Changes between Version 2 and Version 3 of 2015-1-5


Ignore:
Timestamp:
Jan 5, 2015, 1:08:48 PM (11 years ago)
Author:
Scooter Morris
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • 2015-1-5

    v2 v3  
    1 Attendees
     1= Attendees =
    22
    33    TomG, Eric, Greg, TomF, Scooter, Conrad, Elaine
    44
    5 Agenda
     5= Agenda =
    66 * Tool shed update
    77 * Model class proposal and prototype implementation
    88 * Eric & Tom proposal for molecular graphical data
     9
     10= Tool shed update =
     11 * Committed and in the repository
     12 * Currently checks local installs and builds local database
     13 * Significant restructuring of Makefile
     14  * Install into app tree is done by building wheel file in build tree and moving into the app tree
     15  * What's not done is moving the wheels into a standard location
     16 * Users will use Chimera2's built-in web browser to select new tools for install
     17 * Source tree organization for tools -- all tools should have a setup.py (or setup.py.in)
     18  * Will handle the build and move
     19  * Should be a simple template
     20 * Metadata for tools come from the wheel, which include an "info" file that include all of the standard and Chimera2-specific metadata
     21
     22= Model class proposal and prototype implementation =
     23 * Now model class is a sub-class of drawing
     24 * Export function is part of the drawing class
     25 * Next step is making it also a subclass the session state API to support saving and restoring sessions
     26 * Issue: how do we handle reading data from a saved session into a current session?
     27  * Restore new session into it's own session object
     28  * Merge them (will have a lot of "fixup" to do)
     29
     30= Molecular graphical data proposal =
     31 * Approach is to use the same approach as Chimera1.  This avoids some ugly things in the C++ layer.
     32 * How do we decide which attributes are in the C++ layer?
     33  * All core graphical attributes are in C++ layer
     34 * For external developers, need to decide on the recommended approach for adding attributes for high performance
     35  * Not on the list for February
     36  * Need to think more about this
     37   * Would be nice to support in the C++ layer
     38   * Should integrate with Render by attribute
     39
     40= Other notes =
     41 * Greg will be adding a precommit hook to lint the python code before you commit
     42  * We're going to try it, but Greg will send out instructions on how to override the precommit hook
     43
     44= Next Meeting =
     45 * Greg: Model API
     46 * Eric & Tom: Initial proposal for the set of C++ attributes that's part of the molecular graphical data proposal
     47 * Conrad: Tool Shed local repository server demo