Author: bugman
Date: Mon Jan 2 19:29:00 2012
New Revision: 15117
URL: http://svn.gna.org/viewcvs/relax?rev=15117&view=rev
Log:
Resigned the internal structural object atom_loop() method to be up to 20
times faster.
The base model_loop() method is no longer used at all. It was previously
being used in a nested
fashion, making the loop extremely slow when lots of models exist.
Modified:
branches/frame_order_testing/generic_fns/structure/internal.py
Modified: branches/frame_order_testing/generic_fns/structure/internal.py
URL:
http://svn.gna.org/viewcvs/relax/branches/frame_order_testing/generic_fns/structure/internal.py?rev=15117&r1=15116&r2=15117&view=diff
==============================================================================
--- branches/frame_order_testing/generic_fns/structure/internal.py (original)
+++ branches/frame_order_testing/generic_fns/structure/internal.py Mon Jan 2
19:29:00 2012
@@ -669,7 +669,7 @@
self.structural_data[i].mol.add_item(mol_name=name,
mol_cont=MolContainer())
- def atom_loop(self, atom_id=None, str_id=None, model_num=None,
model_num_flag=False, mol_name_flag=False, res_num_flag=False,
res_name_flag=False, atom_num_flag=False, atom_name_flag=False,
element_flag=False, pos_flag=False, ave=False):
+ def atom_loop(self, atom_id=None, str_id=None, model_num=None,
mol_name_flag=False, res_num_flag=False, res_name_flag=False,
atom_num_flag=False, atom_name_flag=False, element_flag=False,
pos_flag=False, ave=False):
"""Generator function for looping over all atoms in the internal
relax structural object.
@keyword atom_id: The molecule, residue, and atom
identifier string. Only atoms matching this selection will be yielded.
@@ -678,8 +678,6 @@
@type str_id: str, int, or None
@keyword model_num: Only loop over a specific model.
@type model_num: int or None
- @keyword model_num_flag: A flag which if True will cause the
model number to be yielded.
- @type model_num_flag: bool
@keyword mol_name_flag: A flag which if True will cause the
molecule name to be yielded.
@type mol_name_flag: bool
@keyword res_num_flag: A flag which if True will cause the
residue number to be yielded.
@@ -709,80 +707,98 @@
if atom_id:
sel_obj = Selection(atom_id)
- # Model loop.
- for model in self.model_loop(model_num):
- # Loop over the molecules.
- for mol_index in range(len(model.mol)):
- mol = model.mol[mol_index]
-
- # Skip non-matching molecules.
- if sel_obj and not sel_obj.contains_mol(mol.mol_name):
+ # Obtain all data from the first model (except the position data).
+ model = self.structural_data[0]
+
+ # Loop over the molecules.
+ for mol_index in range(len(model.mol)):
+ mol = model.mol[mol_index]
+
+ # Skip non-matching molecules.
+ if sel_obj and not sel_obj.contains_mol(mol.mol_name):
+ continue
+
+ # Loop over all atoms.
+ for i in xrange(len(mol.atom_name)):
+ # Skip non-matching atoms.
+ if sel_obj and not sel_obj.contains_spin(mol.atom_num[i],
mol.atom_name[i], mol.res_num[i], mol.res_name[i], mol.mol_name):
continue
- # Loop over all atoms.
- for i in xrange(len(mol.atom_name)):
- # Skip non-matching atoms.
- if sel_obj and not
sel_obj.contains_spin(mol.atom_num[i], mol.atom_name[i], mol.res_num[i],
mol.res_name[i], mol.mol_name):
- continue
-
- # Initialise.
- res_num = mol.res_num[i]
- res_name = mol.res_name[i]
- atom_num = mol.atom_num[i]
- atom_name = mol.atom_name[i]
- element = mol.element[i]
- pos = zeros(3, float64)
-
- # The atom position.
+ # Initialise.
+ res_num = mol.res_num[i]
+ res_name = mol.res_name[i]
+ atom_num = mol.atom_num[i]
+ atom_name = mol.atom_name[i]
+ element = mol.element[i]
+
+ # The atom position.
+ if pos_flag:
+ # Average the position.
if ave:
+ # Initialise.
+ pos = zeros(3, float64)
+
# Loop over the models.
- for model in self.model_loop():
+ for j in range(len(self.structural_data)):
+ # A single model.
+ if model_num != None and
self.structural_data[j].num != model_num:
+ continue
+
# Alias.
- mol = model.mol[mol_index]
+ mol2 = self.structural_data[j].mol[mol_index]
# Some sanity checks.
- if mol.atom_num[i] != atom_num:
+ if mol2.atom_num[i] != atom_num:
raise RelaxError("The loaded structures do
not contain the same atoms. The average structural properties can not be
calculated.")
# Sum the atom positions.
- pos = pos + array([mol.x[i], mol.y[i],
mol.z[i]], float64)
+ pos = pos + array([mol2.x[i], mol2.y[i],
mol2.z[i]], float64)
# Average the position array (divide by the number
of models).
pos = pos / len(self.structural_data)
+
+ # All positions.
else:
- pos = array([mol.x[i], mol.y[i], mol.z[i]], float64)
-
- # The molecule name.
- mol_name = mol.mol_name
-
- # Build the tuple to be yielded.
- atomic_tuple = ()
- if model_num_flag:
- if ave:
- atomic_tuple = atomic_tuple + (None,)
- else:
- atomic_tuple = atomic_tuple + (model.num,)
- if mol_name_flag:
- atomic_tuple = atomic_tuple + (mol_name,)
- if res_num_flag:
- atomic_tuple = atomic_tuple + (res_num,)
- if res_name_flag:
- atomic_tuple = atomic_tuple + (res_name,)
- if atom_num_flag:
- atomic_tuple = atomic_tuple + (atom_num,)
- if atom_name_flag:
- atomic_tuple = atomic_tuple + (atom_name,)
- if element_flag:
- atomic_tuple = atomic_tuple + (element,)
- if pos_flag:
- atomic_tuple = atomic_tuple + (pos,)
-
- # Yield the information.
- yield atomic_tuple
-
- # Break out of the loop if the ave flag is set, as data from
only one model is used.
- if ave:
- break
+ # Initialise.
+ pos = []
+
+ # Loop over the models.
+ for j in range(len(self.structural_data)):
+ # A single model.
+ if model_num != None and
self.structural_data[j].num != model_num:
+ continue
+
+ # Alias.
+ mol2 = self.structural_data[j].mol[mol_index]
+
+ # Append the position.
+ pos.append([mol2.x[i], mol2.y[i], mol2.z[i]])
+
+ # Convert.
+ pos = array(pos, float64)
+
+ # The molecule name.
+ mol_name = mol.mol_name
+
+ # Build the tuple to be yielded.
+ atomic_tuple = ()
+ if mol_name_flag:
+ atomic_tuple = atomic_tuple + (mol_name,)
+ if res_num_flag:
+ atomic_tuple = atomic_tuple + (res_num,)
+ if res_name_flag:
+ atomic_tuple = atomic_tuple + (res_name,)
+ if atom_num_flag:
+ atomic_tuple = atomic_tuple + (atom_num,)
+ if atom_name_flag:
+ atomic_tuple = atomic_tuple + (atom_name,)
+ if element_flag:
+ atomic_tuple = atomic_tuple + (element,)
+ if pos_flag:
+ atomic_tuple = atomic_tuple + (pos,)
+
+ # Yield the information.
+ yield atomic_tuple
def bond_vectors(self, attached_atom=None, model_num=None,
mol_name=None, res_num=None, res_name=None, spin_num=None, spin_name=None,
return_name=False, return_warnings=False):
r15117 - /branches/frame_order_testing/generic_fns/structure/internal.py