Reworked CalculateProjectSASA.py #338

schwancr · 2014-02-07T22:37:07Z

Basically it doesn't make sense to save a single file for an entire dataset because the result of mdtraj.geometry.sasa.shrake_rupley is an array with SASA for each ATOM in the trajectory. So the script will save a file for each trajectory in an input range. Additionally, I removed the atom indices flag because I think most people will use it incorrectly.

…nce as well as the test function

…res passing the ProjectInfo.yaml file as well as the stride used in clustering since the shape of the assignments file depends on the trajectory lengths in the project info

AssignHierarchical.py Fix

…xist. The builder actually assumed that the error was a corrupt xtc file, and so it was stuck in an infinite loop trying to throw away the last frame in the xtc.

…s is harder than I thought it would be ...

…inally in the builder.

… index to save to. Also some other minor changes

…s is harder than I thought it would be ...

…inally in the builder.

… index to save to. Also some other minor changes

…g to read the first file in an empty list. I added a check in the while corrupted loop that will break when there are no files left. This results in a RuntimeError, which is caught by both _update_traj and _load_new_files and correctly just does not add a trajectory to the project. Before the entire script would crash because this error was not caught correctly. I did change it from ValueError, but I have no idea what would be raising that error. This is at the very least more transparent, but some xtc error may still cause a crash.

…to load the pdb from disk for every trajectory

…an discard redundant frames in a reasonable manner. The trouble is, the code that is in there now DOESN'T WORK...

… with redundant frames.

…oj_builder

schwancr · 2014-02-07T22:38:51Z

Personally I think it's better to save a lot of data to disk (especially when your dataset is huge) and then analyze that data after the fact.

i.e. Save the SASA for each atom to disk, and then analyze those results for the total Trp SASA, for instance.

rmcgibbo · 2014-02-07T22:39:26Z

scripts/CalculateProjectSASA.py


 if __name__ == '__main__':
+    parser = arglib.ArgumentParser(description="""Calculates the Solvent Accessible Surface Area


Can you put the parser declaration outside of if __name__ == '__main__? This is how the msmb script is able to collect all the parsers

yea I didn't realize that

On Fri, Feb 7, 2014 at 2:39 PM, Robert McGibbon [email protected]:

In scripts/CalculateProjectSASA.py:

if name == 'main':

parser = arglib.ArgumentParser(description="""Calculates the Solvent Accessible Surface Area

Can you put the parser declaration outside of if name == 'main?
This is how the msmb script is able to collect all the parsershttps://github.com/SimTk/msmbuilder/blob/master/scripts/msmb#L37

Reply to this email directly or view it on GitHubhttps://github.com//pull/338/files#r9558142
.

rmcgibbo · 2014-02-07T22:49:53Z

scripts/CalculateProjectSASA.py

+    trajectory""", default='sasa')
+parser.add_argument('which', help="""which trajectories to calculate the SASA for.
+    This script saves a separate file for each trajectory.""", default=[0, np.inf],
+    nargs=2, type=int)


This isn't supposed to be a list of the indices you want to use, its the start and stop indices. Maybe change the help text to something like:

help="Indices of the initial and final trajectories on which to run SASA calculation. This script saves separate SASA files for each trajectory. Example: '--which 0 10' will calculate the SASA only for the first 10 trajectories in your project."

Or maybe they should just be two separate arguments...

bugfix in CalculateProjectDistance.py that wasn't using mdtraj

Add more description to ergodic trim docstring

Fixed more description to ergodic trim docstring

Simplify setup.py

kyleabeauchamp · 2014-02-11T15:10:49Z

I think this docstring sounds like we're outputting just a single HDF file:

Calculates the Solvent Accessible Surface Area
of all atoms in a subset of the trajectories, or for all trajectories in the project. The
output is a hdf5 file which contains the SASA for each atom in each frame
in each trajectory (or the single trajectory you passed in.

Otherwise looks ok

schwancr · 2014-02-11T15:27:23Z

Yea I didn't change it from the last version. Will do

—
Sent from Mailbox for iPhone

On Tue, Feb 11, 2014 at 7:10 AM, kyleabeauchamp [email protected]
wrote:

I think this docstring sounds like we're outputting just a single HDF file:
Calculates the Solvent Accessible Surface Area
of all atoms in a subset of the trajectories, or for all trajectories in the project. The
output is a hdf5 file which contains the SASA for each atom in each frame
in each trajectory (or the single trajectory you passed in.
Otherwise looks ok

Reply to this email directly or view it on GitHub:
#338 (comment)

rmcgibbo · 2014-03-29T00:15:11Z

bump.

schwancr · 2014-03-29T19:38:30Z

This is going to be a pain to merge because of the divergent commit history issues. Any idea as to the best way to do it?

rmcgibbo · 2014-03-29T19:40:27Z

Get the patchfile and then reapply it on a new branch checked out from
master?

On Sat, Mar 29, 2014 at 12:38 PM, Christian Schwantes <
[email protected]> wrote:

This is going to be a pain to merge because of the divergent commit
history issues. Any idea as to the best way to do it?

Reply to this email directly or view it on GitHubhttps://github.com//pull/338#issuecomment-39006452
.

merge in all of the new msmbuilder changes with merge issues being dealt with by picking SimTk's versions

…ese files won't have their history tracked in the versions in MSMBuilder/

schwancr · 2014-03-29T20:11:25Z

Alright in case anyone was curious, you can use the -X flag with a value of theirs when doing git pull or git merge to just accept all of the remote's changes.

One issue was I needed to delete the src/ directory, which might mean that the library files in MSMBuilder/ are not the same as their previous versions (so their history may be gone forever!)

rmcgibbo · 2014-03-29T20:16:48Z

I don't think anything is gone forever.

On Sat, Mar 29, 2014 at 1:11 PM, Christian Schwantes <
[email protected]> wrote:

Alright in case anyone was curious, you can use the -X flag with a value
of theirs when doing git pull or git merge to just accept all of the
remote's changes.

One issue was I needed to delete the src/ directory, which might mean
that the library files in MSMBuilder/ are not the same as their previous
versions (so their history may be gone forever!)

Reply to this email directly or view it on GitHubhttps://github.com//pull/338#issuecomment-39007411
.

schwancr · 2014-03-29T20:18:02Z

Fair enough. It's a bit annoying that github thinks that I made 1250 commits because of merging the new msmb

I guess it's a good way to pad my stats, but it just adds a lot of bloat to our commit history

schwancr · 2014-03-29T20:19:13Z

Did I do something wrong? I don't really understand why this happened

…ade the entry_point change

schwancr added 30 commits November 19, 2013 14:22

updated assign hierarchical test. I changed WardAssignments.h5 refere…

8312536

…nce as well as the test function

changed AssignHierarchical.py to actually do what we want. This requi…

0c42bcc

…res passing the ProjectInfo.yaml file as well as the stride used in clustering since the shape of the assignments file depends on the trajectory lengths in the project info

Merge pull request msmbuilder#256 from schwancr/assign_hier

f89f208

AssignHierarchical.py Fix

got project tests working on my mac

e227f7a

Added bugfix for a project builder issue when pdb filename does not e…

8d33c9f

…xist. The builder actually assumed that the error was a corrupt xtc file, and so it was stuck in an infinite loop trying to throw away the last frame in the xtc.

first attempts at getting the project builder to update the file. Thi…

7e9af3f

…s is harder than I thought it would be ...

small fix in ProjectBuilder docstring

13216ba

added logger prints as well as added the error checking that was orig…

a941c76

…inally in the builder.

bugfix in builder

d596323

bugfix in updating project that was giving new trajectories the wrong…

1ed6934

… index to save to. Also some other minor changes

fixed typo in comment

b37854a

first attempts at getting the project builder to update the file. Thi…

78fdcd5

…s is harder than I thought it would be ...

small fix in ProjectBuilder docstring

49d2a82

added logger prints as well as added the error checking that was orig…

81536da

…inally in the builder.

bugfix in builder

a7eb3c6

bugfix in updating project that was giving new trajectories the wrong…

21b7b27

… index to save to. Also some other minor changes

fixed typo in comment

205b6ea

cleaned up a little bit.

da1ace1

bugfixes and got things to work.

c965574

removed import IPython from builder.py

2b4ccad

changed PDBFilename to Conf in _load_traj, which means we don't have …

c22e291

…to load the pdb from disk for every trajectory

bugfix, I think it's all working now, finally.

ffc74f4

added a comment describing what the list comprehension thing is doing

d574539

small fixes

cd35df4

started messing with Trajectory.py in hopes of fixing it so that we c…

37b1be9

…an discard redundant frames in a reasonable manner. The trouble is, the code that is in there now DOESN'T WORK...

added fixes to Trajectory.py as well as project/builder to fix issues…

8b214f8

… with redundant frames.

got Trajectory.load_from_xtc working

ee4ee9c

got project tests working on my mac

2bea664

Merge branch 'proj_builder' of github.com:schwancr/msmbuilder into pr…

2145f00

…oj_builder

rmcgibbo reviewed Feb 7, 2014
View reviewed changes

schwancr added 3 commits February 7, 2014 14:41

moved parser declaration outside the boilerplate

bf6787a

remove chunk_ind

0b8f13b

bugfix in CalculateProjectDistance.py that wasn't using mdtraj

0964b29

rmcgibbo reviewed Feb 7, 2014
View reviewed changes

rmcgibbo and others added 9 commits February 7, 2014 14:51

Merge pull request msmbuilder#339 from schwancr/project_distance

a8bd978

bugfix in CalculateProjectDistance.py that wasn't using mdtraj

add more description to ergodic trim

565b1e0

It's actually a lil matrix

a1fc28b

Merge pull request msmbuilder#342 from mpharrigan/doc

0b41857

Add more description to ergodic trim docstring

Data types

ec0d36e

Being explicit and correct about datatypes

677521d

Merge pull request msmbuilder#344 from mpharrigan/doc

820c846

Fixed more description to ergodic trim docstring

Simplify setup.py

014e7d0

Merge pull request msmbuilder#345 from SimTk/setup

b827cd3

Simplify setup.py

Merge branch 'master' of github.com:SimTk/msmbuilder into sasa

cc1d30a

schwancr added 3 commits March 29, 2014 16:06

Merge remote-tracking branch 'parent/master'

d4e0c4d

merge in all of the new msmbuilder changes with merge issues being dealt with by picking SimTk's versions

the merge didn't delete the src directory. This means I guess that th…

5b4416c

…ese files won't have their history tracked in the versions in MSMBuilder/

Merge branch 'master' into sasa

5cccb38

just brought in the new version of CalculateProjectSASA.py and then m…

d41f69c

…ade the entry_point change

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Reworked CalculateProjectSASA.py #338

Reworked CalculateProjectSASA.py #338

schwancr commented Feb 7, 2014

schwancr commented Feb 7, 2014

rmcgibbo Feb 7, 2014

schwancr Feb 7, 2014

rmcgibbo Feb 7, 2014

rmcgibbo Feb 7, 2014

kyleabeauchamp commented Feb 11, 2014

schwancr commented Feb 11, 2014

Otherwise looks ok

rmcgibbo commented Mar 29, 2014

schwancr commented Mar 29, 2014

rmcgibbo commented Mar 29, 2014

schwancr commented Mar 29, 2014

rmcgibbo commented Mar 29, 2014

schwancr commented Mar 29, 2014

schwancr commented Mar 29, 2014


		if __name__ == '__main__':
		parser = arglib.ArgumentParser(description="""Calculates the Solvent Accessible Surface Area

Reworked CalculateProjectSASA.py #338

Are you sure you want to change the base?

Reworked CalculateProjectSASA.py #338

Conversation

schwancr commented Feb 7, 2014

schwancr commented Feb 7, 2014

rmcgibbo Feb 7, 2014

Choose a reason for hiding this comment

schwancr Feb 7, 2014

Choose a reason for hiding this comment

rmcgibbo Feb 7, 2014

Choose a reason for hiding this comment

rmcgibbo Feb 7, 2014

Choose a reason for hiding this comment

kyleabeauchamp commented Feb 11, 2014

schwancr commented Feb 11, 2014

Otherwise looks ok

rmcgibbo commented Mar 29, 2014

schwancr commented Mar 29, 2014

rmcgibbo commented Mar 29, 2014

schwancr commented Mar 29, 2014

rmcgibbo commented Mar 29, 2014

schwancr commented Mar 29, 2014

schwancr commented Mar 29, 2014