biotite 0.40.0__cp310-cp310-win_amd64.whl → 0.41.0__cp310-cp310-win_amd64.whl

biotite/structure/io/pdb/file.py

@@ -12,7 +12,7 @@ from ...atoms import AtomArray, AtomArrayStack, repeat
 from ...bonds import BondList, connect_via_residue_names
 from ...box import vectors_from_unitcell, unitcell_from_vectors
 from ....file import TextFile, InvalidFileError
-from ..general import _guess_element as guess_element
+from ...repair import infer_elements
 from ...error import BadStructureError
 from ...filter import (
     filter_first_altloc,
@@ -23,6 +23,9 @@ from ...util import matrix_rotate
 from .hybrid36 import encode_hybrid36, decode_hybrid36, max_hybrid36_number
 
 
+_PDB_MAX_ATOMS = 99999
+_PDB_MAX_RESIDUES = 9999
+
 # slice objects for readability
 # ATOM/HETATM
 _record = slice(0, 6)
@@ -460,15 +463,14 @@ class PDBFile(TextFile):
 
         # Replace empty strings for elements with guessed types
         # This is used e.g. for PDB files created by Gromacs
-        if "" in array.element:
-            rep_num = 0
-            for idx in range(len(array.element)):
-                if not array.element[idx]:
-                    atom_name = array.atom_name[idx]
-                    array.element[idx] = guess_element(atom_name)
-                    rep_num += 1
+        empty_element_mask = array.element == ""
+        if empty_element_mask.any():
             warnings.warn(
-                "{} elements were guessed from atom_name.".format(rep_num)
+                f"{np.count_nonzero(empty_element_mask)} elements "
+                "were guessed from atom name"
+            )
+            array.element[empty_element_mask] = infer_elements(
+                array.atom_name[empty_element_mask]
             )
 
         # Fill in coordinates
@@ -574,6 +576,8 @@ class PDBFile(TextFile):
         records are also written for all non-water hetero residues
         and all inter-residue connections.
         """
+        _check_pdb_compatibility(array, hybrid36)
+
         natoms = array.array_length()
         annot_categories = array.get_annotation_categories()
         record = np.char.array(np.where(array.hetero, "HETATM", "ATOM"))
@@ -599,25 +603,6 @@ class PDBFile(TextFile):
         else:
             charge = np.char.array(np.full(natoms, "  ", dtype="U2"))
 
-        # Do checks on atom array (stack)
-        if hybrid36:
-            max_atoms = max_hybrid36_number(5)
-            max_residues = max_hybrid36_number(4)
-        else:
-            max_atoms, max_residues = 99999, 9999
-        if array.array_length() > max_atoms:
-            warnings.warn(f"More then {max_atoms:,} atoms per model")
-        if (array.res_id > max_residues).any():
-            warnings.warn(f"Residue IDs exceed {max_residues:,}")
-        if np.isnan(array.coord).any():
-            raise BadStructureError("Coordinates contain 'NaN' values")
-        if any([len(name) > 1 for name in array.chain_id]):
-            raise BadStructureError("Some chain IDs exceed 1 character")
-        if any([len(name) > 3 for name in array.res_name]):
-            raise BadStructureError("Some residue names exceed 3 characters")
-        if any([len(name) > 4 for name in array.atom_name]):
-            raise BadStructureError("Some atom names exceed 4 characters")
-
         if hybrid36:
             pdb_atom_id = np.char.array(
                 [encode_hybrid36(i, 5) for i in atom_id]
@@ -630,14 +615,14 @@ class PDBFile(TextFile):
             # but negative IDs are also possible
             pdb_atom_id = np.char.array(np.where(
                 atom_id > 0,
-                ((atom_id - 1) % 99999) + 1,
+                ((atom_id - 1) % _PDB_MAX_ATOMS) + 1,
                 atom_id
             ).astype(str))
             # Residue IDs are supported up to 9999,
             # but negative IDs are also possible
             pdb_res_id = np.char.array(np.where(
                 array.res_id > 0,
-                ((array.res_id - 1) % 9999) + 1,
+                ((array.res_id - 1) % _PDB_MAX_RESIDUES) + 1,
                 array.res_id
             ).astype(str))
 
@@ -1184,4 +1169,72 @@ def _apply_transformations(structure, rotations, translations):
         coord += translation
         assembly_coord[i] = coord
 
-    return repeat(structure, assembly_coord)
+    return repeat(structure, assembly_coord)
+
+
+def _check_pdb_compatibility(array, hybrid36):
+    annot_categories = array.get_annotation_categories()
+
+    if hybrid36:
+        max_atoms = max_hybrid36_number(5)
+        max_residues = max_hybrid36_number(4)
+    else:
+        max_atoms, max_residues = _PDB_MAX_ATOMS, _PDB_MAX_RESIDUES
+    if "atom_id" in annot_categories:
+        max_atom_id = np.max(array.atom_id)
+    else:
+        max_atom_id = array.array_length()
+
+    if max_atom_id > max_atoms:
+        warnings.warn(f"Atom IDs exceed {max_atoms:,}, will be wrapped")
+    if (array.res_id > max_residues).any():
+        warnings.warn(f"Residue IDs exceed {max_residues:,}, will be wrapped")
+    if np.isnan(array.coord).any():
+        raise BadStructureError("Coordinates contain 'NaN' values")
+    if any([len(name) > 1 for name in array.chain_id]):
+        raise BadStructureError("Some chain IDs exceed 1 character")
+    if any([len(name) > 3 for name in array.res_name]):
+        raise BadStructureError("Some residue names exceed 3 characters")
+    if any([len(name) > 4 for name in array.atom_name]):
+        raise BadStructureError("Some atom names exceed 4 characters")
+    for i, coord_name in enumerate(["x", "y", "z"]):
+        n_coord_digits = _number_of_integer_digits(array.coord[..., i])
+        if n_coord_digits > 4:
+            raise BadStructureError(
+                f"4 pre-decimal columns for {coord_name}-coordinates are "
+                f"available, but array would require {n_coord_digits}"
+            )
+    if "b_factor" in annot_categories:
+        n_b_factor_digits = _number_of_integer_digits(array.b_factor)
+        if n_b_factor_digits > 3:
+            raise BadStructureError(
+                "3 pre-decimal columns for B-factor are available, "
+                f"but array would require {n_b_factor_digits}"
+            )
+    if "occupancy" in annot_categories:
+        n_occupancy_digits = _number_of_integer_digits(array.occupancy)
+        if n_occupancy_digits > 3:
+            raise BadStructureError(
+                "3 pre-decimal columns for occupancy are available, "
+                f"but array would require {n_occupancy_digits}"
+            )
+    if "charge" in annot_categories:
+        # The sign can be omitted is it is put into the adjacent column
+        n_charge_digits = _number_of_integer_digits(np.abs(array.charge))
+        if n_charge_digits > 1:
+            raise BadStructureError(
+                "1 column for charge is available, "
+                f"but array would require {n_charge_digits}"
+            )
+
+
+def _number_of_integer_digits(values):
+    """
+    Get the maximum number of characters needed to represent the
+    pre-decimal positions of the given numeric values.
+    """
+    values = values.astype(int, copy=False)
+    n_digits = 0
+    n_digits = max(n_digits, len(str(np.min(values))))
+    n_digits = max(n_digits, len(str(np.max(values))))
+    return n_digits

biotite/structure/io/pdbx/__init__.py

@@ -20,5 +20,4 @@ from .bcif import *
 from .cif import *
 from .component import *
 from .encoding import *
-from .error import *
 from .legacy import *

biotite/structure/io/pdbx/bcif.py

@@ -12,9 +12,8 @@ import numpy as np
 import msgpack
 from .component import _Component, _HierarchicalContainer, MaskValue
 from .encoding import decode_stepwise, encode_stepwise, deserialize_encoding, \
-                      create_uncompressed_encoding, ByteArrayEncoding
-from .error import SerializationError
-from ....file import File, is_binary, is_open_compatible
+                      create_uncompressed_encoding
+from ....file import File, is_binary, is_open_compatible, SerializationError
 
 
 class BinaryCIFData(_Component):

biotite/structure/io/pdbx/cif.py

@@ -11,8 +11,8 @@ import shlex
 from collections.abc import MutableMapping, Sequence
 import numpy as np
 from .component import _Component, MaskValue
-from .error import DeserializationError, SerializationError
-from ....file import File, is_open_compatible, is_text
+from ....file import File, is_open_compatible, is_text, DeserializationError, \
+                     SerializationError
 
 
 UNICODE_CHAR_SIZE = 4
@@ -402,7 +402,9 @@ class CIFCategory(_Component, MutableMapping):
                     f"while the first column has row_count {self._row_count}"
                 )
 
-        if self._row_count == 1:
+        if self._row_count == 0:
+            raise ValueError("At least one row is required")
+        elif self._row_count == 1:
             lines = self._serialize_single()
         else:
             lines = self._serialize_looped()
@@ -766,7 +768,7 @@ class CIFFile(_Component, File, MutableMapping):
 
     @property
     def lines(self):
-        return "\n".join(self.serialize())
+        return self.serialize().splitlines()
 
     @property
     def block(self):
@@ -978,7 +980,7 @@ def _to_single(lines, is_looped):
                 processed_lines[out_i - 1] += " " + shlex.quote(multi_line_str)
             in_i = j + 1
 
-        elif not is_looped and lines[in_i][0] in ["'", '"']:
+        elif not is_looped and lines[in_i][0] != "_":
             # Singleline value in the line after the corresponding key
             processed_lines[out_i - 1] += " " + lines[in_i]
             in_i += 1
@@ -1025,4 +1027,6 @@ def _multiline(value):
 def _arrayfy(data):
     if not isinstance(data, (Sequence, np.ndarray)) or isinstance(data, str):
         data = [data]
+    elif len(data) == 0:
+        raise ValueError("Array must contain at least one element")
     return np.asarray(data)

biotite/structure/io/pdbx/component.py

@@ -14,7 +14,7 @@ __all__ = ["MaskValue"]
 from enum import IntEnum
 from abc import ABCMeta, abstractmethod
 from collections.abc import MutableMapping
-from .error import SerializationError, DeserializationError
+from ....file import SerializationError, DeserializationError
 
 
 class MaskValue(IntEnum):
@@ -105,6 +105,9 @@ class _Component(metaclass=ABCMeta):
         """
         raise NotImplementedError()
 
+    def __str__(self):
+        return str(self.serialize())
+
 
 class _HierarchicalContainer(_Component, MutableMapping,
                              metaclass=ABCMeta):