Menu

SVN archive Commit Log


Commit Date  
[r3268] by varioustoxins

almost feature complete mpi4py and uni processor versions of relax

2007-04-17 12:39:25 Tree
[r3267] by varioustoxins

fully operative grid search (uniprocessor broken)

2007-04-16 22:54:48 Tree
[r3266] by varioustoxins

partial integration of distributed grid code (everything broken)

2007-04-12 16:48:28 Tree
[r3265] by varioustoxins

correct version of r3264 correction to bad indenting in fix in r3262

2007-04-10 21:52:14 Tree
[r3264] by varioustoxins

changes to grid search for parallelisation and code cleanup

2007-04-10 21:42:49 Tree
[r3263] by varioustoxins

correction to bad indenting in fix in r3262 and removal of errant print
statement

2007-04-10 21:41:33 Tree
[r3262] by bugman

Bug fix for the model-free results.write() problem.

This bug was reported by Gary Thompson at
https://mail.gna.org/public/relax-devel/2007-03/msg00176.html (Message-id:
<460CE499.9070905@...>). The thread continues into the next month
(https://mail.gna.org/public/relax-devel/) starting at
https://mail.gna.org/public/relax-devel/2007-04/msg00028.html (Message-id:
<460CE499.9070905@...>).

The fix was to synchronise the relaxation data assembly in the value and simulation sections of the
write_columnar_results() function. The 'num_ri' test was missing from the simulation section!

2007-04-10 10:41:42 Tree
[r3261] by varioustoxins

Major update - monte carlo simulations supported (trivially)
also implimented: chunky queues, exception throwing on slaves, and rank
annotated output. note uniprocessor support currently broken.

2007-04-07 20:11:21 Tree
[r3260] by macraild

A further attempt to make boolean mol-res-spin selection work. This is
still some way from a solution.

2007-04-03 17:19:34 Tree
[r3259] by bugman

Blocked revisions 3173-3194,3196-3197,3200-3207,3209,3211-3234,3240,3245,3247,3255-3258 via svnmerge

........
r3173 | varioustoxins | 2007-03-13 20:35:15 +1100 (Tue, 13 Mar 2007) | 2 lines

directory independant running of unit tests, improved docs, code cleanup
........
r3174 | varioustoxins | 2007-03-13 20:58:31 +1100 (Tue, 13 Mar 2007) | 2 lines

remove annoying debug messages
........
r3175 | bugman | 2007-03-13 21:07:52 +1100 (Tue, 13 Mar 2007) | 3 lines

Added the 'test_mol_res_spin' module to the __all__ package variable.
........
r3176 | varioustoxins | 2007-03-13 21:17:40 +1100 (Tue, 13 Mar 2007) | 3 lines

corrected lack of return from run function
........
r3177 | varioustoxins | 2007-03-13 21:56:38 +1100 (Tue, 13 Mar 2007) | 2 lines

improved messages for bad input
........
r3178 | bugman | 2007-03-14 13:45:52 +1100 (Wed, 14 Mar 2007) | 7 lines

Fixed the test suite synopsis (the system/functional tests weren't returning the test result).

All the contents of the System_tests.__init__() method has been placed into the run() method as
__init__() needs to return None. This run() method then returns the global_pass flag to the
Test_suite_runner class or test_suite/runner.py.
........
r3179 | bugman | 2007-03-14 16:07:21 +1100 (Wed, 14 Mar 2007) | 3 lines

Removed the svn:executable property from the 'maths_fns/test_chi2.py' unit test file.
........
r3180 | bugman | 2007-03-14 16:09:52 +1100 (Wed, 14 Mar 2007) | 3 lines

Removed the line '#!/usr/bin/env python' from all the unit test modules.
........
r3181 | bugman | 2007-03-14 17:08:07 +1100 (Wed, 14 Mar 2007) | 3 lines

Replaced the empty chi2 and dchi2 unit tests with true unit tests for those functions.
........
r3182 | bugman | 2007-03-14 17:32:17 +1100 (Wed, 14 Mar 2007) | 3 lines

Improvements to the docstrings and a comment in the chi-squared value and gradient unit tests.
........
r3183 | bugman | 2007-03-14 17:33:10 +1100 (Wed, 14 Mar 2007) | 3 lines

Added a unit test for the chi-squared Hessian function d2chi2.
........
r3184 | bugman | 2007-03-14 17:37:39 +1100 (Wed, 14 Mar 2007) | 3 lines

More docstring improvements for the chi-square function unit tests.
........
r3185 | bugman | 2007-03-14 17:38:11 +1100 (Wed, 14 Mar 2007) | 3 lines

Added a tearDown statement to delete the chi-squared Hessian.
........
r3186 | bugman | 2007-03-14 17:41:18 +1100 (Wed, 14 Mar 2007) | 3 lines

Simplified the chi2 gradient unit test and added delete statements to destroy all unit test data.
........
r3187 | bugman | 2007-03-14 19:17:32 +1100 (Wed, 14 Mar 2007) | 3 lines

Completion of point 9 of the preparation in the data model redesign document.
........
r3188 | bugman | 2007-03-14 19:30:56 +1100 (Wed, 14 Mar 2007) | 7 lines

Added initial support for the residue user function class.

The file 'prompt/residue.py' was copied from 'prompt/sequence.py' and the function 'add()' renamed
to 'create' (generating the 'residue.create()' user function). All the other sequence user
functions still remain unchanged.
........
r3189 | bugman | 2007-03-14 19:36:51 +1100 (Wed, 14 Mar 2007) | 3 lines

Created the 'generic_fns.residue' module and started converting the sequence add() fn to create().
........
r3190 | bugman | 2007-03-14 19:37:58 +1100 (Wed, 14 Mar 2007) | 3 lines

Converted the Residue.create() method into a function of the 'generic_fns.residue' module.
........
r3191 | bugman | 2007-03-14 19:38:21 +1100 (Wed, 14 Mar 2007) | 3 lines

Removed the 'self' argument to the create() function.
........
r3192 | bugman | 2007-03-14 19:42:44 +1100 (Wed, 14 Mar 2007) | 3 lines

Added an import statement for the 'residue' module of the 'generic_fns' package.
........
r3193 | bugman | 2007-03-14 19:53:49 +1100 (Wed, 14 Mar 2007) | 3 lines

A few changes to make the 'residue.create()' user function partially functional.
........
r3194 | varioustoxins | 2007-03-14 21:25:08 +1100 (Wed, 14 Mar 2007) | 4 lines

updates to unit_test_runner, code cleanup, more command line options
changes for the change in name of unit_tests.unit_test_runner.Unit_test_runner
........
r3196 | bugman | 2007-03-15 15:24:38 +1100 (Thu, 15 Mar 2007) | 3 lines

Shifted the unit tests for the relax data storage singleton to the file 'test___init__.py'.
........
r3197 | bugman | 2007-03-15 15:57:50 +1100 (Thu, 15 Mar 2007) | 5 lines

Manually reverted most of r3157.

The 'add()' method of the relax data storage singleton has been reintroduced.
........
r3200 | macraild | 2007-03-16 03:58:23 +1100 (Fri, 16 Mar 2007) | 9 lines

Alternative implimentation of the data object as a singleton.

As proposed here:
https://mail.gna.org/public/relax-devel/2007-03/msg00013.html

and agreed here:
https://mail.gna.org/public/relax-devel/2007-03/msg00058.html#

........
r3201 | bugman | 2007-03-16 17:45:47 +1100 (Fri, 16 Mar 2007) | 2 lines

Fixes for the unit tests of the relax data storage singleton methods.
........
r3202 | bugman | 2007-03-16 17:46:44 +1100 (Fri, 16 Mar 2007) | 2 lines

Wrote two unit tests for testing the generic_fns.residue.create() function.
........
r3203 | bugman | 2007-03-16 17:49:27 +1100 (Fri, 16 Mar 2007) | 2 lines

Modified generic_fns.residue.create() to replace the empty first residue if no residues exist.
........
r3204 | bugman | 2007-03-16 17:50:42 +1100 (Fri, 16 Mar 2007) | 6 lines

Important fix for the PipeContainer data structure.

All of the objects of the PipeContainer where being instantiated as objects of the class rather than
the separate instances. Hence two different PipeContainer instances contained the same objects!
All object instantiation is now done within the __init__() method to avoid the problem.
........
r3205 | bugman | 2007-03-16 17:53:23 +1100 (Fri, 16 Mar 2007) | 4 lines

Fixed the unit tests for the 'generic_fns.pipes' module.

The relax data storage singleton object is now reset at the start of the setUp unit test method.
........
r3206 | bugman | 2007-03-16 17:56:02 +1100 (Fri, 16 Mar 2007) | 4 lines

Created the 'generic_fns.residue.copy()' function.

Two unit tests for this function have been added.
........
r3207 | bugman | 2007-03-16 17:57:46 +1100 (Fri, 16 Mar 2007) | 2 lines

Wrote unit tests for the currently non-existant 'generic_fns.residue.delete()' function.
........
r3209 | macraild | 2007-03-17 00:46:23 +1100 (Sat, 17 Mar 2007) | 6 lines

Further modifications for compatibility with the new relax Data object.

The data unit-tests have been fixed, and all import statements in client
modules updated.
........
r3211 | bugman | 2007-03-17 12:48:15 +1100 (Sat, 17 Mar 2007) | 5 lines

Updated the residue.delete() user function prompt code.

The code was still that of sequence.delete()!
........
r3212 | bugman | 2007-03-17 12:51:44 +1100 (Sat, 17 Mar 2007) | 3 lines

Copied the 'residue.py' files to implement the new 'spin' and 'molecule' user function classes.
........
r3213 | bugman | 2007-03-17 12:52:56 +1100 (Sat, 17 Mar 2007) | 6 lines

Removed the executable property from the test_float.py unit test module.

The command used was:
$ svn propdel svn:executable test_float.py
........
r3214 | bugman | 2007-03-17 12:55:35 +1100 (Sat, 17 Mar 2007) | 3 lines

Removed the 'if __name__ == ...' statements and added a copyright statement for Gary.
........
r3215 | bugman | 2007-03-17 12:57:34 +1100 (Sat, 17 Mar 2007) | 3 lines

Removed to unused comment lines.
........
r3216 | bugman | 2007-03-17 12:58:48 +1100 (Sat, 17 Mar 2007) | 3 lines

Added to empty unit test modules.
........
r3217 | bugman | 2007-03-19 15:47:57 +1100 (Mon, 19 Mar 2007) | 18 lines

Implemented the generic_fns.residue.delete() function (although it doesn't work yet).

This function has been coded to use the UCSF selection syntax as described in Gary Thompson's post
located at https://mail.gna.org/public/relax-devel/2007-01/msg00014.html (Message-id:
<f001463a0701071417w6bd7927cp8fdd052e698575ec@...>).

The function makes use of three new selection functions - tokenise, parse_token, and molecule_loop.
The first two are described at https://mail.gna.org/public/relax-devel/2007-01/msg00036.html
(Message-ID: <1168883717.7569.511.camel@...>). The last function is similar to the spin_loop
function described in the same post. All three are yet to be implemented (although stubs have been
created for the functions).

The RelaxError 'RelaxSpinSelectDisallowError' has been created to prevent spins from being selected
by the user function.

The unit tests for generic_fns.residue.delete() have been updated and two more added:
test_delete_all() and test_delete_shift().
........
r3218 | bugman | 2007-03-19 15:56:04 +1100 (Mon, 19 Mar 2007) | 4 lines

Converted all the class methods of the 'generic_fns.selection' module into functions.

The unsel_*() methods have been renamed to desel_*() and the functions alphabetically sorted.
........
r3219 | bugman | 2007-03-19 15:59:48 +1100 (Mon, 19 Mar 2007) | 2 lines

Wrote a unit test for the generic_fns.selection.reverse() function.
........
r3220 | bugman | 2007-03-19 16:04:31 +1100 (Mon, 19 Mar 2007) | 2 lines

Bug fix. Removed the import of the now non-existant Selection class.
........
r3221 | bugman | 2007-03-19 16:19:33 +1100 (Mon, 19 Mar 2007) | 5 lines

Fixed the select.reverse() user function.

The generic_fns.selection.reverse() function still does not work as the spin_loop() function needs
to be implemented.
........
r3222 | bugman | 2007-03-19 16:20:13 +1100 (Mon, 19 Mar 2007) | 12 lines

Wrote the spin_loop() generator function for looping over and returning spin containers.

For more info, see my post at https://mail.gna.org/public/relax-devel/2006-10/msg00057.html
(Message-id: <1160557041.9523.74.camel@...>). The idea of the
generator function, using the yield statement, to return the spin data containers was given by Gary
Thompson at https://mail.gna.org/public/relax-devel/2007-01/msg00014.html (Message-id:
<f001463a0701071417w6bd7927cp8fdd052e698575ec@...>). More details are given by Chris
MacRaild at https://mail.gna.org/public/relax-devel/2007-01/msg00036.html (Message-ID:
<1168883717.7569.511.camel@...>).

The full generator function is generic_fns.selection.spin_loop().
........
r3223 | bugman | 2007-03-19 16:20:44 +1100 (Mon, 19 Mar 2007) | 4 lines

Wrote the molecule_loop() generator function for looping over and returning molecule containers.

The full generator function is generic_fns.selection.molecule_loop().
........
r3224 | bugman | 2007-03-19 16:21:14 +1100 (Mon, 19 Mar 2007) | 2 lines

Created the RelaxMolSelectDisallowError and RelaxResSelectDisallowError errors.
........
r3225 | bugman | 2007-03-19 16:21:47 +1100 (Mon, 19 Mar 2007) | 4 lines

Multiple unit tests for the spin_loop(), residue_loop(), and molecule_loop() generator functions.

These generator functions belong to the 'generic_fns.selection' module.
........
r3226 | bugman | 2007-03-19 16:22:33 +1100 (Mon, 19 Mar 2007) | 6 lines

Bug fix.

The spin_loop() and molecule_loop() generator functions now accept zero arguments for the selection
of everything. This fixes two of the unit tests. The test_molecule_loop_no_selection() unit test
has also been fixed.
........
r3227 | bugman | 2007-03-19 16:23:18 +1100 (Mon, 19 Mar 2007) | 2 lines

Implemented the generic_fns.selection.residue_loop() generator function.
........
r3228 | bugman | 2007-03-19 16:23:48 +1100 (Mon, 19 Mar 2007) | 2 lines

Small bug fix for the test_residue_loop_no_selection() unit test.
........
r3229 | bugman | 2007-03-19 16:24:23 +1100 (Mon, 19 Mar 2007) | 4 lines

Wrote the generic_fns.selection.tokenise() function and 28 unit tests covering the function!

The unit tests attempt to cover all proper and improper combinations of the selection string.
........
r3230 | bugman | 2007-03-19 16:25:03 +1100 (Mon, 19 Mar 2007) | 5 lines

Wrote the generic_fns.selection.parse_token() function and 13 unit tests covering the function!

The unit tests attempt to cover all proper and improper combinations of the supplied token including
numbers (integers), names, and number ranges.
........
r3231 | bugman | 2007-03-19 16:25:57 +1100 (Mon, 19 Mar 2007) | 5 lines

Updated the data model redesign document.

The spin_loop() generator function and selection syntax parsing functions tokenise() and
parse_token() are now implemented.
........
r3232 | bugman | 2007-03-19 16:26:45 +1100 (Mon, 19 Mar 2007) | 5 lines

Wrote the identification string documentation.

This documentation is the string 'generic_fns.selection.id_string_doc'. The residue.delete() user
function docstring has had this string added to the end.
........
r3233 | bugman | 2007-03-19 16:27:21 +1100 (Mon, 19 Mar 2007) | 4 lines

Fixed the generic_fns.residue.delete() function and its associated unit tests.

An additional unit test for deletion using residue names has been added.
........
r3234 | bugman | 2007-03-19 17:20:17 +1100 (Mon, 19 Mar 2007) | 3 lines

Updated the unit test runner to be cross platform compatible.
........
r3240 | bugman | 2007-03-26 15:25:28 +1000 (Mon, 26 Mar 2007) | 10 lines

Bug fix.

This issue was reported by Gary Thompson at
https://mail.gna.org/public/relax-devel/2007-03/msg00084.html (Message-id:
<45FE68E4.4010000@...>).

The module separator character '.' had been replace by os.curdir in a number of places in the unit
test runner. This has now been fixed.
........
r3245 | macraild | 2007-03-30 02:46:15 +1000 (Fri, 30 Mar 2007) | 13 lines

A proposal for a modified mol-res-spin selection mechanism.

This simple modification, based on a selection object, permits
the use of booleans in selection strings, eg:

':leu@... | :val@...,CG2'

selects leucine and valine methyls.

Also, modification of Selection.__contains__() will allow for
wildcards, regex, etc as desired.
........
r3247 | macraild | 2007-03-30 04:40:50 +1000 (Fri, 30 Mar 2007) | 3 lines

This time booleans actually work!
........
r3255 | macraild | 2007-03-30 23:35:39 +1000 (Fri, 30 Mar 2007) | 3 lines

Improvements and unit-tests for spin selection
........
r3256 | varioustoxins | 2007-03-30 23:55:16 +1000 (Fri, 30 Mar 2007) | 3 lines

unit_test_runner.py no longersilently ignores TestCases with syntax
errors. They now produce a dummy test case that prints the SyntaxError
........
r3257 | varioustoxins | 2007-03-31 00:11:44 +1000 (Sat, 31 Mar 2007) | 3 lines

unit_test_runner emits a better warning messages that no tests from
a module with a syntax error will be run
........
r3258 | macraild | 2007-03-31 02:41:01 +1000 (Sat, 31 Mar 2007) | 3 lines

Further mol-res-spin selection improvements. Work still needed on
boolean logic.
........

2007-04-03 11:21:47 Tree
Older >
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.