Permalink
Switch branches/tags
Commits on Sep 12, 2014
  1. WIP: display

    tomka committed Sep 3, 2014
Commits on Aug 8, 2014
  1. Intersection tables: add basic test cases for PL/pgSQL functions

    tomka committed Jul 7, 2014
    There is one test case to test if the new table is actually created in
    the expected format and the one tests tests the intersections. The tests
    can be run with this:
    
    ./manage.py test catmaid.tests.test_plpgsql
  2. WIP: test case

    tomka committed Jul 7, 2014
  3. Add basic PL/pgSQL function to create intersection table

    tomka committed Jun 16, 2014
    This script currently only creates intersections for treenodes that are on
    individual slices of a XY aligned stack.
  4. Add post_migrate signal handler to load custom SQL

    tomka committed Jun 30, 2014
    This handler can be used to load SQL stored in files ending with .sql stored
    inside the sql directory of the catmaid application folder.
  5. Move tests into own subdirectory

    tomka committed Jul 7, 2014
    Also the name of the test module changes to let it be found by Django's test
    discovery (all files matching test*.py).
  6. Merge branch 'front-end-tests'

    tomka committed Aug 8, 2014
  7. 3D viewer: add fallback file name pattern for mesh loading

    tomka committed Aug 8, 2014
    Instead of only looking for the file "<project_id>_<stack_id'>.hdf", the mesh
    loading function of the back-end now also looks for "<project_id>.hdf" if the
    former one isn't found. This makes it easier to support many stacks in one
    project with the same mesh.
Commits on Aug 7, 2014
  1. Remove a tag's class instance if it isn't used anymore

    tomka committed Aug 7, 2014
    After removing a tag from a treenode, it is now tested whether the tag is still
    linked to another treenode. If it isn't, it is removed (i.e. it's class instance
    is removed). This fixes #422.
  2. Fix permission check on treenode tag removal

    tomka committed Aug 7, 2014
    This commit replaces the code responsible for deleting tags of treenodes
    (like 'ends'). Before it was only tested whether a user is a superuser
    or is owner of this tag. It didn't honor group membership. This is
    is fixed by using the existing permission checking function
    'can_edit_or_fail'. It closes #725.
  3. Add migration to remove duplicate annotations

    tomka committed Aug 7, 2014
    To do all duplicate annotations are retrieved and looked at. If one duplicate
    annotation was created by the same person who created the annotated neuron, this
    annotation is kept, the other(s) are marked for removal. If such a user doesn't
    exist, the oldest annotation link is used.
    
    The administrator is presented a list of all annotations that will be removed
    along with the neuron ID and the user who created that annotation link. One can
    then confirm the removal, ignore these duplicates or cancel the migration.
    
    Issue #724 discusses this issue.
Commits on Aug 4, 2014
  1. Show Nginx example configuration from create_configuration script

    tomka committed Aug 4, 2014
    In addition to the existing Apache configuration, the script now also
    displays a simple example configuration for Nginx. This closes #666.
Commits on Aug 2, 2014
Commits on Jul 31, 2014
  1. Ontology clustering: make 'Average (UPGMA)' first linkage option

    tomka committed Jul 31, 2014
    The UPGMA linkage method seems to be of more use to the folks using the
    ontology clustering. Therefore, this commit makes it implicitly the
    default.
Commits on Jul 30, 2014
  1. Classification export: use feature short name for feature list

    tomka committed Jul 30, 2014
    Using the short name causes the classification name not to show up when
    exporting the whole classification space. It is part of the exported
    object anyway.
  2. Improve performance of some aggregation queries

    tomka committed Jul 30, 2014
    Replacing count(column) with count(*) in SQL queries can lead to a
    performance improvement (e.g. queries like retrieving the number of
    treenodes per skeleton).  This, however, only works correctly if the
    column of interest has a NOT NULL constraint. Without it, count(column)
    and count(*) would have different semantics. The former wouldn't include
    NULL in the given column. The latter in turn would.
  3. Ontologies: test parent class only for first element of path

    tomka committed Jul 29, 2014
    Testing the parent class on all elements of the feature doesn't work.
    The test is pretty naive and assumes too much about the structure of the
    graph.
Commits on Jul 29, 2014
  1. Ontologies: make root node of feature link part of its name

    tomka committed Jul 29, 2014
    This allows clear distinction between displayed feature links if an ontology
    sub-tree is reused underneath different classification roots.
  2. Ontologies: fix text if a classification graph realizes ontology feat…

    tomka committed Jul 29, 2014
    …ure path
    
    Neither the simple nor the complex implementation of this test checked whether
    the class (i.e. the type) of the parent graph component is the correct. This
    is no problem until the first child nodes of different ontologies are different.
    But if a sub-tree is reused in different ontologies, this breaks down. This
    commit fixes this problem.
Commits on Jul 28, 2014
  1. CircuitGraphPlot: added measurement of "Cable of main dendritic shaft"

    acardona committed Jul 28, 2014
    Computed using the cable length of all nodes with 0.9 * max centripetal flow centrality.
  2. CircuitGraphPlot: anatomy prunes away artifactual branches whose end …

    acardona committed Jul 28, 2014
    …nodes are tagged with "not a branch".