Feature/pacemaker implementation

.devcontainer/devcontainer.json

@@ -14,5 +14,5 @@
         "ms-azuretools.vscode-docker"]
     },
   },
-  "postCreateCommand": "pip cache purge && uv pip install --prerelease=allow -e .[strict,docs] && pre-commit install",
+  "postCreateCommand": "pip cache purge && uv pip install --prerelease=allow -e '.[strict,docs,tests]' && pre-commit install",
 }

.github/workflows/docs.yml

@@ -42,6 +42,7 @@ jobs:
             uv pip install --prerelease=allow .[docs,strict,tests]
             uv pip install --upgrade monty
             uv pip install numpy==1.26.4
+
 
       - name: Copy tutorials
         run: |

.github/workflows/pacemaker-test.yml

@@ -0,0 +1,48 @@
+# This workflow will install pacemaker and its dependencies, run tests and lint with Python 3.10
+
+name: Testing Pacemaker (Python 3.10)
+
+on:
+  workflow_dispatch:
+  push:
+    branches: [ main ]
+  pull_request:
+    branches: [ main ]
+
+jobs:
+  pacemaker-test:
+    runs-on: ubuntu-latest
+    env:
+      PYTHON_VERSION: "3.10"
+
+    steps:
+    - name: Fetch Most Recent Docker Image Tag
+      run: |
+        TAG=$(curl -H "Accept: application/vnd.github.v3+json" \
+            -H "Authorization: token ${{ secrets.GITHUB_TOKEN }}" \
+            https://api.github.com/orgs/autoatml/packages/container/autoplex%2Fautoplex-python-3.10/versions \
+            | jq -r 'sort_by(.created_at) | reverse | .[0].metadata.container.tags[0]')
+        echo "VERSION=$TAG" >> $GITHUB_ENV
+
+    - uses: actions/checkout@v4
+      with:
+        fetch-depth: 0
+
+    - name: Run Pacemaker Tests
+      run: |
+        docker pull ghcr.io/autoatml/autoplex/autoplex-python-3.10:${{ env.VERSION }}
+
+        docker run --rm \
+          -v ${{ github.workspace }}:/workspace \
+          -w /workspace \
+          ghcr.io/autoatml/autoplex/autoplex-python-3.10:${{ env.VERSION }} \
+          bash -c "
+          git config --global --add safe.directory /workspace && \
+          python -m pip install --upgrade pip && \
+          python -m uv cache clean && \
+          python -m uv pip install 'protobuf<3.20' tensorflow==2.8.0 && \
+          python -m uv pip install --no-deps git+https://github.com/ICAMS/TensorPotential.git@1e44b2558356800ae070658c0bb856ff9bf74538 && \
+          python -m uv pip install --no-deps git+https://github.com/ICAMS/python-ace.git@d1c213a7d9c5b809a3ae83b2e5a916be26d921f0 && \
+          python -m uv pip install --prerelease=allow '.[strict,tests,aims]' && \
+          OMP_NUM_THREADS=1 pytest --cache-clear -vv -k 'PACE'
+        "

.github/workflows/python-package.yml

@@ -40,6 +40,7 @@ jobs:
     - name: Run tests using Docker image for Python ${{ matrix.python-version }}
       run: |
         docker pull ghcr.io/autoatml/autoplex/autoplex-python-${{ matrix.python-version }}:${{ env.VERSION }}
+
         docker run --rm \
           -v ${{ github.workspace }}:/workspace \
           -w /workspace \
@@ -48,9 +49,10 @@ jobs:
           git config --global --add safe.directory /workspace && \
           python -m pip install --upgrade pip && \
           python -m uv cache clean && \
-          python -m uv pip install --prerelease=allow .[strict,tests,aims] && \
-          OMP_NUM_THREADS=1 pytest --cache-clear --cov=autoplex --cov-report term-missing --cov-append --splits 5 --group ${{ matrix.split }} -vv --durations-path /workspace/tests/test_data/.pytest-split-durations --store-durations 
+          python -m uv pip install --prerelease=allow '.[strict,tests,aims]' && \
+          OMP_NUM_THREADS=1 pytest --cache-clear --cov=autoplex --cov-report term-missing --cov-append --splits 5 --group ${{ matrix.split }} -vv --durations-path /workspace/tests/test_data/.pytest-split-durations --store-durations -k 'not PACE'
         "
+
 
     - name: Upload test durations artifact
       if: matrix.python-version == '3.10'

README.md

@@ -91,8 +91,21 @@ Once installed in the terminal, run the following commands to get Julia ACEpoten
 julia -e 'using Pkg; Pkg.Registry.add("General"); Pkg.Registry.add(Pkg.Registry.RegistrySpec(url="https://github.com/ACEsuit/ACEregistry")); Pkg.add(Pkg.PackageSpec(;name="ACEpotentials", version="0.6.7")); Pkg.add("DataFrames"); Pkg.add("CSV")'
 ```
 
+> ℹ️ To fit and validate `Pacemaker ACE` potentials, one also needs to install `tensorflow`, `tensorpotential`, and `python-ace`.
+> Please note that Pacemaker ACE fitting can be run on both CPU and GPU.
+> ⚠️ Please also note on versioning: to prevent dependency conflicts (e.g., with `pandas` versions) and ensure stability, please install the exact commit hashes listed below using the `--no-deps` flag. These specific versions have been fully tested and validated for use with Autoplex.
+
+```bash
+pip install autoplex[pacemaker]
+pip install tensorflow==2.8.0
+pip install --no-deps git+https://github.com/ICAMS/TensorPotential.git@1e44b2558356800ae070658c0bb856ff9bf74538
+# Ensure CMake is available before running this:
+pip install --no-deps git+https://github.com/ICAMS/python-ace.git@d1c213a7d9c5b809a3ae83b2e5a916be26d921f0
+```
+
 > ℹ️ To fit and validate NEP potentials, one requires an Nvidia GPU card with compute capability no less than 3.5 and CUDA toolkit 9.0 or newer. This potential can only be trained on GPU only and currently interface to NEP potential training is provided via [calorine](https://calorine.materialsmodeling.org/) package that uses `nep` executable from the [GPUMD](https://gpumd.org/index.html) package. To get this executable please follow the compilation instructions [here](https://gpumd.org/installation.html) and add this executable to the system path.
 
+
 ## Enabling RSS workflows
 
 Additionally, `buildcell` as a part of `AIRSS` needs to be installed if one wants to use the RSS functionality:

pyproject.toml

@@ -90,6 +90,14 @@ strict = [
 dev = ["pre-commit>=2.12.1"]
 tests = ["pytest", "pytest-mock", "pytest-split", "pytest-cov", "types-setuptools", "nbmake"]
 aims = ["pyfhiaims"]
+pacemaker = [
+     "numpy==1.26.4",
+     "pandas>=2.2",
+     "scipy",
+     "ruamel.yaml",
+     "psutil",
+     "scikit-learn==1.4.2",
+]
 
 [tool.setuptools_scm]
 

src/autoplex/__init__.py

@@ -16,6 +16,7 @@
     MLIPHypers,
     NEPSettings,
     NEQUIPSettings,
+    PacemakerSettings,
     RssConfig,
 )
 
@@ -27,3 +28,4 @@
 MACE_HYPERS = MACESettings()
 NEQUIP_HYPERS = NEQUIPSettings()
 NEP_HYPERS = NEPSettings()
+PACEMAKER_HYPERS = PacemakerSettings()

src/autoplex/auto/rss/flows.py

@@ -222,7 +222,7 @@ def make(self, **kwargs):
         pre_database_dir: str | None
             Directory where the previous database was saved.
         mlip_type: str
-            Choose one specific MLIP type to be fitted: 'GAP' | 'J-ACE' | 'NEQUIP' | 'M3GNET' | 'MACE'.
+            Choose one specific MLIP type to be fitted: 'GAP' | 'J-ACE' | 'P-ACE' | 'NEQUIP' | 'M3GNET' | 'MACE'.
             Default is 'GAP'.
         ref_energy_name: str
             Reference energy name. Default is 'REF_energy'.

src/autoplex/auto/rss/jobs.py

@@ -95,7 +95,9 @@ def initial_rss(
     force_max: float | None = None,
     force_label: str = "REF_forces",
     pre_database_dir: str | None = None,
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"] = "GAP",
+    mlip_type: Literal[
+        "GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"
+    ] = "GAP",
     ref_energy_name: str = "REF_energy",
     ref_force_name: str = "REF_forces",
     ref_virial_name: str = "REF_virial",
@@ -197,7 +199,7 @@ def initial_rss(
         The label of force values to use for distillation. Default is 'REF_forces'.
     pre_database_dir: str | None
         Directory where the previous database was saved. Default is None.
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
+    mlip_type: Literal["GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
         Choose one specific MLIP type to be fitted. Default is 'GAP'.
     ref_energy_name: str
         Reference energy name. Default is 'REF_energy'.
@@ -366,7 +368,9 @@ def do_rss_iterations(
     distillation: bool = True,
     force_max: float = 200,
     force_label: str = "REF_forces",
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"] = "GAP",
+    mlip_type: Literal[
+        "GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"
+    ] = "GAP",
     ref_energy_name: str = "REF_energy",
     ref_force_name: str = "REF_forces",
     ref_virial_name: str = "REF_virial",
@@ -508,7 +512,7 @@ def do_rss_iterations(
         Maximum force value to exclude structures. Default is 200.
     force_label: str
         The label of force values to use for distillation. Default is 'REF_forces'.
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
+    mlip_type: Literal["GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
         Choose one specific MLIP type to be fitted. Default is 'GAP'.
     ref_energy_name: str
         Reference energy name. Default is 'REF_energy'.

src/autoplex/data/rss/jobs.py

@@ -424,7 +424,7 @@ def _parallel_process(
 
 @job
 def do_rss_single_node(
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"],
+    mlip_type: Literal["GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"],
     mlip_path: str | Path,
     iteration_index: str,
     structures: list[Structure],
@@ -451,7 +451,7 @@ def do_rss_single_node(
 
     Parameters
     ----------
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
+    mlip_type: Literal["GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
         Choose one specific MLIP type to be fitted.
     mlip_path: str | Path
         Path to the MLIP model.
@@ -527,7 +527,7 @@ def do_rss_single_node(
 
 @job
 def do_rss_multi_node(
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"],
+    mlip_type: Literal["GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"],
     mlip_path: str | Path,
     iteration_index: str,
     structure: list[Structure] | list[list[Structure]] | None = None,
@@ -555,7 +555,7 @@ def do_rss_multi_node(
 
     Parameters
     ----------
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
+    mlip_type: Literal["GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
         Choose one specific MLIP type to be fitted.
     mlip_path: str | Path
         Path to the MLIP model.

src/autoplex/data/rss/utils.py

@@ -333,7 +333,7 @@ def process_rss(
         ASE Atoms object representing the atomic configuration.
     mlip_type: str
         Choose one specific MLIP type:
-        'GAP' | 'J-ACE' | 'NequIP' | 'M3GNet' | 'MACE'.
+        'GAP' | 'J-ACE' | 'P-ACE' | 'NequIP' | 'M3GNet' | 'MACE'.
     mlip_path: str | Path
         Path to the MLIP model.
     output_file_name: str
@@ -420,6 +420,29 @@ def process_rss(
             lmpcmds=cmds, atom_types=atom_types, log_file="test.log", keep_alive=True
         )
 
+    elif mlip_type == "P-ACE":
+
+        mlip_path_obj = Path(mlip_path)
+        potential_file = None
+
+        if mlip_path_obj.is_file():
+            potential_file = mlip_path_obj
+        else:
+            target_file = mlip_path_obj / "output_potential.yaml"
+            if target_file.exists():
+                potential_file = target_file
+
+        if potential_file is None:
+            raise FileNotFoundError(
+                f"Could not find 'output_potential.yaml' in {mlip_path} for P-ACE."
+            )
+
+        from autoplex.fitting.common.utils import (  # noqa: PLC0415
+            AutoplexPyACECalculator,
+        )
+
+        pot = AutoplexPyACECalculator(basis_set=str(potential_file.resolve()))
+
     elif mlip_type == "NEQUIP":
         nequip_label = os.path.join(mlip_path, "deployed_nequip_model.pth")
         if isolated_atom_energies:
@@ -539,7 +562,7 @@ def build_traj():
 
 
 def minimize_structures(
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"],
+    mlip_type: Literal["GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"],
     mlip_path: str | Path,
     iteration_index: str,
     structures: list[Structure],
@@ -566,7 +589,7 @@ def minimize_structures(
 
     Parameters
     ----------
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
+    mlip_type: Literal["GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
         Choose one specific MLIP type to be fitted.
     mlip_path: str | Path
         Path to the MLIP model.

src/autoplex/fitting/common/flows.py

@@ -42,7 +42,7 @@ class MLIPFitMaker(Maker):
     ----------
     name : str
         Name of the flows produced by this maker.
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
+    mlip_type: Literal["GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]
         Choose one specific MLIP type to be fitted.
     hyperpara_opt: bool
         Perform hyperparameter optimization using XPOT
@@ -93,7 +93,9 @@ class MLIPFitMaker(Maker):
     """
 
     name: str = "MLpotentialFit"
-    mlip_type: Literal["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"] = "GAP"
+    mlip_type: Literal["GAP", "J-ACE", "P-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"] = (
+        "GAP"
+    )
     hyperpara_opt: bool = False
     ref_energy_name: str = "REF_energy"
     ref_force_name: str = "REF_forces"
@@ -149,10 +151,19 @@ def make(
         fit_kwargs: dict
             Additional keyword arguments for MLIP fitting.
         """
-        if self.mlip_type not in ["GAP", "J-ACE", "NEP", "NEQUIP", "M3GNET", "MACE"]:
+        if self.mlip_type not in [
+            "GAP",
+            "J-ACE",
+            "P-ACE",
+            "NEP",
+            "NEQUIP",
+            "M3GNET",
+            "MACE",
+        ]:
             raise ValueError(
                 "Please correct the MLIP name!"
-                "The current version ONLY supports the following models: GAP, J-ACE, NEP, NEQUIP, M3GNET, and MACE."
+                "The current version ONLY supports the following models: "
+                "GAP, J-ACE, P-ACE, NEP, NEQUIP, M3GNET, and MACE."
             )
 
         if self.apply_data_preprocessing:

src/autoplex/fitting/common/jobs.py

@@ -16,6 +16,7 @@
     mace_fitting,
     nep_fitting,
     nequip_fitting,
+    pace_fitting,
 )
 
 
@@ -67,7 +68,7 @@ def machine_learning_fit(
         List of GPU indices to be used for fitting. Only used for NEP fitting.
     mlip_type: str
         Choose one specific MLIP type to be fitted:
-        'GAP' | 'J-ACE' | 'NEQUIP' | 'NEP' | 'M3GNET' | 'MACE'
+        'GAP' | 'J-ACE' | 'P-ACE' | 'NEQUIP' | 'NEP' | 'M3GNET' | 'MACE'
     ref_energy_name: str
         Reference energy name.
     ref_force_name: str
@@ -164,6 +165,51 @@ def machine_learning_fit(
         )
         mlip_paths.append(train_test_error["mlip_path"])
 
+    elif mlip_type == "P-ACE":
+
+        from autoplex.settings import PacemakerSettings  # noqa: PLC0415
+
+        pace_specific_keys = {
+            "cutoff",
+            "seed",
+            "metadata",
+            "potential",
+            "data",
+            "fit",
+            "backend",
+        }
+        pace_kwargs = {k: v for k, v in fit_kwargs.items() if k in pace_specific_keys}
+
+        pace_hypers = (
+            PacemakerSettings(**pace_kwargs) if pace_kwargs else hyperparameters.P_ACE
+        )
+
+        # if not species_list:
+        #     if pace_hypers.potential and "elements" in pace_hypers.potential:
+        if (
+            not species_list
+            and pace_hypers.potential
+            and "elements" in pace_hypers.potential
+        ):
+            species_list = pace_hypers.potential["elements"]
+
+        remaining_fit_kwargs = {
+            k: v for k, v in fit_kwargs.items() if k not in pace_specific_keys
+        }
+
+        train_test_error = pace_fitting(
+            db_dir=database_dir,
+            species_list=species_list,
+            hyperparameters=pace_hypers,
+            fit_kwargs=remaining_fit_kwargs,  # Pass only non-P-ACE params
+            isolated_atom_energies=isolated_atom_energies,
+            ref_energy_name=ref_energy_name,
+            ref_force_name=ref_force_name,
+            ref_virial_name=ref_virial_name,
+            num_processes_fit=num_processes_fit,
+        )
+        mlip_paths.append(train_test_error["mlip_path"])
+
     elif mlip_type == "NEP":
         if gpu_identifier_indices is None:
             gpu_identifier_indices = [0]