentifier:
                         axis_tol=self.axis_tol,
                         pos_tol=self.pos_tol,
                     )
+                elif self.assign_method == "pymatgen":
+                    st1 = self.pymat.polymlp_to_pymat_st(uniq_str.original_polymlp_st)
+                    st2 = self.pymat.polymlp_to_pymat_st(
+                        struct_reps.original_polymlp_st
+                    )
+                    judge_duplicate = self.pymat.match_str(st1, st2)
                 else:
                     judge_duplicate = struct_match_allrep(
                         uniq_str,
@@ -171,7 +181,7 @@ def identify_unique_structure(
     struct_list: list[dict],
     axis_tol: float = 0.01,
     pos_tol: float = 0.01,
-    assign_method: Literal["redrep", "greedy", "HK"] = "redrep",
+    assign_method: Literal["redrep", "greedy", "HK", "pymatgen"] = "redrep",
     keep_unique: bool = False,
     num_process: int = -1,
     backend: str = "loky",

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/api/cli_rsspolymlp.py

@@ -211,6 +211,13 @@ def run():
         default=None,
         help="List of element symbols",
     )
+    parser.add_argument(
+        "--assign_method",
+        choices=["redrep", "greedy", "HK", "pymatgen"],
+        default="redrep",
+        help="Method for matching atomic positions: "
+        "reduced representation, greedy, or Hopcroft-Karp",
+    )
     parser.add_argument(
         "--primitive_symprecs",
         nargs="*",
@@ -314,6 +321,7 @@ def run():
             element_order=args.element_order,
             num_process=args.num_process,
             backend=args.backend,
+            assign_method=args.assign_method,
             primitive_symprecs=args.primitive_symprecs,
             output_poscar=args.output_poscar,
             thresholds=args.thresholds,

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/api/cli_rsspolymlp_utils.py

@@ -43,7 +43,7 @@ def run():
     )
     parser.add_argument(
         "--assign_method",
-        choices=["redrep", "greedy", "HK"],
+        choices=["redrep", "greedy", "HK", "pymatgen"],
         default="redrep",
         help="Method for matching atomic positions: "
         "reduced representation, greedy, or Hopcroft-Karp",

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/api/rsspolymlp.py

@@ -1,6 +1,7 @@
 import glob
 import os
 import time
+from typing import Literal
 
 from rsspolymlp.analysis.ghost_minima import (
     detect_actual_ghost_minima,
@@ -182,6 +183,7 @@ def rss_summarize(
     element_order: list = None,
     num_process=-1,
     backend="loky",
+    assign_method: Literal["redrep", "greedy", "HK", "pymatgen"] = "redrep",
     primitive_symprecs: list[float] = [1e-5, 1e-4, 1e-3, 1e-2],
     output_poscar: bool = False,
     thresholds: list[float] = None,
@@ -195,6 +197,7 @@ def rss_summarize(
         element_order=element_order,
         num_process=num_process,
         backend=backend,
+        assign_method=assign_method,
         primitive_symprecs=primitive_symprecs,
         output_poscar=output_poscar,
         thresholds=thresholds,

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/api/rsspolymlp_utils.py

@@ -1,7 +1,6 @@
 from typing import Literal, Optional
 
 from pypolymlp.utils.spglib_utils import SymCell
-
 from rsspolymlp.analysis.rss_summarize import (
     log_all_unique_structures,
     log_unique_structures,
@@ -23,7 +22,7 @@ def struct_matcher(
     struct_reps_list: Optional[list[dict]] = None,
     axis_tol: float = 0.05,
     pos_tol: float = 0.03,
-    assign_method: Literal["redrep", "greedy", "HK"] = "redrep",
+    assign_method: Literal["redrep", "greedy", "HK", "pymatgen"] = "redrep",
     num_process: int = -1,
     backend: str = "loky",
     primitive_symprecs: list[float] = [1e-5, 1e-4, 1e-3, 1e-2],
@@ -148,7 +147,7 @@ def struct_compare(
     poscar_paths,
     axis_tol: float = 0.01,
     pos_tol: float = 0.01,
-    assign_method: Literal["redrep", "greedy", "HK"] = "redrep",
+    assign_method: Literal["redrep", "greedy", "HK", "pymatgen"] = "redrep",
     refine_symprec: float = 1e-3,
     reduced_symprecs: list[float] = None,
     normalize_axis: bool = False,

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/common/atomic_energy.py

@@ -26,8 +26,10 @@ def atomic_energy(element):
         "P": -1.88664411,
         "Pb": -0.58498619,
         "Rb": -0.14189970,
+        "S": -0.94100986,
         "Sb": -1.43005001,
         "Sc": -1.98188993,
+        "Se": -0.76510990,
         "Si": -0.87162900,
         "Sn": -0.48437348,  # dftu s-orbital 1.0 eV, dft: -0.64316709
         "Sr": -0.02805273,

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/common/property.py

@@ -1,8 +1,8 @@
 import numpy as np
 
 from pypolymlp.core.data_format import PolymlpStructure
-from pypolymlp.utils.spglib_utils import SymCell
 from rsspolymlp.common.composition import compute_composition
+from rsspolymlp.utils.spglib_utils import SymCell
 
 
 class PropUtil:

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/mlp_dev/dataset/split_dataset.py

@@ -32,7 +32,8 @@ def parse_vasp_results(elements, vasprun_paths):
     for vasprun_path in vasprun_paths:
         try:
             structure, (energy, force, stress) = parse_properties_from_vasprun(
-                vasprun_path, units="GPa",
+                vasprun_path,
+                units="GPa",
             )
         except Exception:
             print(vasprun_path, "failed")
@@ -119,8 +120,8 @@ def split_dataset(
             e_tag = ""
 
         # Filter by stress tensor components
-        if max_stress > threshold_s_large * 10 or (
-            threshold_s_small is not None and min_stress < threshold_s_small * 10
+        if max_stress > threshold_s_large or (
+            threshold_s_small is not None and min_stress < threshold_s_small
         ):
             vasprun_dict[f"s_large{e_tag}"].append(vasprun_path)
             return

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/utils/matplot_util/make_plot.py

@@ -230,14 +230,17 @@ class MakePlot:
                     rasterized=rasterized,
                 )
 
-    def ax_hist(self, dist, bins=50, range=None):
+    def ax_hist(self, dist, bins=50, range=None, density=False, fraction=False):
         if range is None:
             range = [min(dist), max(dist)]
+        weights = np.ones(len(dist)) / len(dist) if fraction else None
         self.ax.hist(
             dist,
             bins=bins,
             range=range,
             color=self.color,
+            density=density,
+            weights=weights,
         )
 
     def finalize_ax(
@@ -288,7 +291,9 @@ class MakePlot:
         """
         self._set_axis_scale(xlog, ylog, xlog_grid, ylog_grid)
         self._set_axis_limits(x_limits, x_limits_hide, y_limits, y_limits_hide)
-        self._set_axis_grid(x_limits, y_limits, x_grid, y_grid)
+        self._set_axis_grid(
+            x_limits, y_limits, x_grid, y_grid, x_limits_hide, y_limits_hide
+        )
         self._set_axis_labels(xlabel, ylabel, title)
         self._set_twiny(twiny, x_limits, x_grid)
         if legend_ncol:
@@ -323,7 +328,9 @@ class MakePlot:
         if y_limits_hide:
             self.ax.tick_params(labelleft=False)
 
-    def _set_axis_grid(self, x_limits, y_limits, x_grid, y_grid):
+    def _set_axis_grid(
+        self, x_limits, y_limits, x_grid, y_grid, x_limits_hide, y_limits_hide
+    ):
         """Set grid spacing for axes."""
         if x_grid:
             if x_grid[1]:
@@ -341,6 +348,10 @@ class MakePlot:
                     np.append(np.arange(*y_limits, y_grid[0]), y_limits[1]), 6
                 )
                 self.ax.set_yticks(y_values)
+        if x_limits_hide:
+            self.ax.set_xticks([])
+        if y_limits_hide:
+            self.ax.set_yticks([])
 
     def _set_axis_labels(self, xlabel, ylabel, title):
         """Set labels and title for axes."""

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4}/src/rsspolymlp/utils/spglib_utils.py

@@ -117,7 +117,11 @@ class SymCell:
 
     def get_spacegroup(self):
         """Return space group."""
-        return spglib.get_spacegroup(self.cell, symprec=self.symprec)
+        try:
+            spg = spglib.get_spacegroup(self.cell, symprec=self.symprec)
+        except spglib._spglib.SpglibCppError:
+            spg = None
+        return spg
 
     def get_spacegroup_multiple_prec(self, symprecs=[1e-2, 1e-3, 1e-4, 1e-5]):
         """Return list of space groups using multiple precisions."""

{rsspolymlp-0.4.3 → rsspolymlp-0.4.4/src/rsspolymlp.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: rsspolymlp
-Version: 0.4.3
+Version: 0.4.4
 Summary: A framework for random structure search using polynomial MLPs
 Author-email: Hayato Wakai <wakai@cms.mtl.kyoto-u.ac.jp>
 Maintainer-email: Hayato Wakai <wakai@cms.mtl.kyoto-u.ac.jp>
@@ -79,11 +79,6 @@ pip install rsspolymlp
    - Global RSS with polynomial MLPs
    - Unique structure identification and RSS result summarization
    - Phase stability analysis
- - [Development kit for polynomial MLPs](docs/rsspolymlp_dev.md)
-   - MLP dataset generation
-   - DFT dataset division
-   - Polynomial MLP development
-   - Pareto-optimal MLP selection
  - Python API
    - [RSS workflow](docs/api_rsspolymlp.md)
    - [VASP calculation utilities](src/rsspolymlp/utils/vasp_util/readme.md)