biotite 1.5.0__cp313-cp313-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl

biotite/structure/io/gro/file.py

@@ -0,0 +1,343 @@
+# This source code is part of the Biotite package and is distributed
+# under the 3-Clause BSD License. Please see 'LICENSE.rst' for further
+# information.
+
+__name__ = "biotite.structure.io.gro"
+__author__ = "Daniel Bauer, Patrick Kunzmann"
+__all__ = ["GROFile"]
+
+import copy
+from datetime import datetime
+import numpy as np
+from biotite.file import InvalidFileError, TextFile
+from biotite.structure.atoms import AtomArray, AtomArrayStack
+from biotite.structure.box import is_orthogonal
+from biotite.structure.error import BadStructureError
+from biotite.structure.repair import infer_elements
+
+_atom_records = {
+    "res_id": (0, 5),
+    "res_name": (5, 10),
+    "atom_name": (10, 15),
+    "atom_id": (15, 20),
+    "coord_x": (20, 28),
+    "coord_y": (28, 36),
+    "coord_z": (36, 44),
+    "v_x": (44, 52),
+    "v_y": (52, 60),
+    "v_z": (60, 68),
+}
+
+
+class GROFile(TextFile):
+    r"""
+    This class represents a GRO file.
+
+    This class only provides support for reading/writing the pure atom
+    information
+
+    Examples
+    --------
+    Load a `\\*.gro` file, modify the structure and save the new
+    structure into a new file:
+
+    >>> import os.path
+    >>> file = GROFile.read(os.path.join(path_to_structures, "1l2y.gro"))
+    >>> array_stack = file.get_structure()
+    >>> array_stack_mod = rotate(array_stack, [1,2,3])
+    >>> file = GROFile()
+    >>> file.set_structure(array_stack_mod)
+    >>> file.write(os.path.join(path_to_directory, "1l2y_mod.gro"))
+    """
+
+    def get_model_count(self):
+        """
+        Get the number of models contained in this GRO file.
+
+        Returns
+        -------
+        model_count : int
+            The number of models.
+        """
+        model_count = 0
+        for line in self.lines:
+            if _is_int(line):
+                model_count += 1
+        return model_count
+
+    def get_structure(self, model=None):
+        """
+        Get an :class:`AtomArray` or :class:`AtomArrayStack` from the
+        GRO file.
+
+        Parameters
+        ----------
+        model : int, optional
+            If this parameter is given, the function will return an
+            :class:`AtomArray` from the atoms corresponding to the given
+            model number (starting at 1).
+            Negative values are used to index models starting from the
+            last model insted of the first model.
+            If this parameter is omitted, an :class:`AtomArrayStack`
+            containing all models will be returned, even if the
+            structure contains only one model.
+
+        Returns
+        -------
+        array : AtomArray or AtomArrayStack
+            The return type depends on the `model` parameter.
+        """
+
+        def _get_atom_line_i(model_start_i, model_atom_counts):
+            """
+            Helper function to get the indices of all atoms for a model
+            """
+            return np.arange(model_start_i + 1, model_start_i + 1 + model_atom_counts)
+
+        def _set_box_dimen(box_param):
+            """
+            Helper function to create the box vectors from the values
+            in the GRO file
+
+            Parameters
+            ----------
+            box_param : list of float
+                The box dimensions in the GRO file.
+
+            Returns
+            -------
+            box_vectors : ndarray, dtype=float, shape=(3,3)
+                The atom array compatible box vectors.
+            """
+            if not any(box_param):
+                return None
+            if len(box_param) == 3:
+                x, y, z = box_param
+                return np.array([[x, 0, 0], [0, y, 0], [0, 0, z]], dtype=float)
+            elif len(box_param) == 9:
+                x1, y2, z3, x2, x3, y1, y3, z1, z2 = box_param
+                return np.array([[x1, x2, x3], [y1, y2, y3], [z1, z2, z3]], dtype=float)
+            else:
+                raise InvalidFileError(
+                    f"Invalid amount of box parameters: {len(box_param)}"
+                )
+
+        # Line indices where a new model starts
+        model_start_i = np.array(
+            [i for i in range(len(self.lines)) if _is_int(self.lines[i])], dtype=int
+        )
+
+        # Number of atoms in each model
+        model_atom_counts = np.array([int(self.lines[i]) for i in model_start_i])
+
+        if model is None:
+            # Check if all models have the same length
+            if np.all(model_atom_counts != model_atom_counts[0]):
+                raise BadStructureError(
+                    "The models in the file have unequal "
+                    "amount of atoms, give an explicit "
+                    "model instead"
+                )
+            depth = len(model_start_i)
+            length = model_atom_counts[0]
+            array = AtomArrayStack(depth, length)
+
+            # Line indices for annotation determination is determined
+            # from model 1
+            annot_i = _get_atom_line_i(model_start_i[0], length)
+        else:
+            if model == 0:
+                raise ValueError("The model index must not be 0")
+            # Negative models mean index starting from last model
+            model = len(model_start_i) + model + 1 if model < 0 else model
+            if model > len(model_start_i):
+                raise ValueError(
+                    f"The file has {len(model_start_i)} models, "
+                    f"the given model {model} does not exist"
+                )
+
+            length = model_atom_counts[model - 1]
+            array = AtomArray(length)
+
+            annot_i = _get_atom_line_i(model_start_i[model - 1], length)
+
+        # Replace empty strings for elements with guessed types
+        # i is index in array, line_i is line index
+        for i, line_i in enumerate(annot_i):
+            line = self.lines[line_i]
+            array.res_id[i] = int(line[0:5])
+            array.res_name[i] = line[5:10].strip()
+            array.atom_name[i] = line[10:15].strip()
+        array.element = infer_elements(array.atom_name)
+
+        # Fill in coordinates and boxes
+        if isinstance(array, AtomArray):
+            atom_i = annot_i
+            for i, line_i in enumerate(atom_i):
+                line = self.lines[line_i]
+                # gro files use nm instead of A
+                array.coord[i, 0] = float(line[20:28]) * 10
+                array.coord[i, 1] = float(line[28:36]) * 10
+                array.coord[i, 2] = float(line[36:44]) * 10
+            # Box is stored in last line (after coordinates)
+            box_i = atom_i[-1] + 1
+            box_param = [float(e) * 10 for e in self.lines[box_i].split()]
+            array.box = _set_box_dimen(box_param)
+
+        elif isinstance(array, AtomArrayStack):
+            for m in range(len(model_start_i)):
+                atom_i = _get_atom_line_i(model_start_i[m], model_atom_counts[m])
+                for i, line_i in enumerate(atom_i):
+                    line = self.lines[line_i]
+                    array.coord[m, i, 0] = float(line[20:28]) * 10
+                    array.coord[m, i, 1] = float(line[28:36]) * 10
+                    array.coord[m, i, 2] = float(line[36:44]) * 10
+                # Box is stored in last line (after coordinates)
+                box_i = atom_i[-1] + 1
+                box_param = [float(e) * 10 for e in self.lines[box_i].split()]
+                box = _set_box_dimen(box_param)
+                # Create a box in the stack if not already existing
+                # and the box is not a dummy
+                if box is not None:
+                    if array.box is None:
+                        array.box = np.zeros((array.stack_depth(), 3, 3))
+                    array.box[m] = box
+
+        return array
+
+    def set_structure(self, array):
+        """
+        Set the :class:`AtomArray` or :class:`AtomArrayStack` for the
+        file.
+
+        Parameters
+        ----------
+        array : AtomArray or AtomArrayStack
+            The array or stack to be saved into this file. If a stack
+            is given, each array in the stack is saved as separate
+            model.
+        """
+
+        def get_box_dimen(array):
+            """
+            GRO files have the box dimensions as last line for each
+            model.
+            In case, the `box` attribute of the atom array is
+            `None`, we simply use the min and max coordinates in xyz
+            to get the correct size
+
+            Parameters
+            ----------
+            array : AtomArray
+                The atom array to get the box dimensions from.
+
+            Returns
+            -------
+            box : str
+                The box, properly formatted for GRO files.
+            """
+            if array.box is None:
+                coord = array.coord
+                bx, by, bz = (coord.max(axis=0) - coord.min(axis=0)) / 10
+                return f"{bx:>8.3f} {by:>8.3f} {bz:>8.3f}"
+            else:
+                box = array.box
+                if is_orthogonal(box):
+                    bx, by, bz = np.diag(box) / 10
+                    return f"{bx:>9.5f} {by:>9.5f} {bz:>9.5f}"
+                else:
+                    box = box / 10
+                    box_elements = (
+                        box[0, 0],
+                        box[1, 1],
+                        box[2, 2],
+                        box[0, 1],
+                        box[0, 2],
+                        box[1, 0],
+                        box[1, 2],
+                        box[2, 0],
+                        box[2, 1],
+                    )
+                    return " ".join([f"{e:>9.5f}" for e in box_elements])
+
+        if "atom_id" in array.get_annotation_categories():
+            atom_id = array.atom_id
+        else:
+            atom_id = np.arange(1, array.array_length() + 1)
+        # Atom IDs are supported up to 99999,
+        # but negative IDs are also possible
+        gro_atom_id = np.where(atom_id > 0, ((atom_id - 1) % 99999) + 1, atom_id)
+        # Residue IDs are supported up to 9999,
+        # but negative IDs are also possible
+        gro_res_id = np.where(
+            array.res_id > 0, ((array.res_id - 1) % 99999) + 1, array.res_id
+        )
+
+        if isinstance(array, AtomArray):
+            self.lines = [None] * (array.array_length() + 3)
+
+            # Write header lines
+            self.lines[0] = f"Generated by Biotite at {datetime.now()}"
+            self.lines[1] = str(array.array_length())
+
+            # Write atom lines
+            fmt = "{:>5d}{:5s}{:>5s}{:>5d}{:>8.3f}{:>8.3f}{:>8.3f}"
+            for i in range(array.array_length()):
+                # gro format is in nm -> multiply coords by 10
+                self.lines[i + 2] = fmt.format(
+                    gro_res_id[i],
+                    array.res_name[i],
+                    array.atom_name[i],
+                    gro_atom_id[i],
+                    array.coord[i, 0] / 10,
+                    array.coord[i, 1] / 10,
+                    array.coord[i, 2] / 10,
+                )
+            # Write box lines
+            self.lines[-1] = get_box_dimen(array)
+        elif isinstance(array, AtomArrayStack):
+            self.lines = []
+            # The entire information, but the coordinates,
+            # is equal for each model
+            # Therefore template lines are created
+            # which are afterwards applied for each model
+            templines = [None] * array.array_length()
+            fmt = "{:>5d}{:5s}{:>5s}{:5d}"
+            for i in range(array.array_length()):
+                templines[i] = fmt.format(
+                    gro_res_id[i], array.res_name[i], array.atom_name[i], gro_atom_id[i]
+                )
+
+            for i in range(array.stack_depth()):
+                self.lines.append(
+                    f"Generated by Biotite at {datetime.now()}, model={i + 1}"
+                )
+                self.lines.append(str(array.array_length()))
+
+                # Fill in coordinates for each model
+                modellines = copy.copy(templines)
+                for j, line in enumerate(modellines):
+                    # Insert coordinates
+                    line = line + "{:>8.3f}{:>8.3f}{:>8.3f}".format(
+                        array.coord[i, j, 0] / 10,
+                        array.coord[i, j, 1] / 10,
+                        array.coord[i, j, 2] / 10,
+                    )
+                    modellines[j] = line
+                self.lines.extend(modellines)
+                self.lines.append(get_box_dimen(array[i]))
+        else:
+            raise TypeError("An atom array or stack must be provided")
+        # Add terminal newline, since PyMOL requires it
+        self.lines.append("")
+
+
+def _is_int(string):
+    """
+    Return ``True`, if the string can be parsed to an int.
+    """
+    try:
+        int(string)
+        return True
+    except ValueError:
+        return False

biotite/structure/io/mol/__init__.py

@@ -0,0 +1,20 @@
+# This source code is part of the Biotite package and is distributed
+# under the 3-Clause BSD License. Please see 'LICENSE.rst' for further
+# information.
+
+"""
+The MOL format is used to depict atom positions and bonds for small
+molecules.
+This subpackage is used for reading and writing an :class:`AtomArray`
+in this format.
+Additionally, reading data from the SDF format, which is a wrapper
+around MOL, is also supported.
+"""
+
+__name__ = "biotite.structure.io.mol"
+__author__ = "Patrick Kunzmann"
+
+from .convert import *
+from .header import *
+from .mol import *
+from .sdf import *

biotite/structure/io/mol/convert.py

@@ -0,0 +1,112 @@
+# This source code is part of the Biotite package and is distributed
+# under the 3-Clause BSD License. Please see 'LICENSE.rst' for further
+# information.
+
+__name__ = "biotite.structure.io.mol"
+__author__ = "Patrick Kunzmann"
+__all__ = ["get_structure", "set_structure"]
+
+from biotite.structure.bonds import BondType
+from biotite.structure.io.mol.mol import MOLFile
+from biotite.structure.io.mol.sdf import SDFile, SDRecord
+
+
+def get_structure(mol_file, record_name=None):
+    """
+    Get an :class:`AtomArray` from the MOL file.
+
+    Ths function is a thin wrapper around
+    :meth:`MOLFile.get_structure()`.
+
+    Parameters
+    ----------
+    mol_file : MOLFile or SDFile or SDRecord
+        The file.
+    record_name : str, optional
+        Has only an effect when `mol_file` is a :class:`SDFile`.
+        The name of the record in the SD file.
+        By default, the first record is used.
+
+    Returns
+    -------
+    array : AtomArray
+        This :class:`AtomArray` contains the optional ``charge``
+        annotation and has an associated :class:`BondList`.
+        All other annotation categories, except ``element`` are
+        empty.
+    """
+    record = _get_record(mol_file, record_name)
+    return record.get_structure()
+
+
+def set_structure(
+    mol_file, atoms, default_bond_type=BondType.ANY, version=None, record_name=None
+):
+    """
+    Set the :class:`AtomArray` for the MOL file.
+
+    Ths function is a thin wrapper around
+    :meth:`MOLFile.set_structure()`.
+
+    Parameters
+    ----------
+    mol_file : MOLFile
+        The MOL file.
+    atoms : AtomArray
+        The array to be saved into this file.
+        Must have an associated :class:`BondList`.
+        Bond type fallback for the *Bond block*, if a
+        :class:`BondType` has no CTAB counterpart.
+        By default, each such bond is treated as
+        :attr:`BondType.ANY`.
+    default_bond_type : BondType, optional
+        Bond type fallback for the *Bond block*, if a :class:`BondType`
+        has no CTAB counterpart.
+        By default, each such bond is treated as :attr:`BondType.ANY`.
+    version : {"V2000", "V3000"}, optional
+        The version of the CTAB format.
+        ``"V2000"`` uses the *Atom* and *Bond* block, while ``"V3000"``
+        uses the *Properties* block.
+        By default, ``"V2000"`` is used, unless the number of atoms or
+        bonds exceeds 999, in which case ``"V3000"`` is used.
+    record_name : str, optional
+        Has only an effect when `mol_file` is a :class:`SDFile`.
+        The name of the record.
+        Default is the first record of the file.
+        If the file is empty, a new record will be created.
+    """
+    record = _get_or_create_record(mol_file, record_name)
+    record.set_structure(atoms, default_bond_type, version)
+
+
+def _get_record(file, record_name):
+    if isinstance(file, (MOLFile, SDRecord)):
+        return file
+    elif isinstance(file, SDFile):
+        # Determine record
+        if record_name is None:
+            return file.record
+        else:
+            return file[record_name]
+    else:
+        raise TypeError(f"Unsupported file type '{type(file).__name__}'")
+
+
+def _get_or_create_record(file, record_name):
+    if isinstance(file, (MOLFile, SDRecord)):
+        return file
+    elif isinstance(file, SDFile):
+        if record_name is None:
+            if len(file) > 0:
+                # Choose first record by default
+                record_name = next(iter(file.keys()))
+            else:
+                # File is empty -> invent a new record name
+                record_name = "Molecule"
+
+        if record_name not in file:
+            record = SDRecord()
+            file[record_name] = record
+        return file[record_name]
+    else:
+        raise TypeError(f"Unsupported file type '{type(file).__name__}'")