basic guesser features

benchmarks/benchmarks/topology.py

@@ -1,6 +1,6 @@
 import MDAnalysis
 import numpy as np
-from MDAnalysis.topology import guessers
+from MDAnalysis.guesser import DefaultGuesser
 
 try:
     from MDAnalysisTests.datafiles import GRO
@@ -26,7 +26,7 @@ def setup(self, num_atoms):
 
     def time_guessbonds(self, num_atoms):
         """Benchmark for guessing bonds"""
-        guessers.guess_bonds(self.ag, self.ag.positions,
+        DefaultGuesser(None).guess_bonds(self.ag, self.ag.positions,
                              box=self.ag.dimensions,
                              vdwradii=self.vdwradii)
 

package/CHANGELOG

@@ -16,13 +16,14 @@ The rules for this file:
 -------------------------------------------------------------------------------
 ??/??/?? IAlibay, HeetVekariya, marinegor, lilyminium, RMeli,
          ljwoods2, aditya292002, pstaerk, PicoCentauri, BFedder,
-         tyler.je.reddy, SampurnaM, leonwehrhan, kainszs, orionarcher,
+         tyler.je.reddy, SampurnaM, leonwehrhan, kainszs, orionarcher, 
          yuxuanzhuang, PythonFZ, laksh-krishna-sharma, orbeckst, MattTDavies,
-         talagayev
+         talagayev, aya9aladdin
 
  * 2.8.0
 
 Fixes
+ * Fix Bohrium (Bh) atomic mass in tables.py (PR #3753)
  * set `n_parts` to the total number of frames being analyzed if
    `n_parts` is bigger. (Issue #4685)
  * Catch higher dimensional indexing in GroupBase & ComponentBase (Issue #4647)
@@ -56,6 +57,17 @@ Fixes
  * Fix groups.py doctests using sphinx directives (Issue #3925, PR #4374)
 
 Enhancements
+ * Removed type and mass guessing from all topology parsers (PR #3753)
+ * Added guess_TopologyAttrs() API to the Universe to handle attribute
+   guessing (PR #3753)
+ * Added the DefaultGuesser class, which is a general-purpose guesser with
+   the same functionalities as the existing guesser.py methods (PR #3753)
+ * Removed type guessing from RDKitParser and assigned type attribute
+   to be equal to the element attribute, if type is not provided (PR #3753)
+ * Added is_value_missing() to `TopologyAttrs` to check for missing
+   values (PR #3753)
+ * Added guessed `Element` attribute to the ITPParser to preserve old mass
+   partial guessing behavior from being broken (PR #3753)
  * Introduce parallelization API to `AnalysisBase` and to `analysis.rms.RMSD` class
    (Issue #4158, PR #4304)
  * Enables parallelization for analysis.gnm.GNMAnalysis (Issue #4672)
@@ -96,6 +108,9 @@ Changes
    numpy.testing.assert_allclose #4438)
 
 Deprecations
+ * Unknown masses are set to 0.0 for current version, this will be depracated
+   in version 3.0.0 and replaced by :class:`Masses`' no_value_label attribute(np.nan)
+   (PR #3753)
  * The MMTF Reader is deprecated and will be removed in version 3.0
    as the MMTF format is no longer supported (Issue #4634)
  * The MDAnalysis.analysis.waterdynamics module has been deprecated in favour

package/MDAnalysis/__init__.py

@@ -181,7 +181,7 @@
 _TOPOLOGY_ATTRS: Dict = {}   # {attrname: cls}
 _TOPOLOGY_TRANSPLANTS: Dict = {}   # {name: [attrname, method, transplant class]}
 _TOPOLOGY_ATTRNAMES: Dict = {}   # {lower case name w/o _ : name}
-
+_GUESSERS: Dict = {}
 
 # custom exceptions and warnings
 from .exceptions import (

package/MDAnalysis/analysis/bat.py

@@ -175,7 +175,7 @@ class to calculate dihedral angles for a given set of atoms or residues
 
 
 def _sort_atoms_by_mass(atoms, reverse=False):
-    r"""Sorts a list of atoms by name and then by index
+    r"""Sorts a list of atoms by mass and then by index
 
     The atom index is used as a tiebreaker so that the ordering is reproducible.
 

package/MDAnalysis/converters/OpenMMParser.py

@@ -25,6 +25,9 @@
 ===================================================================
 
 .. versionadded:: 2.0.0
+.. versionchanged:: 2.8.0
+   Removed type and mass guessing (attributes guessing takes place
+   now through universe.guess_TopologyAttrs() API)
 
 
 Converts an
@@ -59,8 +62,7 @@
 import warnings
 
 from ..topology.base import TopologyReaderBase
-from ..topology.tables import SYMB2Z
-from ..topology.guessers import guess_types, guess_masses
+from ..guesser.tables import SYMB2Z
 from ..core.topology import Topology
 from ..core.topologyattrs import (
     Atomids,
@@ -108,11 +110,6 @@
         -------
         top : MDAnalysis.core.topology.Topology
 
-        Note
-        ----
-        When none of the elements are present in the openmm topolgy, their
-        atomtypes are guessed using their names and their masses are
-        then guessed using their atomtypes.
 
         When partial elements are present, values from available elements
         are used whereas the absent elements are assigned an empty string
@@ -184,21 +181,32 @@
                 warnings.warn("Element information missing for some atoms. "
                               "These have been given an empty element record ")
                 if any(i == 'X' for i in atomtypes):
-                    warnings.warn("For absent elements, atomtype has been  "
-                                  "set to 'X' and mass has been set to 0.0. "
-                                  "If needed these can be guessed using "
-                                  "MDAnalysis.topology.guessers.")
+                    warnings.warn(
+                        "For absent elements, atomtype has been  "
+                        "set to 'X' and mass has been set to 0.0. "
+                        "If needed these can be guessed using "
+                        "universe.guess_TopologyAttrs("
+                        "to_guess=['masses', 'types']). "
+                        "(for MDAnalysis version 2.x "
+                        "this is done automatically,"
+                        " but it will be removed in 3.0).")
+
                 attrs.append(Elements(np.array(validated_elements,
                                                dtype=object)))
 
             else:
-                atomtypes = guess_types(atomnames)
-                masses = guess_masses(atomtypes)
-                wmsg = ("Element information is missing for all the atoms. "
-                        "Elements attribute will not be populated. "
-                        "Atomtype attribute will be guessed using atom "
-                        "name and mass will be guessed using atomtype."
-                        "See MDAnalysis.topology.guessers.")
+                wmsg = (
+                    "Element information is missing for all the atoms. "
+                    "Elements attribute will not be populated. "
+                    "Atomtype attribute will be guessed using atom "
+                    "name and mass will be guessed using atomtype."
+                    "For MDAnalysis version 2.x this is done automatically, "
+                    "but it will be removed in MDAnalysis v3.0. "
+                    "These can be guessed using "
+                    "universe.guess_TopologyAttrs("
+                    "to_guess=['masses', 'types']) "
+                    "See MDAnalysis.guessers.")
+
                 warnings.warn(wmsg)
         else:
             attrs.append(Elements(np.array(validated_elements, dtype=object)))

package/MDAnalysis/converters/ParmEd.py

@@ -81,12 +81,12 @@
 import itertools
 import warnings
 
+from ..guesser.tables import SYMB2Z
 import numpy as np
 from numpy.lib import NumpyVersion
 
 from . import base
 from ..coordinates.base import SingleFrameReaderBase
-from ..topology.tables import SYMB2Z
 from ..core.universe import Universe
 from ..exceptions import NoDataError
 

package/MDAnalysis/converters/ParmEdParser.py

@@ -90,7 +90,7 @@
 import numpy as np
 
 from ..topology.base import TopologyReaderBase, change_squash
-from ..topology.tables import Z2SYMB
+from ..guesser.tables import Z2SYMB
 from ..core.topologyattrs import (
     Atomids,
     Atomnames,

package/MDAnalysis/converters/RDKit.py

@@ -255,9 +255,7 @@ class RDKitConverter(base.ConverterBase):
         from MDAnalysisTests.datafiles import PSF, DCD
         from rdkit.Chem.Descriptors3D import Asphericity
 
-        u = mda.Universe(PSF, DCD)
-        elements = mda.topology.guessers.guess_types(u.atoms.names)
-        u.add_TopologyAttr('elements', elements)
+        u = mda.Universe(PSF, DCD, to_guess=['elements'])
         ag = u.select_atoms("resid 1-10")
 
         for ts in u.trajectory:

package/MDAnalysis/converters/RDKitParser.py

@@ -1,5 +1,5 @@
 # -*- Mode: python; tab-width: 4; indent-tabs-mode:nil; coding:utf-8 -*-
-# vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4 
+# vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
 #
 # MDAnalysis --- https://www.mdanalysis.org
 # Copyright (c) 2006-2017 The MDAnalysis Development Team and contributors
@@ -47,7 +47,6 @@
 import numpy as np
 
 from ..topology.base import TopologyReaderBase, change_squash
-from ..topology import guessers
 from ..core.topologyattrs import (
     Atomids,
     Atomnames,
@@ -90,16 +89,14 @@
      - Atomnames
      - Aromaticities
      - Elements
+     - Types
      - Masses
      - Bonds
      - Resids
      - Resnums
      - RSChirality
      - Segids
 
-    Guesses the following:
-     - Atomtypes
-
     Depending on RDKit's input, the following Attributes might be present:
      - Charges
      - Resnames
@@ -156,6 +153,12 @@
     .. versionadded:: 2.0.0
     .. versionchanged:: 2.1.0
        Added R/S chirality support
+    .. versionchanged:: 2.8.0
+       Removed type guessing (attributes guessing takes place now
+       through universe.guess_TopologyAttrs() API). If atoms types is not
+       present in the input rdkit molecule as a _TriposAtomType property,
+       the type attribute get the same values as the element attribute.
+
     """
     format = 'RDKIT'
 
@@ -303,8 +306,7 @@
         if atomtypes:
             attrs.append(Atomtypes(np.array(atomtypes, dtype=object)))
         else:
-            atomtypes = guessers.guess_types(names)
-            attrs.append(Atomtypes(atomtypes, guessed=True))
+            atomtypes = np.char.upper(elements)
 
         # Partial charges
         if charges:

package/MDAnalysis/coordinates/PDB.py

@@ -155,7 +155,6 @@
 from ..lib.util import store_init_arguments
 from . import base
 from .timestep import Timestep
-from ..topology.core import guess_atom_element
 from ..exceptions import NoDataError
 
 

package/MDAnalysis/core/groups.py

@@ -3453,7 +3453,6 @@ def guess_bonds(self, vdwradii=None, fudge_factor=0.55, lower_bound=0.1):
         ----------
         vdwradii : dict, optional
             Dict relating atom types: vdw radii
-
         fudge_factor : float, optional
             The factor by which atoms must overlap each other to be considered
             a bond.  Larger values will increase the number of bonds found. [0.55]
@@ -3477,8 +3476,8 @@ def guess_bonds(self, vdwradii=None, fudge_factor=0.55, lower_bound=0.1):
            Corrected misleading docs, and now allows passing of `fudge_factor`
            and `lower_bound` arguments.
         """
-        from ..topology.core import guess_bonds, guess_angles, guess_dihedrals
         from .topologyattrs import Bonds, Angles, Dihedrals
+        from ..guesser.default_guesser import DefaultGuesser
 
         def get_TopAttr(u, name, cls):
             """either get *name* or create one from *cls*"""
@@ -3490,22 +3489,20 @@ def get_TopAttr(u, name, cls):
                 return attr
 
         # indices of bonds
-        b = guess_bonds(
-            self.atoms,
-            self.atoms.positions,
-            vdwradii=vdwradii,
-            box=self.dimensions,
-            fudge_factor=fudge_factor,
-            lower_bound=lower_bound,
-        )
-        bondattr = get_TopAttr(self.universe, "bonds", Bonds)
+        guesser = DefaultGuesser(None, fudge_factor=fudge_factor,
+                                 lower_bound=lower_bound,
+                                 box=self.dimensions,
+                                 vdwradii=vdwradii)
+        b = guesser.guess_bonds(self.atoms, self.atoms.positions)
+
+        bondattr = get_TopAttr(self.universe, 'bonds', Bonds)
         bondattr._add_bonds(b, guessed=True)
 
-        a = guess_angles(self.bonds)
+        a = guesser.guess_angles(self.bonds)
         angleattr = get_TopAttr(self.universe, 'angles', Angles)
         angleattr._add_bonds(a, guessed=True)
 
-        d = guess_dihedrals(self.angles)
+        d = guesser.guess_dihedrals(self.angles)
         diheattr = get_TopAttr(self.universe, 'dihedrals', Dihedrals)
         diheattr._add_bonds(d)
 

package/MDAnalysis/core/topologyattrs.py

@@ -518,6 +518,18 @@ def set_segments(self, sg, values):
         """Set segmentattributes for a given SegmentGroup"""
         raise NotImplementedError
 
+    @classmethod
+    def are_values_missing(cls, values):
+        """check if an attribute has a missing value
+
+        .. versionadded:: 2.8.0
+        """
+        missing_value_label = getattr(cls, 'missing_value_label', None)
+
+        if missing_value_label is np.nan:
+            return np.isnan(values)
+        else:
+            return values == missing_value_label
 
 # core attributes
 
@@ -1441,6 +1453,7 @@ class Masses(AtomAttr):
     attrname = 'masses'
     singular = 'mass'
     per_object = 'atom'
+    missing_value_label = np.nan
     target_classes = [AtomGroup, ResidueGroup, SegmentGroup,
                       Atom, Residue, Segment]
     transplants = defaultdict(list)