Author: bugman
Date: Wed Oct 8 12:10:04 2014
New Revision: 26207
URL: http://svn.gna.org/viewcvs/relax?rev=26207&view=rev
Log:
Intermediate optimisation results are now stored by the frame order
auto-analysis.
The results from each minimise.grid_search and minimise.execute user function
call are now stored in
specially named directories located in the 'intermediate_results' directory,
which itself is located
in the auto-analysis results_dir directory. This allows intermediate results
to be more easily
analysed later on, which can be useful for optimising the optimisation steps.
These directories can
also be used for the pre_run_dir auto-analysis argument for subsequent
refinements from earlier
steps in the optimisation. The results stored include everything from the
results_output() method
and the count_sobol_points() and summarise() functions.
To allow this to work, the auto-analysis functions count_sobol_points() and
summarise() required
modification. Results files are now always loaded into a temporary data
pipe, rather switching to
the corresponding pipe, and the temporary data pipe is deleted after the data
has been extracted.
The original data pipe name is also stored and a switch back to that pipe
occurs at the end of each
function.
Modified:
branches/frame_order_cleanup/auto_analyses/frame_order.py
Modified: branches/frame_order_cleanup/auto_analyses/frame_order.py
URL:
http://svn.gna.org/viewcvs/relax/branches/frame_order_cleanup/auto_analyses/frame_order.py?rev=26207&r1=26206&r2=26207&view=diff
==============================================================================
--- branches/frame_order_cleanup/auto_analyses/frame_order.py (original)
+++ branches/frame_order_cleanup/auto_analyses/frame_order.py Wed Oct 8
12:10:04 2014
@@ -89,6 +89,9 @@
@type force: bool
"""
+ # Store the current data pipe name.
+ original_pipe = pipes.cdp_name()
+
# The model names, titles and directories, including axis permutations.
models = []
model_titles = []
@@ -126,17 +129,12 @@
if not access(dirs[i]+sep+'results.bz2', F_OK):
continue
- # Switch to the data pipe if it already exists.
- if pipes.has_pipe(models[i]):
- pipes.switch(models[i])
-
- # Otherwise load the data.
- else:
- # Create a data pipe.
- pipes.create(models[i], 'frame order')
-
- # Load the data.
- results.read(file='results', dir=dirs[i])
+ # Create a data pipe.
+ pipe_name = 'temp %s' % models[i]
+ pipes.create(pipe_name, 'frame order')
+
+ # Load the data.
+ results.read(file='results', dir=dirs[i])
# SciPy quadratic integration has been used.
if hasattr(cdp, 'quad_int') and cdp.quad_int:
@@ -151,6 +149,9 @@
count[models[i]] = cdp.sobol_points_used
count_total[models[i]] = cdp.sobol_max_points
percentage[models[i]] = "%10.3f" % (float(cdp.sobol_points_used) /
float(cdp.sobol_max_points) * 100.0) + '%'
+
+ # Delete the temporary data pipe.
+ pipes.delete(pipe_name)
# Initialise the output string.
string = "Quasi-random Sobol' numerical PCS integration point
counting:\n\n"
@@ -175,6 +176,10 @@
file.write(string)
file.close()
+ # Switch back to the original data pipe, if it exists.
+ if original_pipe:
+ pipes.switch(original_pipe)
+
def model_directory(model, base_dir=None):
"""Return the directory to be used for the model.
@@ -212,6 +217,9 @@
@keyword force: A flag which if True will cause any preexisting
file to be overwritten.
@type force: bool
"""
+
+ # Store the current data pipe name.
+ original_pipe = pipes.cdp_name()
# The model names, titles and directories, including axis permutations.
models = []
@@ -260,17 +268,12 @@
if not access(dirs[i]+sep+'results.bz2', F_OK):
continue
- # Switch to the data pipe if it already exists.
- if pipes.has_pipe(models[i]):
- pipes.switch(models[i])
-
- # Otherwise load the data.
- else:
- # Create a data pipe.
- pipes.create(models[i], 'frame order')
-
- # Load the data.
- results.read(file='results', dir=dirs[i])
+ # Create a data pipe.
+ pipe_name = 'temp %s' % models[i]
+ pipes.create(pipe_name, 'frame order')
+
+ # Load the data.
+ results.read(file='results', dir=dirs[i])
# Number of params.
k = len(cdp.params)
@@ -297,7 +300,7 @@
# Convert the axis alpha angle to spherical angles for comparison.
if hasattr(cdp, 'axis_alpha') and cdp.model in [MODEL_ROTOR,
MODEL_FREE_ROTOR]:
- axis_theta, axis_phi =
convert_axis_alpha_to_spherical(alpha=cdp.axis_alpha,
pivot=generate_pivot(order=1, pipe_name=models[i]),
point=pipe_centre_of_mass(verbosity=0))
+ axis_theta, axis_phi =
convert_axis_alpha_to_spherical(alpha=cdp.axis_alpha,
pivot=generate_pivot(order=1, pipe_name=pipe_name),
point=pipe_centre_of_mass(verbosity=0))
contents1[-1][5] = wrap_angles(axis_theta, 0.0, 2.0*pi)
contents1[-1][6] = wrap_angles(axis_phi, 0.0, 2.0*pi)
@@ -338,6 +341,9 @@
contents2[-1][8] = cdp.pivot_y
contents2[-1][9] = cdp.pivot_z
+ # Delete the temporary data pipe.
+ pipes.delete(pipe_name)
+
# Add the tables.
string += format_table(headings=headings1, contents=contents1,
custom_format=[None, None, "%.2f", "%.2f", "%.3f", "%.3f", "%.3f", "%.2f",
"%.2f", "%.2f", "%.2f"])
string += format_table(headings=headings2, contents=contents2,
custom_format=[None, "%.3f", "%.3f", "%.3f", "%.3f", "%.3f", "%.3f", "%.3f",
"%.3f", "%.3f"])
@@ -350,6 +356,10 @@
file = open_write_file(file_name=file_name, dir=dir, force=force)
file.write(string)
file.close()
+
+ # Switch back to the original data pipe, if it exists.
+ if original_pipe:
+ pipes.switch(original_pipe)
@@ -923,34 +933,72 @@
# Printout.
subsubtitle(file=sys.stdout, text="Optimisation using the
PCS subset")
+ # Results directory stub for intermediate results.
+ intermediate_stub = self.results_dir + sep +
'intermediate_results' + sep + 'pcs_subset'
+
# Zooming grid search.
for i in opt.loop_grid():
+ # The intermediate results directory.
+ intermediate_dir = intermediate_stub + '_grid%i' % i
+
# Set the zooming grid search level.
zoom = opt.get_grid_zoom_level(i)
if zoom != None:
self.interpreter.minimise.grid_zoom(level=zoom)
-
- # The numerical optimisation settings.
-
self.interpreter.frame_order.quad_int(opt.get_grid_quad_int(i))
- self.sobol_setup(opt.get_grid_sobol_info(i))
+ intermediate_dir += '_zoom%i' % zoom
# Set up the custom grid increments.
incs = self.custom_grid_incs(model,
inc=opt.get_grid_inc(i))
+ intermediate_dir += '_inc%i' % opt.get_grid_inc(i)
+
+ # The numerical optimisation settings.
+ quad_int = opt.get_grid_quad_int(i)
+ if quad_int:
+ self.interpreter.frame_order.quad_int()
+ intermediate_dir += '_quad_int'
+ else:
+ sobol_num = opt.get_grid_sobol_info(i)
+ self.sobol_setup(sobol_num)
+ intermediate_dir += '_sobol%i' % sobol_num[0]
# Perform the grid search.
self.interpreter.minimise.grid_search(inc=incs)
+ # Store the intermediate results and statistics.
+ self.results_output(model=model,
dir=model_directory(model, base_dir=intermediate_dir), results_file=True)
+ count_sobol_points(dir=intermediate_dir, force=True)
+ summarise(dir=intermediate_dir, force=True)
+
# Minimise (for the PCS data subset and full RDC set).
for i in opt.loop_min():
+ # The intermediate results directory.
+ func_tol = opt.get_min_func_tol(i)
+ max_iter = opt.get_min_max_iter(i)
+ intermediate_dir = intermediate_stub +
'_min%i_ftol%g_max_iter%i' % (i, func_tol, max_iter)
+
# The numerical optimisation settings.
-
self.interpreter.frame_order.quad_int(opt.get_min_quad_int(i))
- self.sobol_setup(opt.get_min_sobol_info(i))
+ quad_int = opt.get_min_quad_int(i)
+ if quad_int:
+ self.interpreter.frame_order.quad_int()
+ intermediate_dir += '_quad_int'
+ else:
+ sobol_num = opt.get_min_sobol_info(i)
+ self.sobol_setup(sobol_num)
+ intermediate_dir += '_sobol%i' % sobol_num[0]
# Perform the optimisation.
-
self.interpreter.minimise.execute(min_algor=opt.get_min_algor(i),
func_tol=opt.get_min_func_tol(i), max_iter=opt.get_min_max_iter(i))
+
self.interpreter.minimise.execute(min_algor=opt.get_min_algor(i),
func_tol=func_tol, max_iter=max_iter)
+
+ # Store the intermediate results.
+ self.results_output(model=model,
dir=model_directory(model, base_dir=intermediate_dir), results_file=True)
+ count_sobol_points(dir=intermediate_dir, force=True)
+ summarise(dir=intermediate_dir, force=True)
# Printout.
subsubtitle(file=sys.stdout, text="Optimisation using the full
data set")
+
+ # Results directory stub for intermediate results.
+ intermediate_stub = self.results_dir + sep +
'intermediate_results' + sep + 'all_data'
# Copy the PCS data.
self.interpreter.pcs.copy(pipe_from=self.data_pipe_full,
pipe_to=self.pipe_name_dict[model])
@@ -967,12 +1015,28 @@
opt = self.opt_full
if opt != None:
for i in opt.loop_min():
+ # The intermediate results directory.
+ func_tol = opt.get_min_func_tol(i)
+ max_iter = opt.get_min_max_iter(i)
+ intermediate_dir = intermediate_stub +
'_min%i_ftol%g_max_iter%i' % (i, func_tol, max_iter)
+
# The numerical optimisation settings.
-
self.interpreter.frame_order.quad_int(opt.get_min_quad_int(i))
- self.sobol_setup(opt.get_min_sobol_info(i))
+ quad_int = opt.get_min_quad_int(i)
+ if quad_int:
+ self.interpreter.frame_order.quad_int()
+ intermediate_dir += '_quad_int'
+ else:
+ sobol_num = opt.get_min_sobol_info(i)
+ self.sobol_setup(sobol_num)
+ intermediate_dir += '_sobol%i' % sobol_num[0]
# Perform the optimisation.
-
self.interpreter.minimise.execute(min_algor=opt.get_min_algor(i),
func_tol=opt.get_min_func_tol(i), max_iter=opt.get_min_max_iter(i))
+
self.interpreter.minimise.execute(min_algor=opt.get_min_algor(i),
func_tol=func_tol, max_iter=max_iter)
+
+ # Store the intermediate results.
+ self.results_output(model=model,
dir=model_directory(model, base_dir=intermediate_dir), results_file=True)
+ count_sobol_points(dir=intermediate_dir, force=True)
+ summarise(dir=intermediate_dir, force=True)
# Results printout.
self.print_results()
r26207 - /branches/frame_order_cleanup/auto_analyses/frame_order.py