Merge pull request MDAnalysis#3683 from hmacdope/CythonTimestep

Cythonized Timestep [Core]

package/CHANGELOG

@@ -13,7 +13,7 @@ The rules for this file:
   * release numbers follow "Semantic Versioning" http://semver.org
 
 ------------------------------------------------------------------------------
-??/??/?? IAlibay, PicoCentauri, orbeckst
+??/??/?? IAlibay, PicoCentauri, orbeckst, hmacdope
 
  * 2.3.0
 
@@ -24,6 +24,8 @@ Enhancements
   * Add `norm` parameter to InterRDF, InterRDF_s to normalize as rdf,
     number density or do not normalize at all. (Issue #3687)
   * Additional logger.info output when per-frame analysis starts (PR #3710)
+  * Timestep has been converted to a Cython Extension type
+    (CZI Performance track, PR #3683)
 
 Changes
 

package/MDAnalysis/analysis/distances.py

@@ -80,7 +80,7 @@ def contact_matrix(coord, cutoff=15.0, returntype="numpy", box=None):
        * ``"sparse"``: return as a :class:`scipy.sparse.lil_matrix`
     box : array-like or ``None``, optional, default ``None``
        Simulation cell dimensions in the form of
-       :attr:`MDAnalysis.trajectory.base.Timestep.dimensions` when
+       :attr:`MDAnalysis.trajectory.timestep.Timestep.dimensions` when
        periodic boundary conditions should be taken into account for
        the calculation of contacts.
 

package/MDAnalysis/analysis/pca.py

@@ -446,7 +446,7 @@ def project_single_frame(self, components=None, group=None, anchor=None):
         -------
         function
             The resulting function f(ts) takes as input a
-            :class:`~MDAnalysis.coordinates.base.Timestep` ts,
+            :class:`~MDAnalysis.coordinates.timestep.Timestep` ts,
             and returns ts with the projected structure
 
             .. warning::

package/MDAnalysis/auxiliary/__init__.py

@@ -112,7 +112,7 @@
 
 'Assignment' of auxiliary steps to trajectory timesteps is determined from the time 
 of the auxiliary step, ``dt`` of the trajectory and time at the first frame of the 
-trajectory (obtained through a :class:`~MDAnalysis.coordinates.base.Timestep` 
+trajectory (obtained through a :class:`~MDAnalysis.coordinates.timestep.Timestep` 
 instance from the trajectory), as::
 
     frame = floor((time_at_step - time_at_frame_0 + dt/2)/dt)

package/MDAnalysis/auxiliary/base.py

@@ -355,13 +355,13 @@ def update_ts(self, ts):
 
         Parameters
         ----------
-        ts : :class:`~MDAnalysis.coordinates.base.Timestep` object
+        ts : :class:`~MDAnalysis.coordinates.timestep.Timestep` object
             The trajectory timestep for which corresponding auxiliary data is
             to be read and updated.
 
         Returns
         -------
-        :class:`~MDAnalysis.coordinates.base.Timestep`
+        :class:`~MDAnalysis.coordinates.timestep.Timestep`
             *ts* with the representative auxiliary
             value in ``ts.aux`` be updated appropriately.
 
@@ -395,7 +395,7 @@ def read_ts(self, ts):
 
         Parameters
         ----------
-        ts : :class:`~MDAnalysis.coordinates.base.Timestep` object
+        ts : :class:`~MDAnalysis.coordinates.timestep.Timestep` object
             The trajectory timestep for which corresponding auxiliary data is
             to be read.
 
@@ -449,7 +449,7 @@ def step_to_frame(self, step, ts, return_time_diff=False):
         step : int
             Number of the auxiliary step to calculate closest trajectory frame
             for.
-        ts : :class:`~MDAnalysis.coordinates.base.Timestep` object
+        ts : :class:`~MDAnalysis.coordinates.timestep.Timestep` object
             (Any) timestep from the trajectory the calculated frame number is to
             correspond to.
         return_time_diff : bool, optional
@@ -492,7 +492,7 @@ def move_to_ts(self, ts):
 
         Parameters
         ----------
-        ts : :class:`~MDAnalysis.coordinates.base.Timestep` object
+        ts : :class:`~MDAnalysis.coordinates.timestep.Timestep` object
             The trajectory timestep before which the auxiliary reader is to
             be positioned.
         """
@@ -526,7 +526,7 @@ def next_nonempty_frame(self, ts):
 
         Parameters
         ----------
-        ts : :class:`~MDAnalysis.coordinates.base.Timestep` object
+        ts : :class:`~MDAnalysis.coordinates.timestep.Timestep` object
             Any timestep from the trajectory for which the next 'non-empty'
             frame is to be found.
 

package/MDAnalysis/coordinates/DLPoly.py

@@ -31,7 +31,6 @@
 import numpy as np
 
 from . import base
-from .base import Timestep
 from . import core
 from ..lib import util
 from ..lib.util import cached, store_init_arguments

package/MDAnalysis/coordinates/DMS.py

@@ -38,8 +38,6 @@
 from . import base
 from .core import triclinic_box
 
-from .base import Timestep
-
 
 class DMSReader(base.SingleFrameReaderBase):
     """

package/MDAnalysis/coordinates/FHIAIMS.py

@@ -116,9 +116,6 @@
 from ..lib import mdamath
 
 
-from .base import Timestep
-
-
 class FHIAIMSReader(base.SingleFrameReaderBase):
     """Reader for the FHIAIMS geometry format.
 

package/MDAnalysis/coordinates/GRO.py

@@ -108,11 +108,10 @@
 import numpy as np
 
 from . import base
-from .base import Timestep
 from .core import triclinic_box, triclinic_vectors
 from ..exceptions import NoDataError
 from ..lib import util
-
+from .timestep import Timestep
 
 """unitcell dimensions (A, B, C, alpha, beta, gamma)
 

package/MDAnalysis/coordinates/H5MD.py

@@ -243,11 +243,11 @@ class H5MDReader(base.ReaderBase):
 
     Additional data in the *observables* group of the H5MD file are
     loaded into the :attr:`Timestep.data
-    <MDAnalysis.coordinates.base.Timestep.data>` dictionary.
+    <MDAnalysis.coordinates.timestep.Timestep.data>` dictionary.
 
     Only 3D-periodic boxes or no periodicity are supported; for no
     periodicity, :attr:`Timestep.dimensions
-    <MDAnalysis.coordinates.base.Timestep.dimensions>` will return ``None``.
+    <MDAnalysis.coordinates.timestep.Timestep.dimensions>` will return ``None``.
 
     Although H5MD can store varying numbers of particles per time step
     as produced by, e.g., GCMC simulations, MDAnalysis can currently
@@ -826,7 +826,7 @@ class H5MDWriter(base.WriterBase):
     H5MD trajectories are automatically recognised by the
     file extension ".h5md".
 
-    All data from the input :class:`~MDAnalysis.coordinates.base.Timestep` is
+    All data from the input :class:`~MDAnalysis.coordinates.timestep.Timestep` is
     written by default. For detailed information on how :class:`H5MDWriter`
     handles units, compression, and chunking, see the Notes section below.
 
@@ -934,7 +934,7 @@ class H5MDWriter(base.WriterBase):
 
     By default, the writer will write all available data (positions,
     velocities, and forces) if detected in the input
-    :class:`~MDAnalysis.coordinates.base.Timestep`. In addition, the settings
+    :class:`~MDAnalysis.coordinates.timestep.Timestep`. In addition, the settings
     for `compression` and `compression_opts` will be read from
     the first available group of positions, velocities, or forces and used as
     the default value. To write a file without any one of these datsets,

package/MDAnalysis/coordinates/PDB.py

@@ -153,6 +153,7 @@
 from ..lib import util
 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
 
@@ -870,13 +871,13 @@ def _update_frame(self, obj):
         * :attr:`PDBWriter.trajectory` (the underlying trajectory
           :class:`~MDAnalysis.coordinates.base.Reader`)
         * :attr:`PDBWriter.timestep` (the underlying trajectory
-          :class:`~MDAnalysis.coordinates.base.Timestep`)
+          :class:`~MDAnalysis.coordinates.timestep.Timestep`)
 
         Before calling :meth:`_write_next_frame` this method **must** be
         called at least once to enable extracting topology information from the
         current frame.
         """
-        if isinstance(obj, base.Timestep):
+        if isinstance(obj, Timestep):
             raise TypeError("PDBWriter cannot write Timestep objects "
                             "directly, since they lack topology information ("
                             "atom names and types) required in PDB files")
@@ -970,14 +971,14 @@ def _write_next_frame(self, ts=None, **kwargs):
 
         :Keywords:
           *ts*
-             :class:`base.Timestep` object containing coordinates to be written to trajectory file;
+             :class:`timestep.Timestep` object containing coordinates to be written to trajectory file;
              if ``None`` then :attr:`PDBWriter.ts`` is tried.
           *multiframe*
              ``False``: write a single frame (default); ``True`` behave as a trajectory writer
 
         .. Note::
 
-           Before using this method with another :class:`base.Timestep` in the *ts*
+           Before using this method with another :class:`timestep.Timestep` in the *ts*
            argument, :meth:`PDBWriter._update_frame` *must* be called
            with the :class:`~MDAnalysis.core.groups.AtomGroup.Universe` as
            its argument so that topology information can be gathered.

package/MDAnalysis/coordinates/TRJ.py

@@ -41,28 +41,6 @@
 * lengths in Angstrom (Å)
 * time in ps (but see below)
 
-AMBER trajectory coordinate frames are based on a custom :class:`Timestep`
-object.
-
-.. autoclass:: Timestep
-   :members:
-
-   .. attribute:: _pos
-
-      coordinates of the atoms as a :class:`numpy.ndarray` of shape `(n_atoms, 3)`
-
-   .. attribute:: _velocities
-
-      velocities of the atoms as a :class:`numpy.ndarray` of shape `(n_atoms, 3)`;
-      only available if the trajectory contains velocities or if the
-      *velocities* = ``True`` keyword has been supplied.
-
-   .. attribute:: _forces
-
-      forces of the atoms as a :class:`numpy.ndarray` of shape `(n_atoms, 3)`;
-      only available if the trajectory contains forces or if the
-      *forces* = ``True`` keyword has been supplied.
-
 
 .. _netcdf-trajectories:
 
@@ -159,6 +137,7 @@
 from math import isclose
 
 import MDAnalysis
+from .timestep import Timestep
 from . import base
 from ..lib import util
 from ..lib.util import store_init_arguments
@@ -172,19 +151,6 @@
     logger.warning("netCDF4 is not available. Writing AMBER ncdf files will be slow.")
 
 
-class Timestep(base.Timestep):
-    """AMBER trajectory Timestep.
-
-    The Timestep can be initialized with `arg` being an integer
-    (the number of atoms) and an optional keyword argument `velocities` to
-    allocate space for both coordinates and velocities;
-
-    .. versionchanged:: 0.10.0
-       Added ability to contain Forces
-    """
-    order = 'C'
-
-
 class TRJReader(base.ReaderBase):
     """AMBER trajectory reader.
 

package/MDAnalysis/coordinates/TRZ.py

@@ -51,7 +51,8 @@
 import errno
 
 from . import base
-from .base import Timestep
+from ..exceptions import NoDataError
+from .timestep import Timestep
 from ..lib import util
 from ..lib.util import cached, store_init_arguments
 from .core import triclinic_box, triclinic_vectors
@@ -62,8 +63,8 @@ class TRZReader(base.ReaderBase):
 
     Attributes
     ----------
-    ts : base.Timestep
-         :class:`~MDAnalysis.coordinates.base.Timestep` object containing
+    ts : timestep.Timestep
+         :class:`~MDAnalysis.coordinates.timestep.Timestep` object containing
          coordinates of current frame
 
     Note
@@ -83,6 +84,8 @@ class TRZReader(base.ReaderBase):
     .. versionchanged:: 2.1.0
        TRZReader now returns a default :attr:`dt` of 1.0 when it cannot be
        obtained from the difference between two frames.
+    .. versionchanged:: 2.3.0
+       _frame attribute moved to `ts.data` dictionary.
     """
 
     format = "TRZ"
@@ -198,7 +201,7 @@ def _read_next_timestep(self, ts=None):
                                  "Maybe `topology` is wrong?".format(
                                                              self.n_atoms))
             ts.frame = data['nframe'][0] - 1  # 0 based for MDA
-            ts._frame = data['ntrj'][0]
+            ts.data['frame'] = data['ntrj'][0] # moved from attr to data
             ts.time = data['treal'][0]
             ts.dimensions = triclinic_box(*(data['box'].reshape(3, 3)))
             ts.data['pressure'] = data['pressure']
@@ -295,9 +298,9 @@ def skip_timestep(self):
         """Timesteps between trajectory frames"""
         curr_frame = self.ts.frame
         try:
-            t0 = self.ts._frame
+            t0 = self.ts.data['frame']
             self.next()
-            t1 = self.ts._frame
+            t1 = self.ts.data['frame']
             skip_timestep = t1 - t0
         except StopIteration:
             return 0
@@ -520,8 +523,8 @@ def _write_next_frame(self, obj):
                     data[att] = 0.0
                 faked_attrs.append(att)
         try:
-            data['step'] = ts._frame
-        except AttributeError:
+            data['step'] = ts.data['frame']
+        except KeyError:
             data['step'] = ts.frame
             faked_attrs.append('step')
         try:
@@ -543,8 +546,8 @@ def _write_next_frame(self, obj):
             unitcell = np.zeros(9, dtype=np.float32)
 
         try:
-            vels = ts._velocities
-        except AttributeError:
+            vels = ts.velocities
+        except NoDataError:
             vels = np.zeros((self.n_atoms, 3), dtype=np.float32, order='F')
             warnings.warn("Timestep didn't have velocity information, "
                           "this will be set to zero in output trajectory. ")

package/MDAnalysis/coordinates/TXYZ.py

@@ -52,7 +52,7 @@
 from ..lib import util
 from . import base
 from ..lib.util import openany, cached, store_init_arguments
-
+from .timestep import Timestep
 
 class TXYZReader(base.ReaderBase):
     """Reads from a TXYZ file"""
@@ -61,7 +61,7 @@ class TXYZReader(base.ReaderBase):
     format = ['TXYZ', 'ARC']
     # these are assumed!
     units = {'time': 'ps', 'length': 'Angstrom'}
-    _Timestep = base.Timestep
+    _Timestep = Timestep
 
     @store_init_arguments
     def __init__(self, filename, **kwargs):

package/MDAnalysis/coordinates/XYZ.py

@@ -90,6 +90,7 @@
 logger = logging.getLogger('MDAnalysis.coordinates.XYZ')
 
 from . import base
+from .timestep import Timestep
 from ..lib import util
 from ..lib.util import cached, store_init_arguments
 from ..exceptions import NoDataError
@@ -334,7 +335,7 @@ class XYZReader(base.ReaderBase):
     format = "XYZ"
     # these are assumed!
     units = {'time': 'ps', 'length': 'Angstrom'}
-    _Timestep = base.Timestep
+    _Timestep = Timestep
 
     @store_init_arguments
     def __init__(self, filename, **kwargs):