Contributors: Chris Barnes, Albert Cardona, Andrew Champion, Tom Kazimiers
Notes
- A new NODE_PROVIDER option can be added to settings.py, which can be used to
configure the way node queries are executed. Options are 'classic',
'postgis3d' and 'postgis2d'. Depending on your environment, 'postgis2d' might
be fastest, but 'postgis3d' is the current default (i.e. what has been used so
far).
- The new PREPARED_STATEMENTS option can be added to settings.py and set to
True to improve node query performance when connection pooling is used.
Without connection pooling, this setting hurts performance, which is why it is
set to False by default.
- Stack mirrors (see below) are not displayed by default in CATMAID's standard
data views. To show them, '"show_mirrors": true' (including double quotes) has
to be added to the respective data view configuration in CATMAID's admin
interface.
Features and enhancements
Stack mirrors:
- To reduce loading time of image data from different places around the world,
it is common to copy the image data set and make it available from a server
closer to its users. These "stack mirrors" are now handled differently by
CATMAID. Until now it was common to select the closest mirror manually by
clicking on an image on CATMAID's project list or by clicking on the
respective link. To select a particular "stack mirror" one now only links can
be used, individual images are not shown anymore. If no images were used
before and only links were shown on CATMAID's home page, only small visual
change was mode: an extra stack link on top of all mirrors is added. As an
alternative to selecting individual mirrors, both, images and stack links can
now be used to open a stack as well and CATMAID selects the fastest mirror
automatically.
- When using links to a particular view in CATMAID ("URL to this view"), CATMAID
will automatically select the fastest available stack mirror.
- Once a stack is displayed the stack mirror in use can be changed: After
clicking on the white-on-blue square in the lower left corner of the image
data display, the image display settings will show up. In there the active
mirror can be selected from a drop-down list.
Treenode Viewer:
- A new tool similar to the connector viewer, which allows users to
quickly view the nodes in a treenode table.
- The viewer is opened from the treenode table.
- Nodes are shown only if they are filtered in the table, and appear in
the order that they appear in the table.
- When the sorting and filtering in the table is changed, users can
refresh the viewer from the table to reflect these changes.
Connector Viewer enhancements:
- Connectors can now be constrained by their pre-synaptic and post-synaptic
skeletons.
- If either is unpopulated, it is unconstrained on that side (i.e. all
outgoing connectors can be found by clearing the 'Post- skeletons' set).
- Sorting algorithms based on treenodes associated with the connector
(e.g. depth) can be applied either to presynaptic or postsynaptic
skeletons.
- The 'Reverse' button switches the contents of the two skeleton sets.
- The 'Sync' checkbox sets the pre- and post- synaptic sets to have the
same contents (the union of their contents when the box is checked)
and keeps them that way: this is useful for inspecting all connectors
acting within a set of skeletons, such as reciprocal connections.
- For undirected connectors (gap junctions, abutting etc.), you can use
either the 'pre' or the 'post' set. If both sets are populated, only
connectors touching a pre- skeleton and a post- skeleton (with 2
distinct edges) will be shown.
- When using the connector viewer as a skeleton source, the output is the
union of the two skeleton sets.
- The controls have been separated out into tabs.
- N.B.: Because the Connector Viewer is now populated in a different way
to the Connector Table, you can now only open a Table from a Viewer if
one of the skeleton sets is empty, and a connector type other than 'All'
is selected.
Miscellaneous:
- Bookmarks are now persistent for each project.
- The Graph Widget now offers two options when exporting SVGs: regular ones and
Adobe Illustrator compatible ones (because Adobe isn't SVG standard conform).
An option dialog is shown after clicking "Export SVG".
- From the Graph Widget exported SVG files preserve now the view of the widget
(zoom and pan).
- The tag table can now be constrained by a set of skeletons.
- The tag table now gives the user information on how many skeletons are
being used as constraints, and how many tags/skeletons/nodes are
selected.
- When the tag table is refreshed, the filters and the last sort are
persisted.
- The treenode table can now be filtered by node confidence, creator and
reviewer.
- When creating new volumes with the help of the Volume Manager, the color and
opacity of the 3D preview can now be adjusted.
Bug fixes
- 3D Viewer: custom tag highlighting now also works for tags with upper case
letters.
- Layer settings: checkboxes now have correct default value.
- Tag table: any tags with identical names are now treated as identical.
- Tag table: fixed off-by-one error in node count.