Merge pull request #96 from dwhswenson/release-0.7

Release 0.7

README.md

@@ -7,7 +7,7 @@
 
 [![Codacy Badge](https://api.codacy.com/project/badge/Grade/f7f3cf53698e4655ac8895f13fa5dea6)](https://www.codacy.com/app/dwhswenson/contact_map?utm_source=github.com&utm_medium=referral&utm_content=dwhswenson/contact_map&utm_campaign=Badge_Grade)
 [![Maintainability](https://api.codeclimate.com/v1/badges/84768756d594176d8da6/maintainability)](https://codeclimate.com/github/dwhswenson/contact_map/maintainability)
-
+[![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/dwhswenson/contact_map/master)
 # Contact Map Explorer
 
 This package provides tools for analyzing and exploring contacts
@@ -38,6 +38,8 @@ representing the fraction of trajectory time that the contact was present.
 
 Full documentation is at http://contact-map.readthedocs.io/.
 
+Try it out online: [![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/dwhswenson/contact_map/master?filepath=%2Fexamples) (Note: the performance of the online servers can vary widely.)
+
 ## Installation
 
 The easiest way to install is with `conda`. Conda is a powerful package and

ci/conda-recipe/meta.yaml

@@ -1,7 +1,7 @@
 package:
   name: contact_map
   # add ".dev0" for unreleased versions
-  version: "0.6.0"
+  version: "0.7.0"
 
 source:
   path: ../../

contact_map/__init__.py

@@ -6,7 +6,9 @@
 __version__ = version.version
 
 from .contact_map import (
-    ContactMap, ContactFrequency, ContactDifference
+    ContactMap, ContactFrequency, ContactDifference,
+    AtomMismatchedContactDifference, ResidueMismatchedContactDifference,
+    OverrideTopologyContactDifference
 )
 
 from .contact_count import ContactCount

contact_map/atom_indexer.py

@@ -0,0 +1,106 @@
+import collections
+import numpy as np
+import mdtraj as md
+
+def _atom_slice(traj, indices):
+    """Mock MDTraj.atom_slice without rebuilding topology"""
+    xyz = np.array(traj.xyz[:, indices], order='C')
+    topology = traj.topology.copy()
+    if traj._have_unitcell:
+        unitcell_lengths = traj._unitcell_lengths.copy()
+        unitcell_angles = traj._unitcell_angles.copy()
+    else:
+        unitcell_lengths = None
+        unitcell_angles = None
+    time = traj._time.copy()
+
+    # Hackish to make the smart slicing work
+    topology._atoms = indices
+    topology._numAtoms = len(indices)
+    return md.Trajectory(xyz=xyz, topology=topology, time=time,
+                         unitcell_lengths=unitcell_lengths,
+                         unitcell_angles=unitcell_angles)
+
+def residue_query_atom_idxs(sliced_query, atom_idx_to_residue_idx):
+    residue_query_atom_idxs = collections.defaultdict(list)
+    for sliced_idx in sliced_query:
+        residue_idx = atom_idx_to_residue_idx[sliced_idx]
+        residue_query_atom_idxs[residue_idx].append(sliced_idx)
+    return residue_query_atom_idxs
+
+
+class AtomSlicedIndexer(object):
+    """Indexer when using atom slicing.
+    """
+    def __init__(self, topology, real_query, real_haystack, all_atoms):
+        self.all_atoms = all_atoms
+        self.sliced_idx = {
+            real_idx : sliced_idx
+            for sliced_idx, real_idx in enumerate(all_atoms)
+        }
+        self.real_idx = {
+            sliced_idx: real_idx
+            for real_idx, sliced_idx in self.sliced_idx.items()
+        }
+        self.query = set([self.sliced_idx[q] for q in real_query])
+        self.haystack = set([self.sliced_idx[h] for h in real_haystack])
+
+        # atom_idx_to_residue_idx
+        self.real_atom_idx_to_residue_idx = {atom.index: atom.residue.index
+                                             for atom in topology.atoms}
+        self.atom_idx_to_residue_idx = {
+            sliced_idx: self.real_atom_idx_to_residue_idx[real_idx]
+            for sliced_idx, real_idx in enumerate(all_atoms)
+        }
+        self.residue_query_atom_idxs = residue_query_atom_idxs(
+            self.query, self.atom_idx_to_residue_idx
+        )
+
+    def ignore_atom_idx(self, atoms, all_atoms_set):
+        result = set(atom.index for atom in atoms)
+        result &= all_atoms_set
+        result = set(self.sliced_idx[a] for a in result)
+        return result
+
+    def convert_atom_contacts(self, atom_contacts):
+        result =  {frozenset(map(self.real_idx.__getitem__, pair)): value
+                   for pair, value in atom_contacts.items()}
+        return collections.Counter(result)
+
+    def slice_trajectory(self, trajectory):
+        # Prevent (memory) expensive atom slicing if not needed.
+        # This check is also needed here because ContactFrequency slices the
+        # whole trajectory before calling this function.
+        if len(self.all_atoms) < trajectory.topology.n_atoms:
+            sliced = _atom_slice(trajectory, self.all_atoms)
+        else:
+            sliced = trajectory
+        return sliced
+
+
+class IdentityIndexer(object):
+    """Indexer when not using atom slicing.
+    """
+    def __init__(self, topology, real_query, real_haystack, all_atoms):
+        self.all_atoms = all_atoms
+        self.topology = topology
+        identity_mapping = {a: a for a in range(topology.n_atoms)}
+        self.sliced_idx = identity_mapping
+        self.real_idx = identity_mapping
+        self.query = set(real_query)
+        self.haystack = set(real_haystack)
+        self.real_atom_idx_to_residue_idx = {atom.index: atom.residue.index
+                                             for atom in topology.atoms}
+        self.atom_idx_to_residue_idx = self.real_atom_idx_to_residue_idx
+        self.residue_query_atom_idxs = residue_query_atom_idxs(
+            self.query, self.atom_idx_to_residue_idx
+        )
+
+    def ignore_atom_idx(self, atoms, all_atoms_set):
+        return set(atom.index for atom in atoms)
+
+    def convert_atom_contacts(self, atom_contacts):
+        return atom_contacts
+
+    def slice_trajectory(self, trajectory):
+        return trajectory

contact_map/contact_count.py

@@ -1,3 +1,4 @@
+import collections
 import scipy
 import numpy as np
 import pandas as pd
@@ -310,3 +311,15 @@ def most_common_idx(self):
         most_common : same thing, using objects as key
         """
         return self._counter.most_common()
+
+    def filter(self, idx):
+        """New ContactCount filtered to idx.
+
+        Returns a new ContactCount with the only the counter keys/values
+        where both the keys are in idx
+        """
+        dct = {k: v for k, v in self._counter.items()
+               if all([i in idx for i in k])}
+        new_count = collections.Counter()
+        new_count.update(dct)
+        return ContactCount(new_count, self._object_f, self.n_x, self.n_y)