[Work Log] cleanup

August 21, 2013
Project Tulips
Subproject Data Association v3
Working path projects/​tulips/​trunk/​src/​matlab/​data_association_3
Unless otherwise noted, all filesystem paths are relative to the "Working path" named above.

Note new working path: data_association_3

Started realizing that there are a huge number of dead/obsolete files in the data_association_2 directory. It's time to migrate to a clean code-base, data_association_3.



Realized that tmp_setup_workspace.m is very valuable, but will probably need to change and evolve over time. Created new directory called setups/, where setup scripts will be stored, organized by date. Related files that the setup scrip needs will be stored in the same directory, with a similar name to the script itself.

tmp_setup_workspace.m is now in setups/setup_workspace_2013_08_21.m.

Format is: setup_<type>_<date>.m. Currently only type is "workspace". Related "load" files have format: setup_<type>_<date>.mat or if multiple files, setup_<type>_<date>.<N>.mat, where N is an increasing integer starting at 1.

Mex files

Propose creating a new file compile_mex_scripts.m. If called with no arguments, it will compile all scripts that are uncompiled. Optional "force_recompile" parameter.

Misc Notes/Issues


split_correspondence.m is no longer available. It became obsolete some time ago, around when we moved from clean_correspondence.m to corr_to_likelihood.m.

When we return to end-to-end sampling, we'll need something like it. Alternatively, maybe full-rebuilding of the split curves is fast enough, now that we've mexed the bottlenecks?


Not totally done, but we can at least run the workspace setup file.

Will save the current workspace, re-open matlab, and see what issues arise as we continue to work toward the short-term goal of implementing branching-curve marginal likelihood.

Workspace saved to tmp/workspace_2013_08_21.mat.

Posted by Kyle Simek
blog comments powered by Disqus