hillclimber 0.1.6__cp314-cp314t-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl

hillclimber/interfaces.py

@@ -0,0 +1,133 @@
+import pathlib
+from typing import Protocol
+
+import ase
+from ase.calculators.calculator import Calculator
+from PIL import Image
+
+
+class NodeWithCalculator(Protocol):
+    """Any class with a `get_calculator` method returning an ASE Calculator."""
+
+    def get_calculator(
+        self, *, directory: str | pathlib.Path | None = None, **kwargs
+    ) -> Calculator: ...
+
+
+class AtomSelector(Protocol):
+    """Protocol for selecting atoms within a single ASE Atoms object.
+
+    This interface defines the contract for selecting atoms based on various
+    criteria within an individual frame/structure.
+    """
+
+    def select(self, atoms: ase.Atoms) -> list[list[int]]:
+        """Select atoms based on the implemented criteria.
+
+        Parameters
+        ----------
+        atoms : ase.Atoms
+            The atomic structure to select from.
+
+        Returns
+        -------
+        list[list[int]]
+            Groups of indices of selected atoms. All indices in the inner lists
+            are representative of the same group, e.g. one molecule.
+        """
+        ...
+
+
+class PlumedGenerator(Protocol):
+    """Protocol for generating PLUMED strings from collective variables."""
+
+    def to_plumed(self, atoms: ase.Atoms) -> list[str]: ...
+
+
+class CollectiveVariable(Protocol):
+    """Protocol for collective variables (CVs) that can be used in PLUMED."""
+
+    prefix: str
+
+    def get_img(self, atoms: ase.Atoms) -> Image.Image: ...
+
+    def to_plumed(self, atoms: ase.Atoms) -> tuple[list[str], list[str]]:
+        """
+        Convert the collective variable to a PLUMED string.
+
+        Parameters
+        ----------
+        atoms : ase.Atoms
+            The atomic structure to use for generating the PLUMED string.
+
+        Returns
+        -------
+        tuple[list[str], str]
+            - List of distance labels.
+            - list of PLUMED strings representing the CV.
+        """
+        ...
+
+
+class BiasProtocol(Protocol):
+    """Protocol for individual bias potentials that act on collective variables.
+
+    Individual bias potentials (restraints, walls) operate on collective variables
+    and generate their own PLUMED commands via to_plumed(). These biases are added
+    to MetaDynamicsModel via the `actions` parameter.
+
+    This is distinct from MetadynamicsBias, which is used for configuring the
+    collective METAD command itself (via the `bias_cvs` parameter).
+    """
+
+    cv: CollectiveVariable
+
+    def to_plumed(self, atoms: ase.Atoms) -> list[str]:
+        """Generate PLUMED input strings for this bias.
+
+        Parameters
+        ----------
+        atoms : ase.Atoms
+            The atomic structure to use for generating the PLUMED string.
+
+        Returns
+        -------
+        list[str]
+            List of PLUMED command strings.
+        """
+        ...
+
+
+class MetadynamicsBias(Protocol):
+    """Protocol for metadynamics bias configuration objects.
+
+    This protocol defines the configuration structure for per-CV metadynamics
+    parameters (sigma, grid settings, etc.). These configuration objects are
+    used in the `bias_cvs` parameter of MetaDynamicsModel to build the collective
+    METAD command.
+
+    This is distinct from BiasProtocol, which is for individual bias potentials
+    (restraints, walls) that generate their own PLUMED commands and are added
+    via the `actions` parameter.
+    """
+
+    cv: CollectiveVariable
+    sigma: float | None = None
+    grid_min: float | str | None = None
+    grid_max: float | str | None = None
+    grid_bin: int | None = None
+
+
+__all__ = [
+    "NodeWithCalculator",
+    "AtomSelector",
+    "PlumedGenerator",
+    "CollectiveVariable",
+    "BiasProtocol",
+    "MetadynamicsBias",
+]
+
+
+def interfaces() -> dict[str, list[str]]:
+    """Return a dictionary of available interfaces."""
+    return {"plumed-nodes": __all__}

hillclimber/metadynamics.py

@@ -0,0 +1,416 @@
+import dataclasses
+import sys
+import typing as t
+from pathlib import Path
+
+import ase.units
+import zntrack
+
+from hillclimber.calc import NonOverwritingPlumed
+from hillclimber.interfaces import (
+    CollectiveVariable,
+    NodeWithCalculator,
+    PlumedGenerator,
+)
+from hillclimber.pycv import PyCV
+
+
+@dataclasses.dataclass
+class MetadBias:
+    """Metadynamics bias configuration for a collective variable.
+
+    Parameters
+    ----------
+    cv : CollectiveVariable
+        The collective variable to bias.
+    sigma : float, optional
+        The width of the Gaussian potential in the same units as the CV
+        (e.g., Å for distances, radians for angles), by default None.
+    grid_min : float | str, optional
+        The minimum value of the grid in CV units (or PLUMED expression like "-pi"),
+        by default None.
+    grid_max : float | str, optional
+        The maximum value of the grid in CV units (or PLUMED expression like "pi"),
+        by default None.
+    grid_bin : int, optional
+        The number of bins in the grid, by default None.
+
+    Resources
+    ---------
+    - https://www.plumed.org/doc-master/user-doc/html/METAD/
+    """
+
+    cv: CollectiveVariable
+    sigma: float | None = None
+    grid_min: float | str | None = None
+    grid_max: float | str | None = None
+    grid_bin: int | None = None
+
+
+@dataclasses.dataclass
+class MetaDynamicsConfig:
+    """Base configuration for metadynamics.
+
+    This contains only the global parameters that apply to all CVs.
+
+    Units
+    -----
+    hillclimber uses ASE units throughout. The UNITS line in the PLUMED input tells
+    PLUMED to interpret all values in ASE units:
+    - Distances: Ångström (Å)
+    - Energies: electronvolt (eV) - including HEIGHT, SIGMA for energy-based CVs, etc.
+    - Time: femtoseconds (fs)
+    - Temperature: Kelvin (K)
+
+    Parameters
+    ----------
+    height : float, optional
+        The height of the Gaussian potential in eV, by default 1.0.
+    pace : int, optional
+        The frequency of Gaussian deposition in MD steps, by default 500.
+    biasfactor : float, optional
+        The bias factor for well-tempered metadynamics, by default None.
+    temp : float, optional
+        The temperature of the system in Kelvin, by default 300.0.
+    file : str, optional
+        The name of the hills file, by default "HILLS".
+    adaptive : t.Literal["GEOM", "DIFF"] | None, optional
+        The adaptive scheme to use, by default None.
+        If None, no ADAPTIVE parameter is written to PLUMED.
+    flush : int | None
+        The frequency of flushing the output files in MD steps.
+        If None, uses the plumed default.
+
+    Resources
+    ---------
+    - https://www.plumed.org/doc-master/user-doc/html/METAD/
+    - https://www.plumed.org/doc-master/user-doc/html/FLUSH/
+    """
+
+    height: float = 1.0  # kJ/mol
+    pace: int = 500
+    biasfactor: float | None = None
+    temp: float = 300.0
+    file: str = "HILLS"
+    adaptive: t.Literal["GEOM", "DIFF"] | None = None
+    flush: int | None = None
+
+
+# THERE SEEMS to be an issue with MetaDynamicsModel changing but ASEMD not updating?!?!
+
+
+class MetaDynamicsModel(zntrack.Node, NodeWithCalculator):
+    """A metadynamics model.
+
+    Parameters
+    ----------
+    config : MetaDynamicsConfig
+        The configuration for the metadynamics simulation.
+    data : list[ase.Atoms]
+        The input data for the simulation.
+    data_idx : int, optional
+        The index of the data to use, by default -1.
+    bias_cvs : list[MetaDBiasCV], optional
+        The collective variables to bias, by default [].
+    actions : list[PlumedGenerator], optional
+        A list of actions to perform during the simulation, by default [].
+    timestep : float, optional
+        The timestep of the simulation in fs, by default 1.0.
+    model : NodeWithCalculator
+        The model to use for the simulation.
+
+    Example
+    -------
+    >>> import hillclimber as pn
+    >>> import ipsuite as ips
+    >>>
+    >>> data = ips.AddData("seed.xyz")
+    >>> cv1 = pn.DistanceCV(
+    ...     x1=pn.SMARTSSelector(pattern="[H]O[H]"),
+    ...     x2=pn.SMARTSSelector(pattern="CO[C:1]"),
+    ...     prefix="d",
+    ... )
+    >>> metad_cv1 = pn.MetadBias(
+    ...     cv=cv1, sigma=0.1, grid_min=0.0, grid_max=2.0, grid_bin=200
+    ... )
+    >>> model = pn.MetaDynamicsModel(
+    ...     config=pn.MetaDynamicsConfig(height=0.25, temp=300, pace=2000, biasfactor=10),
+    ...     bias_cvs=[metad_cv1],
+    ...     data=data.frames,
+    ...     model=ips.MACEMP(),
+    ...     timestep=0.5
+    ... )
+    >>> md = ips.ASEMD(
+    ...     model=model, data=data.frames, ...
+    ... )
+    """
+
+    config: MetaDynamicsConfig = zntrack.deps()
+    data: list[ase.Atoms] = zntrack.deps()
+    data_idx: int = zntrack.params(-1)
+    bias_cvs: list[MetadBias] = zntrack.deps(default_factory=list)
+    actions: list[PlumedGenerator] = zntrack.deps(default_factory=list)
+    timestep: float = zntrack.params(1.0)  # in fs, default is 1 fs
+    model: NodeWithCalculator = zntrack.deps()
+
+    figures: Path = zntrack.outs_path(zntrack.nwd / "figures", independent=True)
+
+    def run(self):
+        self.figures.mkdir(parents=True, exist_ok=True)
+        for cv in self.bias_cvs:
+            img = cv.cv.get_img(self.data[self.data_idx])
+            img.save(self.figures / f"{cv.cv.prefix}.png")
+
+    def get_calculator(
+        self, *, directory: str | Path | None = None, **kwargs
+    ) -> NonOverwritingPlumed:
+        if directory is None:
+            raise ValueError("Directory must be specified for PLUMED input files.")
+        directory = Path(directory)
+        directory.mkdir(parents=True, exist_ok=True)
+
+        atoms = self.data[self.data_idx]
+
+        # Write adapter scripts for any PyCV instances
+        self._write_pycv_adapter_scripts(directory, atoms)
+
+        lines = self.to_plumed(atoms)
+        # Replace FILE= with FILE={directory}/ for output files (HILLS, COLVAR, etc.)
+        # but NOT for LOAD commands (which have absolute paths to plugins)
+        processed_lines = []
+        for line in lines:
+            if line.startswith("LOAD FILE="):
+                # Don't modify LOAD commands - they have absolute paths
+                processed_lines.append(line)
+            else:
+                processed_lines.append(line.replace("FILE=", f"FILE={directory}/"))
+        lines = processed_lines
+
+        # Write plumed input file
+        with (directory / "plumed.dat").open("w") as file:
+            for line in lines:
+                file.write(line + "\n")
+
+        # Add directory to sys.path so PLUMED can import PyCV adapter scripts
+        # This is required because PLUMED's pycv plugin imports Python modules
+        # and needs to find the generated _pycv_*.py adapter scripts
+        directory_str = str(directory.resolve())
+        if directory_str not in sys.path:
+            sys.path.insert(0, directory_str)
+
+        kT = ase.units.kB * self.config.temp
+
+        return NonOverwritingPlumed(
+            calc=self.model.get_calculator(directory=directory),
+            atoms=atoms,
+            input=lines,
+            timestep=float(self.timestep * ase.units.fs),
+            kT=float(kT),
+            log=(directory / "plumed.log").as_posix(),
+        )
+
+    def _write_pycv_adapter_scripts(
+        self, directory: Path, atoms: ase.Atoms
+    ) -> list[Path]:
+        """Write adapter scripts for all PyCV instances.
+
+        Parameters
+        ----------
+        directory : Path
+            Directory to write scripts to.
+        atoms : ase.Atoms
+            Reference structure for atom selection.
+
+        Returns
+        -------
+        list[Path]
+            Paths to written adapter scripts.
+        """
+        scripts = []
+
+        # Collect all PyCV instances from bias_cvs
+        for bias_cv in self.bias_cvs:
+            if isinstance(bias_cv.cv, PyCV):
+                script_path = self._write_single_pycv_script(
+                    bias_cv.cv, directory, atoms
+                )
+                scripts.append(script_path)
+
+        return scripts
+
+    def _write_single_pycv_script(
+        self, pycv: PyCV, directory: Path, atoms: ase.Atoms
+    ) -> Path:
+        """Write adapter script for a single PyCV instance.
+
+        Parameters
+        ----------
+        pycv : PyCV
+            The PyCV instance.
+        directory : Path
+            Directory to write to.
+        atoms : ase.Atoms
+            Reference structure.
+
+        Returns
+        -------
+        Path
+            Path to written script.
+        """
+        cv_class = type(pycv)
+
+        return pycv.write_adapter_script(
+            directory=directory,
+            atoms=atoms,
+            cv_class_module=cv_class.__module__,
+            cv_class_name=cv_class.__name__,
+            cv_init_args=pycv.get_init_args(),
+        )
+
+    def _has_pycv(self) -> bool:
+        """Check if any bias CV is a PyCV instance."""
+        return any(isinstance(bias_cv.cv, PyCV) for bias_cv in self.bias_cvs)
+
+    def to_plumed(self, atoms: ase.Atoms) -> list[str]:
+        """Generate PLUMED input string for the metadynamics model."""
+        # check for duplicate CV prefixes
+        cv_labels = set()
+        for bias_cv in self.bias_cvs:
+            if bias_cv.cv.prefix in cv_labels:
+                raise ValueError(f"Duplicate CV prefix found: {bias_cv.cv.prefix}")
+            cv_labels.add(bias_cv.cv.prefix)
+
+        plumed_lines = []
+        all_labels = []
+
+        sigmas, grid_mins, grid_maxs, grid_bins = [], [], [], []
+
+        # PLUMED UNITS line specifies conversion factors from ASE units to PLUMED's native units:
+        # - LENGTH=A: ASE uses Ångström (A), PLUMED native is nm → A is a valid PLUMED unit
+        # - TIME: ASE uses fs, PLUMED native is ps → 1 fs = 0.001 ps
+        # - ENERGY: ASE uses eV, PLUMED native is kJ/mol → 1 eV = 96.485 kJ/mol
+        # See: https://www.plumed.org/doc-master/user-doc/html/ (MD engine integration docs)
+        plumed_lines.append(
+            f"UNITS LENGTH=A TIME={1 / 1000} ENERGY={ase.units.mol / ase.units.kJ}"
+        )
+
+        # Add LOAD command for pycv plugin if any PyCV instances are present
+        if self._has_pycv():
+            import plumed
+
+            pycv_path = plumed.get_pycv_path()
+            plumed_lines.append(f"LOAD FILE={pycv_path}")
+
+        for bias_cv in self.bias_cvs:
+            labels, cv_str = bias_cv.cv.to_plumed(atoms)
+            plumed_lines.extend(cv_str)
+            all_labels.extend(labels)
+
+            # Collect per-CV parameters for later - repeat for each label
+            # PLUMED requires one parameter value per ARG, so if a CV generates
+            # multiple labels, we need to repeat the parameter values
+            for _ in labels:
+                sigmas.append(str(bias_cv.sigma) if bias_cv.sigma is not None else None)
+                grid_mins.append(
+                    str(bias_cv.grid_min) if bias_cv.grid_min is not None else None
+                )
+                grid_maxs.append(
+                    str(bias_cv.grid_max) if bias_cv.grid_max is not None else None
+                )
+                grid_bins.append(
+                    str(bias_cv.grid_bin) if bias_cv.grid_bin is not None else None
+                )
+
+        metad_parts = [
+            "METAD",
+            f"ARG={','.join(all_labels)}",
+            f"HEIGHT={self.config.height}",
+            f"PACE={self.config.pace}",
+            f"TEMP={self.config.temp}",
+            f"FILE={self.config.file}",
+        ]
+        if self.config.adaptive is not None:
+            metad_parts.append(f"ADAPTIVE={self.config.adaptive}")
+        if self.config.biasfactor is not None:
+            metad_parts.append(f"BIASFACTOR={self.config.biasfactor}")
+
+        # Add SIGMA, GRID_MIN, GRID_MAX, GRID_BIN only if any value is set
+        if any(v is not None for v in sigmas):
+            # When using ADAPTIVE, PLUMED requires only one sigma value
+            if self.config.adaptive is not None:
+                # Validate that all sigma values are the same when adaptive is set
+                unique_sigmas = set(v for v in sigmas if v is not None)
+                if len(unique_sigmas) > 1:
+                    raise ValueError(
+                        f"When using ADAPTIVE={self.config.adaptive}, all CVs must have the same sigma value. "
+                        f"Found different sigma values: {unique_sigmas}"
+                    )
+                # Use the first non-None sigma value
+                sigma_value = next(v for v in sigmas if v is not None)
+                metad_parts.append(f"SIGMA={sigma_value}")
+            else:
+                # Standard mode: one sigma per CV
+                metad_parts.append(
+                    f"SIGMA={','.join(v if v is not None else '0.0' for v in sigmas)}"
+                )
+        if any(v is not None for v in grid_mins):
+            metad_parts.append(
+                f"GRID_MIN={','.join(v if v is not None else '0.0' for v in grid_mins)}"
+            )
+        if any(v is not None for v in grid_maxs):
+            metad_parts.append(
+                f"GRID_MAX={','.join(v if v is not None else '0.0' for v in grid_maxs)}"
+            )
+        if any(v is not None for v in grid_bins):
+            metad_parts.append(
+                f"GRID_BIN={','.join(v if v is not None else '0' for v in grid_bins)}"
+            )
+
+        plumed_lines.append(f"metad: {' '.join(metad_parts)}")
+
+        # Track defined commands to detect duplicates and conflicts
+        # Map label -> full command for labeled commands (e.g., "d: DISTANCE ...")
+        defined_commands = {}
+        for line in plumed_lines:
+            # Check if this is a labeled command (format: "label: ACTION ...")
+            if ": " in line:
+                label = line.split(": ", 1)[0]
+                defined_commands[label] = line
+
+        # Add any additional actions (restraints, walls, print actions, etc.)
+        for action in self.actions:
+            action_lines = action.to_plumed(atoms)
+
+            # Filter out duplicate CV definitions, but detect conflicts
+            filtered_lines = []
+            for line in action_lines:
+                # Check if this is a labeled command
+                if ": " in line:
+                    label = line.split(": ", 1)[0]
+
+                    # Check if this label was already defined
+                    if label in defined_commands:
+                        # If the command is identical, skip (deduplication)
+                        if defined_commands[label] == line:
+                            continue
+                        # If the command is different, raise error (conflict)
+                        else:
+                            raise ValueError(
+                                f"Conflicting definitions for label '{label}':\n"
+                                f"  Already defined: {defined_commands[label]}\n"
+                                f"  New definition:  {line}"
+                            )
+                    else:
+                        # New labeled command, track it
+                        defined_commands[label] = line
+                        filtered_lines.append(line)
+                else:
+                    # Unlabeled command, always add
+                    filtered_lines.append(line)
+
+            plumed_lines.extend(filtered_lines)
+
+        # Add FLUSH if configured
+        if self.config.flush is not None:
+            plumed_lines.append(f"FLUSH STRIDE={self.config.flush}")
+        return plumed_lines

hillclimber/nodes.py

@@ -0,0 +1,6 @@
+from hillclimber import __all__
+
+
+def nodes() -> dict[str, list[str]]:
+    """Return all available nodes."""
+    return {"hillclimber": __all__}