Use ResourceResolvers in lammps/vasp/sphinx

.ci_support/environment-old.yml

@@ -13,8 +13,8 @@ dependencies:
 - pandas =2.0.3
 - phonopy =2.20.0
 - pint =0.18
-- pyiron_base =0.9.4
-- pyiron_snippets =0.1.1
+- pyiron_base =0.9.12
+- pyiron_snippets =0.1.3
 - pylammpsmpi =0.2.18
 - pyscal3 =3.2.5
 - monty =2024.3.31

pyiron_atomistics/atomistics/job/potentials.py

@@ -8,9 +8,11 @@
 """
 
 import os
+from abc import ABC, abstractmethod
 
 import pandas
 from pyiron_base import state
+from pyiron_snippets.resources import ResourceNotFound, ResourceResolver
 
 __author__ = "Martin Boeckmann, Jan Janssen"
 __copyright__ = (
@@ -24,7 +26,7 @@
 __date__ = "Sep 1, 2017"
 
 
-class PotentialAbstract(object):
+class PotentialAbstract(ABC):
     """
     The PotentialAbstract class loads a list of available potentials and sorts them. Afterwards the potentials can be
     accessed through:
@@ -36,6 +38,14 @@ class PotentialAbstract(object):
         selected_atoms:
     """
 
+    @property
+    @abstractmethod
+    def resource_plugin_name(self) -> str:
+        """Return the name of the folder of this plugin/code in the pyiron resources.
+
+        One of lammps/vasp/sphinx, to be overriden in the specific sub classes."""
+        pass
+
     def __init__(self, potential_df, default_df=None, selected_atoms=None):
         self._potential_df = potential_df
         self._default_df = default_df
@@ -102,108 +112,75 @@ def __getitem__(self, item):
     def __str__(self):
         return str(self.list())
 
-    @staticmethod
-    def _get_potential_df(plugin_name, file_name_lst):
+    @classmethod
+    def _get_resolver(cls):
+        """Return a ResourceResolver that can be searched for potential files or potential dataframes.
+
+        This exists primarily so that the lammps and sphinx sub classes can overload it to add their conda package
+        specific resource paths.
+
+        Returns:
+            :class:`.ResourceResolver`
+        """
+        return ResourceResolver(
+            state.settings.resource_paths,
+            cls.resource_plugin_name,
+            "potentials",
+        )
+
+    @classmethod
+    def _get_potential_df(cls, file_name_lst):
         """
 
         Args:
-            plugin_name (str):
             file_name_lst (set):
 
         Returns:
             pandas.DataFrame:
         """
         env = os.environ
-        resource_path_lst = state.settings.resource_paths
-        for conda_var in ["CONDA_PREFIX", "CONDA_DIR"]:
-            if conda_var in env.keys():  # support iprpy-data package
-                path_to_add = os.path.join(env[conda_var], "share", "iprpy")
-                if path_to_add not in resource_path_lst:
-                    resource_path_lst += [path_to_add]
-        df_lst = []
-        for resource_path in resource_path_lst:
-            if os.path.exists(os.path.join(resource_path, plugin_name, "potentials")):
-                resource_path = os.path.join(resource_path, plugin_name, "potentials")
-            if "potentials" in resource_path or "iprpy" in resource_path:
-                for path, folder_lst, file_lst in os.walk(resource_path):
-                    for periodic_table_file_name in file_name_lst:
-                        if (
-                            periodic_table_file_name in file_lst
-                            and periodic_table_file_name.endswith(".csv")
-                        ):
-                            df_lst.append(
-                                pandas.read_csv(
-                                    os.path.join(path, periodic_table_file_name),
-                                    index_col=0,
-                                    converters={
-                                        "Species": lambda x: x.replace("'", "")
-                                        .strip("[]")
-                                        .split(", "),
-                                        "Config": lambda x: x.replace("'", "")
-                                        .replace("\\n", "\n")
-                                        .strip("[]")
-                                        .split(", "),
-                                        "Filename": lambda x: x.replace("'", "")
-                                        .strip("[]")
-                                        .split(", "),
-                                    },
-                                )
-                            )
-        if len(df_lst) > 0:
-            return pandas.concat(df_lst)
-        else:
-            raise ValueError("Was not able to locate the potential files.")
 
-    @staticmethod
+        def read_csv(path):
+            return pandas.read_csv(
+                path,
+                index_col=0,
+                converters={
+                    "Species": lambda x: x.replace("'", "").strip("[]").split(", "),
+                    "Config": lambda x: x.replace("'", "")
+                    .replace("\\n", "\n")
+                    .strip("[]")
+                    .split(", "),
+                    "Filename": lambda x: x.replace("'", "").strip("[]").split(", "),
+                },
+            )
+
+        files = cls._get_resolver().search(file_name_lst)
+        return pandas.concat(map(read_csv, files), ignore_index=True)
+
+    @classmethod
     def _get_potential_default_df(
-        plugin_name,
+        cls,
         file_name_lst={"potentials_vasp_pbe_default.csv"},
     ):
         """
 
         Args:
-            plugin_name (str):
             file_name_lst (set):
 
         Returns:
             pandas.DataFrame:
         """
-        for resource_path in state.settings.resource_paths:
-            pot_path = os.path.join(resource_path, plugin_name, "potentials")
-            if os.path.exists(pot_path):
-                resource_path = pot_path
-            if "potentials" in resource_path:
-                for path, folder_lst, file_lst in os.walk(resource_path):
-                    for periodic_table_file_name in file_name_lst:
-                        if (
-                            periodic_table_file_name in file_lst
-                            and periodic_table_file_name.endswith(".csv")
-                        ):
-                            return pandas.read_csv(
-                                os.path.join(path, periodic_table_file_name),
-                                index_col=0,
-                            )
-                        elif (
-                            periodic_table_file_name in file_lst
-                            and periodic_table_file_name.endswith(".h5")
-                        ):
-                            return pandas.read_hdf(
-                                os.path.join(path, periodic_table_file_name), mode="r"
-                            )
-        raise ValueError("Was not able to locate the potential files.")
-
-
-def find_potential_file_base(path, resource_path_lst, rel_path):
-    if path is not None:
-        for resource_path in resource_path_lst:
-            path_direct = os.path.join(resource_path, path)
-            path_indirect = os.path.join(resource_path, rel_path, path)
-            if os.path.exists(path_direct):
-                return path_direct
-            elif os.path.exists(path_indirect):
-                return path_indirect
-    raise ValueError(
-        "Either the filename or the functional has to be defined.",
-        path,
-        resource_path_lst,
-    )
+        try:
+            return pandas.read_csv(
+                cls._get_resolver().first(file_name_lst), index_col=0
+            )
+        except ResourceNotFound:
+            raise ValueError("Was not able to locate the potential files.") from None
+
+    @classmethod
+    def find_potential_file(cls, path):
+        res = cls._get_resolver()
+        try:
+            return res.first(path)
+        except ResourceNotFound:
+            raise ValueError(f"Could not find file '{path}' in {res}!") from None

pyiron_atomistics/lammps/potential.py

@@ -8,11 +8,9 @@
 
 import pandas as pd
 from pyiron_base import GenericParameters, state
+from pyiron_snippets.resources import ResourceResolver
 
-from pyiron_atomistics.atomistics.job.potentials import (
-    PotentialAbstract,
-    find_potential_file_base,
-)
+from pyiron_atomistics.atomistics.job.potentials import PotentialAbstract
 from pyiron_atomistics.atomistics.structure.atoms import Atoms
 
 __author__ = "Joerg Neugebauer, Sudarsan Surendralal, Jan Janssen"
@@ -66,6 +64,7 @@ def remove_structure_block(self):
 
     @property
     def files(self):
+        env = os.environ
         if len(self._df["Filename"].values[0]) > 0 and self._df["Filename"].values[
             0
         ] != [""]:
@@ -77,27 +76,11 @@ def files(self):
                 for files in list(self._df["Filename"])[0]
                 if not os.path.isabs(files)
             ]
-            env = os.environ
-            resource_path_lst = state.settings.resource_paths
-            for conda_var in ["CONDA_PREFIX", "CONDA_DIR"]:
-                if conda_var in env.keys():  # support iprpy-data package
-                    path_to_add = state.settings.convert_path_to_abs_posix(
-                        os.path.join(env[conda_var], "share", "iprpy")
-                    )
-                    if path_to_add not in resource_path_lst:
-                        resource_path_lst.append(path_to_add)
             for path in relative_file_paths:
                 absolute_file_paths.append(
-                    find_potential_file_base(
-                        path=path,
-                        resource_path_lst=resource_path_lst,
-                        rel_path=os.path.join("lammps", "potentials"),
-                    )
+                    LammpsPotentialFile.find_potential_file(path)
                 )
-            if len(absolute_file_paths) != len(list(self._df["Filename"])[0]):
-                raise ValueError("Was not able to locate the potentials.")
-            else:
-                return absolute_file_paths
+            return absolute_file_paths
 
     def copy_pot_files(self, working_directory):
         if self.files is not None:
@@ -251,10 +234,26 @@ class LammpsPotentialFile(PotentialAbstract):
         selected_atoms:
     """
 
+    resource_plugin_name = "lammps"
+
+    @classmethod
+    def _get_resolver(cls):
+        env = os.environ
+        return (
+            super()
+            ._get_resolver()
+            .chain(
+                ResourceResolver(
+                    [env[var] for var in ("CONDA_PREFIX", "CONDA_DIR") if var in env],
+                    "share",
+                    "iprpy",
+                )
+            )
+        )
+
     def __init__(self, potential_df=None, default_df=None, selected_atoms=None):
         if potential_df is None:
             potential_df = self._get_potential_df(
-                plugin_name="lammps",
                 file_name_lst={"potentials_lammps.csv"},
             )
         super(LammpsPotentialFile, self).__init__(

pyiron_atomistics/sphinx/base.py

@@ -38,9 +38,6 @@
     collect_spins_dat,
 )
 from pyiron_atomistics.sphinx.potential import SphinxJTHPotentialFile
-from pyiron_atomistics.sphinx.potential import (
-    find_potential_file as find_potential_file_jth,
-)
 from pyiron_atomistics.sphinx.structure import read_atoms
 from pyiron_atomistics.sphinx.util import sxversions
 from pyiron_atomistics.sphinx.volumetric_data import SphinxVolumetricData
@@ -49,9 +46,6 @@
     VaspPotentialSetter,
     strip_xc_from_potential_name,
 )
-from pyiron_atomistics.vasp.potential import (
-    find_potential_file as find_potential_file_vasp,
-)
 
 __author__ = "Osamu Waseda, Jan Janssen"
 __copyright__ = (
@@ -1293,11 +1287,9 @@ def _get_potential_path(
 
         if potformat == "JTH":
             potentials = SphinxJTHPotentialFile(xc=xc)
-            find_potential_file = find_potential_file_jth
             pot_path_dict.setdefault("PBE", "jth-gga-pbe")
         elif potformat == "VASP":
             potentials = VaspPotentialFile(xc=xc)
-            find_potential_file = find_potential_file_vasp
             pot_path_dict.setdefault("PBE", "paw-gga-pbe")
             pot_path_dict.setdefault("LDA", "paw-lda")
         else:
@@ -1313,7 +1305,7 @@ def _get_potential_path(
             if "pseudo_potcar_file" in species_obj.tags.keys():
                 new_element = species_obj.tags["pseudo_potcar_file"]
                 potentials.add_new_element(parent_element=elem, new_element=new_element)
-                potential_path = find_potential_file(
+                potential_path = potentials.find_potential_file(
                     path=potentials.find_default(new_element)["Filename"].values[0][0]
                 )
                 assert os.path.isfile(
@@ -1327,14 +1319,14 @@ def _get_potential_path(
                     potentials.add_new_element(
                         parent_element=elem, new_element=new_element
                     )
-                    potential_path = find_potential_file(
+                    potential_path = potentials.find_potential_file(
                         path=potentials.find_default(new_element)["Filename"].values[0][
                             0
                         ]
                     )
             else:
                 ori_paths.append(
-                    find_potential_file(
+                    potentials.find_potential_file(
                         path=potentials.find_default(elem)["Filename"].values[0][0]
                     )
                 )