2020.02.15

2020.02.15

Contributors: Chris Barnes, Andrew Champion, Stephan Gerhard, Pat Gunn, Tom Kazimiers

Notes

• Python 3.5 is not supported anymore. Use Python 3.6, 3.7 or 3.8.

• Postgres 11 and PostGIS 2.5 is required, Postgres 12 and PostGIS 3 is
  recommended. If Postgres needs to be updated, update directly to Postgres 12.
  If both needs to be updated, update PostGIS first and run
  ALTER EXTENSION postgis UPDATE; in every database. For docker-compose
  setups this database update is performed automatically. If a replication setup is in
  use, the database configuration changes for Postgres 12. CATMAID's replication
  documentation explains what needs to be done.

• The next version of CATMAID will require Postgres 12 and PostGIS 3.

• A virtualenv update is required. Before you start it, please remove some
  packages that are not needed anymore first:

  pip uninstall asgi-ipc asgi-rabbitmq

• If ASGI was set up before, make sure to install channels_rabbitmq or
  channels_redis (depending on what yous use). The older asgi_rabbitmq and
  asgi_redis packages aren't supported anymore. This also requires an update of
  the CHANNELS_LAYERS in settings.py. The channels_rabbitmq documentation for an
  example: https://github.com/CJWorkbench/channels_rabbitmq/. This variable
  isn't defined by default anymore. Therefore you likely have to replace any
  CHANNELS_LAYERS[…] = … with something like CHANNELS_LAYERS = { … }. The
  new format is (use custom credentials on any production system!):

  CHANNEL_LAYERS = {
    "default": {
        "BACKEND": "channels_rabbitmq.core.RabbitmqChannelLayer",
        "CONFIG": {
            "host": "amqp://guest:[email protected]/asgi",
        },
    },
  }
  

  Also, if supervisord is in use to manage CATMAID process groups, the main
  Daphne process needs an adjustment: instead of calling daphne with the
  mysite.asgi:channel_layer parameter, use mysite.asgi:application. A
  complete supervisord entry would then look something like this:

  [program:catmaid-daphne]
  directory = /home/catmaid/catmaid/django/projects/
  command = /home/catmaid/catmaid/django/env/bin/daphne --unix-socket=/var/run/daphne/catmaid.sock mysite.asgi:application
  user = www-data
  stdout_logfile = /var/log/daphne/catmaid-server.log
  redirect_stderr = true
  

  As last step, the supervisor entry for the daphne worker process has to be
  removed. New types of workers can be added, but are not needed in most cases.
  The channels documentation has more information on this.

• GDAL v2 or newer is now needed. If your Ubuntu version doesn't support this
  yet, there is an official PPA:

  sudo add-apt-repository ppa:ubuntugis/ppa`
  sudo apt-get update`
  sudo apt-get install gdal-bin
  

• The management command gdal-bincatmaid_populate_summary_tables is now known as
  catmaid_refresh_node_statistics.

• When enabling and disabling both history tracking and spatial update events on
  startup, advisory locks are now used. This should make the startup of parallel
  CATMAID workers more robust.

• The application of migrations 88-91 and 98-99 might take a while to complete,
  because they rewrite potentially big database table (treenode_edge, treenode,
  class_instance, and more). Therefore, make also sure that there is enough
  space available at the database storage location (25% of database data
  directory should be plenty). If no replication is used, setting the following
  Postgres options can speed up the process: wal_level = minimal,
  archive_mode = off and max_wal_senders = 0.

  Due to this database update data consistency and correctness was improved,
  because additional foreign key relationships have been added that were missing
  before.

• Both configuration.py.example and create_configuration.py support now the
  option catmaid_default_enabled_tools, which defines the list of front-end
  tools that are enabled by default for new users. The following options are
  supported: cropping, tagging, textlabel, tracing, ontology and roi. By default
  the tracing tool is now enabled for new users, because it seems in many setups
  users want to have this enabled by default.

• Back-end errors result now in actual HTTP error status codes. Third-party
  clients need possibly some adjustments to handle API errors. In case of an
  error, status 400 is returned if an input data or parameter problem, 401 for
  permission problems and 500 otherwise.

• On startup, CATMAID will now test if the output directory is accessible as
  well as if its expected folder layout is present. If expected subfolders are
  missing, they will now be created.

• Python code is now linted by flake8. The config is fairly relaxed
  (see .travis.flake8), but contributions which do not conform with the selected
  rules will fail CI checks.

Features and enhancements

Permissions:

• If data is imported by a user, e.g. from a remote CATMAID instance, this user
  has all permissions over the imported data. If data was imported or created by
  someone else, the regular rules apply: permissions are granted if the user is
  creator, superuser or has permission over the data's creator. If user A has
  permissions over user B's data, user A will also have permission to edit user
  B's imported skeletons. This is done in order to be able to maintain the
  original creator and editor data of the imported data, should it be available.

Node filters:

• Binary split: a new toggle labeled "inclusive" allows to decide whether to
  include the split not in an upstream sub-arbor (default: true).

CLI importer:

• Performance improvement: only the skeleton summary entries of the imported
  skeletons are now created.

• Only update treenode edges and skeleton summary if treenodes were actually
  imported (rather than e.g. only volumes).

• Add --auto-name-unknown-users option, to automatically generate names for
  users that are not available from an import, but referenced in it (by ID).

CLI exporter:

• The new --annotation-annotation option allows to specify a meta-annotation
  that all exported annotations need to share transitively. If for instance all
  all top-level annotations that should be exported are annotated with
  "exportable" and "--annotation-annotation exportable" is used, only
  annotations from those annotation hierarchies marked as "exportable" will be
  exported. Without this option, all annotations are exported.

• If users should be exported, also export users of volumes.

Project statistics:

• Requests to the back-end our now performed in parallel, resulting in faster
  loading times.

• It is now possible to configure how many largest neurons should be displayed
  as well as an optional name pattern.

• The user/time action table will now also list imported nodes and imported
  cable length. These values are subtracted from the corresponding regular
  counters. If "Include imports" is unchecked, the same information that was
  displayed before this commit is shown, i.e. the regular node count and cable
  length field include the imported data.

Tracing tool:

• The tool bar button to refresh caches has been replaced with a button to open
  a context menu titled "More tools". The context menu has an entry to refresh
  cashes and a list of tracing data in other projects.

• Tracing data from other CATMAID instances (or other local projects) can now be
  added to the active stack viewer by selecting the respective remote project
  from the "More tools" menu item "Remote data".

• To open a remote CATMAID instance at the same location as the current view,
  find the remote project in the "Remote data" section of the "More tools" menu,
  and click on "Open remote view".

• The marker type used fore remote tracing data can be configured through the
  Settings Widget in the Tracing Overlay section, option "Remote node marker". A
  ring is used by default.

• Merges will now by default respect the special annotation "stable". If a
  neuron is marked as such it will always win a merge and if both merge partners
  are marked "stable", the merge is canceled. This can be disabled by setting
  the project or instance setting "Stable join annotation" in the Tracing
  section of the Settings Widget to an empty string. The merge dialog will show
  a warning if stable merge annotations have any bearing on the merge direction.

• Fast merge mode can now optionally be applied only to neurons/fragments with
  a particular name pattern (regular expression). This can be configured in the
  Fast Merge section of the Tracing section in the Settings Widget.

• Alt + G will now select the closest node in the active skeleton. If no
  skeleton is selected, the globally closest node is select. Regular G works as
  before in the current section.

• Skeleton splitting: if a skeleton sampler was in use, a split could so far
  only be successfully executed if the split node is within a sampler domain.
  This is fixed now and sampled skeletons can now be split upstream or
  downstream of a domain as well. If split upstream, the domain will be deleted,
  because it is part of the split-off fragment.

• The default interaction mode (skeleton tracing, synapse dropping, select,
  move, import) can now be configured in the Tracing section of the Settings
  Widget. This interaction mode will be enabled when opening the Tracing tool.

Tracing overlay:

• The remote mirror CATMAID instance configuration is now taken from the central
  configuration, which can be edited in the "Other CATMAID instances" section in
  the Settings Widget. The "Read-only mirror index" settings is kept but refers
  to list elements in the central configuration list.

• Joining two skeletons without confirmation ("fast mode"), will now attach an
  annotation to the resulting skeleton that references the merged in skeleton by
  name. The name of this annotation as well as the suggested reference visible
  in the confirmation dialog have the form "Merged: ".

• It is now possible order loaded tracing data by skeleton size, both in
  ascending and descending order. This is useful with other node count limiting
  filters. The layer settings accessible through the blue/white box in the lower
  left corner of a stack viewer provides the "Data ordering" select box to
  change this setting. By default no ordering is applied.

• When merging a fragment that contains a sampler, it can now be merged if the
  respective skeleton doesn't need to be rerooted.

• Visibility groups can now use a generic node creator test that references
  the current user (logged in our anonymous). It is called "(self)" in the list
  of creators for all three visibility groups. This is useful to define
  visibility groups on the project or instance level that include a hiding
  pattern while guaranteeing that users see their own nodes.

• Shift + Click on a remote node from a local node (or vice versa) will now
  bring up an import preview dialog which is used as a merge preview at the same
  time. Confirming the change will import the remote skeleton and merge it at
  the selected location, works with virtual nodes.

Navigator:

• Two new buttons and a text input field on the right side of the toolbar allow
  easier handling of locations: the first icon (the map marker) is a button to
  copy the current location of the center of the view in project (physical)
  coordinates. If clicked while the Alt key is pressed, stack (pixel)
  coordinates will be copied. Both modes will copy the coordinates in the form
  "X, Y, Z" to the clipboard.

• The text input box allows to enter coordinates and other information. Pressing
  enter or clicking the angle brackets on its right side will make CATMAID try
  to interpret the text and act accordingly. The following interpretations are
  attempted: 1. Project coordinates of the form "X, Y, Z" and "X, Y".
  Parentheses and brackets are removed, 2. Skeleton ID, 3. Location ID
  (treenode, connector, etc.), 4. Neuron ID, 5. Bookmark letter from the
  bookmark system accessible through ; and 6. Project coordinates from
  CATMAID URL. These are tested in order and whatever matches first, wins.

• Additionally, particular interpretations can be enforced by using any of the
  following prefixes "stack:", "px:", "neuron:", "skeleton:", "node:",
  "connector:", "url:" and "bookmark:". It is possible to use only the beginning
  of a prefix, i.e. "stack: 1,2,3" is the same as "st 1,2,3". This will only try
  the referenced parser. It also makes stack-coordinate parsing available, which
  is otherwise not checked for.

• Annotation lists show now also annotation IDs.

Administration:

• The catmaid_update_cache_tables management command accepts now a --jobs=n
  parameter, which allows it to run in parallel when set to values n larger one.
  By default a single process is used. This can speed up cache generation
  significantly. With the help of the option --chunk-size=n it is possible to
  define how many cells should be processed per CPU task. For large dataset it
  might be useful to raise the default of 10.

• The catmaid_update_cache_tables management command accepts now a
  --depth-step=n parameter, which makes the update process reevaluate the
  number of grid cells to look at n times during a single run. For instance,
  if set to 2, the upper bound of cells to look at is reevaluated when the
  process is through with half the depth dimension. In larger datasets, this can
  speed up cache generation significantly.

• The catmaid_update_cache_tables management command accepts now an optional
  --order parameter, which can be set to either "cable-asc" or "cable-desc".
  By default no order is applied.

• The group list will now also show a sortable column containing the number of
  group members. It also provides a more useful filtering option on the right.
  Clicking "Only non-user groups" will now only show groups with a name that
  isn't a username. Additionally, it is possible to filter by a minimum and
  maximum number of group members.

• Settings: add CLIENT_SETTINGS and FORCE_CLIENT_SETTINGS options so that
  an initial instance-wide client configuration can be specified.

• Setup: the create_configuration.py script will now create backups of existing
  settings.py and django.wsgi files.

Volumes:

• Volumes have now their surface area and volume associated with them as well as
  whether they are watertight. This information is visible in the Volume Manager
  widget and can be recomputed on demand if needed, using the respective button
  in the action column.

• The base unit used for the ara and volume columns can be adjusted using the
  "Reference unit" select box in the widget controls.

Neuron Search:

• Name queries can now be marked as exact, which prevents matching against parts
  of other names. This is ignored for regular expressions.

• Name queries can now be marked as case sensitive using the respective checkbox
  next to the name field. By default name queries are case insensitive.

• The search can now be further constrained to only skeletons that include at
  least partially imported fragments or to skeletons that are fully imported.

Graph widget:

• Node filters are now supported. Like in many widgets, they available through
  the funnel icon in the title bar. Only nodes that are allowed by the selected
  filters, can contribute connections and the graph. At the moment, this comes
  with a few caveats: if filters are in use, no subgraphs (e.g. axon/dendrite,
  synapse clusters) can be used. Grouping isn't available either and in
  addition, skeletons are removed from the widget if they don't comply with the
  filters. All this is likely going to chagne in future versions.

• The "Nodes" tab has now three new buttons to hide, show and remove unconnected
  nodes. Show will nodes visible that have been hidden before. Removing nodes
  will remove the skeletons/nodes completely from the widget.

• The node size specified in an imported .graphml file is now respected. This is
  useful if graph fine tuning (layout and node size) is done in other tools,
  like Gephi.

Review widget:

• The new tab "Synapse completeness" can list synaptic connector nodes don't
  have a presynaptic node associated with them as well the ones without a
  postsynaptic node.

• The Skeleton analytics tab now also lists low confidence nodes/edges with a
  confidence lower than 4. The label of each entry shows the respective
  confidence value.

Import/Export widget:

• It is now possible to import neurons from other CATMAID instances. The "Import
  from CATMAID" tab provides tools to search remote skeletons by name or
  annotation. Imported neurons can automatically be annotated with a set of
  annotations. If no annotation are provided, CATMAID will add the annotation
  "Import" to imported skeletons.

• A custom set of default annotations that will be added to imported skeletons
  can be defined through the "Default skeleton import annotations" option in the
  Settings Widget. At the moment, this is set to "Import" as well as "{source}
  upload {group}". The second annotation uses two optional placeholders, which
  are replaced by the skeleton import source reference and the user's primary
  group, respectively. Should no primary group be set, the username name is
  used.

• The original ID and source URL of all imported skeletons is tracked. This
  allows the widget to display local skeleton IDs, if the remote ID has already
  been imported from the current source URL.

• If the active skeleton is a remote skeleton (in a remote data tracing layer),
  it can now be imported by clicking the "Import active skeleton" button in the
  "Import from CATMAID" tab of the widget. Like the other import, this iwll
  first ask for confirmation with a 3D dialog.

• The NRRD export can now use the JRC2018U (unisex) target template brain space
  as target.

• The NRRD export can now export more than one skeleton at a time. All skeleton
  sources are now displayed.

• The NRRD export will now by default create a Zip file if multiple skeletons
  are exported both in a synchronous and asynchronous export.

• SWC imports can now have a user defined initial name.

Neuron navigator:

• The new button "Imported nodes" will list all imported treenodes for the
  displayed skeleton.

• The detail view of an annotation will now list the time a neuron was annotated
  next to the listed neuron name. The last time this link was edited is shown as
  well. Both columns can be used for sorting as well.

3D viewer:

• The new shading mode named "Imported fragments" will make only imported edges
  visible. The new coloring mode "Imported (green)" will highlight all imported
  fragments in green.

• Stored views (View tab > Save view) are now shared between all CATMAID
  instances, not only between all 3D Viewers like it is done now. This means
  they are also persisted on the back-end and can be retrieved at a later point.

• 3D viewer: the line width has now an upper limit of 25px to prevent crashes of
  the 3D viewer with larger values and many skeletons.

• Desmosome links can now be displayed. The color assigned in the Settings
  Widget section for desmosomes in the Tracing Overlay is used here as well.

Neuroglancer widget:

• CATMAID now includes a copy of Neuroglancer that can be accessed at
  <catmaid_base_url>/neuroglancer

• The Neurglancer widget allows to open an instance of Neuroglancer in a new
  window and its navigation (center location, zoom level) is synchronized with
  the navigation happening in CATMAID. An URL to a Neuroglancer instance containing
  the same dataset as shown in the CATMAID stack viewer needs to be provided.

• The widget's info page has a step-by-step guide on how to use it for the
  FAFB dataset

Treenode table:

• The UI elements to select active filter options like node type, confidence or
  confidence operator now have a colored background if an active filter is
  selected. This is useful to draw more attention to the fact that leaf nodes
  are selected by default.

• Remote skeletons are now supported. They can be mixed with local skeletons.

• A new column "i" represents whether or not a particular treenode was imported.
  This information is loaded separately, parallel to the regular information.
  Which, is why initially "…" is shown. If a node wasn't imported, the field is
  empty. If it was imported, "Y" is shown.

Morphology plot:

• Two new center computation modes have been added: Tagged node closest to root,
  Tagged node most distant from root. Both use the tag input field to the right
  of the center mode selector. Either pressing enter or Redraw after a tag
  change will recompute the plot. If no node is found with the respective tag,
  the root is used and a warning is displayed.

• It is now possible to select the interpolation mode of the plot. So far the
  plotted data was always interpolated with a basis spline. This isn't
  necessarily what is required for a particular plot, e.g. if the sample values
  have to be intersected. To allow this, a drop-down menu with the following
  interpolation options has been added: basis spline, linear, cardinal spline,
  monotone, step before and step after.

Spaces:

• Users with the new "can_fork" permission have a new user menu entry in the
  upper right corner: "Create own space". This allows to create a personal copy
  of a project, without any tracing data.

Miscellaneous:

• The active project title is now shown in the window title.

• There is a new menu item named "User info" in the menu displayed when hovering
  the mouse cursor over the user's name in the upper right corner. When clicked,
  it will show the basic information on a user: name, username, user id, primary
  group and color.

• Projects can now be kept open if all last stack-viewers are closed. To not
  automatically close the open project with the last stack-viewer, uncheck the
  "Project closes with last stack viewer" option in the Settings Widget.

• Neuron similarity widget: besides a regular arithmetic mean, the geometric
  mean is now an additional option for how to combine (normalized) forward and
  reverse scores.

• Neuron similarity detail widget: make CSV export of object IDs and names
  configurable.

• Retrieving review information is now faster for bigger skeletons.

• Name search: the speed of search queries is now improved.

• Selection table: the summary info button now shows also the number of imported
  nodes within the selected skeletons.

• Neuron search: search fields will now expand if their content is bigger than
  the default size. Also copied fields don't contain the content of the source
  field anymore.

• A CATMAID deep-link URL can now include the layout that should be displayed
  using the layout parameter. It takes a layout specification as it is used in
  the Settings Widget and the Layout menu.

• The speed of bulk operations on skeletons has changed (e.g. splitting, joins,
  reroot).

• Settings widget: it is now possible to clear both front-end and back-end
  settings through the "Clear settings" button. It will ask for confirmation and
  whether both front-end and back-end settings should be cleared.

• The top toolbar of the web UI won't add line breaks anymore if only little
  horizontal space is available. The part of the toolbar that contains the
  widget buttons for the selected tool will shrink instead. Only elements that
  fit the available width are displayed there now. Invisible items are not a
  problem, because all widgets can be reached through the "Open Widget" dialog
  or Ctrl + Space.

• Potential WebGL performance problems are now detected when stacks and stack
  groups are loaded. If problems are expected by the browser, a warning is
  shown.

• Spatial update worker: cache cells are not queued for updates anymore if the
  source project of a spatial event doesn't have any caches.

• Data views: add option "show_empty_projects" to simple project list data view.
  It is false by default and can be turned on to show empty projects, including
  the classification dummy project.

• Project statistics: clicking the "User analytics" button with user selected
  will open the respective user analytics views for all of them (max. the first
  three).

Bug fixes

• Connector list: dates are now display correctly.

• Tracing tool: don't allow moving nodes in select-only mode.

• Tracing layer: merged-in single nodes will now also have their name remembered
  as a new annotation in a join if their name doesn't follow the default pattern
  "neuron ".

• Project importer: the UNIX "glob" filter pattern that can be used to filter
  imported projects is now also applied to imports from URLs and files.

• Nblast widget: store tangent neighbor count specified in UI for scoring matrix
  imported from CSV.

• Celery can now be restarted again without restarting all of CATMAID.

• Connectivity widget: fix error showing when doing many subsequent Refresh
  button clicks.

• 3D viewer: the handling of intersecting transparent objects has improved and
  works like expected in most situations.

• Synapse plot: a missing skeleton source interface method was added, making it
  now possible to also remove skeletons from the widget.

• Synapse plot: skeletons can now be added to the widget again. This was
  temporarily broken for CATMAID instances using a sub-folder in their URL.

• Management command "catmaid_find_node_provider_config" to generate
  NODE_PROVIDERS configurations for the settings.py file:The default value for
  the "--orientation" parameter is now "xy". It was accidentally "xz" before.

• Graph widget: pressing the Delete key with a node selected in an active Graph
  Widget will now ask for the removal of the node from the graph, not anymore
  whether the active node should be deleted completely.

• Graph widget: showing fractions of inbound/outbound fractions of connections
  as edge labels works again.

• Graph widget: refreshing a graph with undirected edges (e.g. sub-graphs like
  axon/dendrite split, desmosomes, etc.) doesn't lead to suddenly appearing
  arrow heads.

• Graph widget: undirected links like desmosomes will now only appear once
  between two nodes. They would be shown twice before, once for each of the
  partners being the "source" or "target".

• Graph widget: apply currently selected layout after graph update. This fixes
  occasional fall-backs to the force-directed layout.

• SWC skeleton import: if force = true, treenodes of a replaced skeleton are
  now correctly removed on a successful import.

• CLI Importer: creating unknown users manually works now as expected.

• Node filters: the application of node filters with rules applied in an
  intersected fashion is fixed now and the filtered result of all skeletons is
  shown.

• Spatial update notifications: only try to update the database configuration if
  the database isn't running as a database replica.

• Synapse plot: fix read-out of settings input fields.

• Connectivity matrix: don't raise an error in "fraction" display mode on a
  widget refresh (e.g. caused by a neuron name update).

• Connectivity matrix: don't show error when enabling "fractions" display
  without neurons being loaded.

• Connectivity widget: fix error caused by quickly refreshing the table
  successively.

• Graph widget: fix error caused by quickly refreshing the graph successively.

• Neuron navigator: fix name display after rename of neuron.

• Neuron search: the tab key focus order within one widget is now maintained if
  multiple search windows are open.

• Neuron search: sort results ascending by name by default and keep the sort
  order through table redraws.

• Settings widget: the width of multiline select elements (e.g. annotation based
  neuron naming) is now adjusted automatically to fit the content.

• 3D Viewer: fix skeleton vs. name CSV in "Synapses as CSV" export.

• 3D Viewer: fix accidental hiding of soma node sphere when hiding radius
  meshes. This happened if the node tagged 'soma' wasn't the root node. Now the
  test for whether to hide radius meshes for a node now requires that the node
  isn't tagged with 'soma', which is more robust than root node alone.

• Measurement table: fix CSV export for aggregated multi-fragment results.

API changes

Additions

• GET /{project_id}/stats/cable-length:
  Accepts now an optional 'name_pattern' argument that allows case insensitive
  pattern matching in the respective neuron name. If a '/' prefix is used, the
  pattern is interpreted as a regular expression.

• GET|POST /{project_id}/skeletons/validity:
  Obtain a list of valid or invalid skeletons in a set pf passed in skeleton
  IDs.

• GET|POST /{project_id}/skeletons/from-origin:
  Looks up skeleton IDs based on a set of source_ids, a source_url and a
  source_project_id..

• GET|POST /{project_id}/skeletons/origin:
  Gets origin information on a list of skeleton_ids.

• GET|POST /{project_id}/skeletons/import-info:
  Gets information on which and how many nodes are imported in a set of
  skeletons.

• POST /{project_id}/annotations/query-targets:
  Accepts now an optional boolean "with_timestamps" parameter, which is set to
  false by default. If set to true, the returned annotations will include their
  creation_time and edition_time.

• POST /{project_id}/annotations/query-targets:
  Accepts now two additional sorting options: annotated_on,
  last_annotation_link_edit. The first one sorts by the date the matched
  annotations were added to the target. The second option sorts by when this
  link was edited last.

• GET|POST /{project_id}/skeletons/completeness:
  Obtain compleness information on a list of skeletons.

• POST /{project_id}/skeletons/{skeleton_id}/find-labels:
  The new optional bolean only_leaves parameter (default false) can be used to
  return only tagged leaf nodes (including root).

Modifications

• All APIs: errors during a request are now indicated with a more useful HTTP
  status code than 200: errors related to the request, input data and the client
  will result in status 400, permission errors in status 403, unavailable
  resources in status 404 and internal server errors in status 500.

• POST|GET /{project_id}/node/list:
  Offers a new optional parameter
  "ordering", which can be used to order the result set of nodes. The values
  cable-asc and cable-desc are allowed. By default no ordering is applied.

• POST|GET /{project_id}/volumes/:
  Information on area, volume and watertightness is now returned for each
  volume as well as information on whether these details have been computed
  (meta_computed). This meta data can be computed for each volume using the
  {project_id}/volumes/{volume_id}/update-meta-info API.

• GET /{project_id}/volumes/{volum_id}/update-meta-info:
  Try to compute area, volume and watertightness for a volume. Sets
  meta_computed to true. If this operation fails, all meta data fields are set
  to NULL/None.

• POST /{project_id}/annotations/query-targets:
  Accepts now an optional boolean "name_exact" parameter, which is set to false
  by default. If set to true, the passed in name has to match exactly. This is
  faster than using a regular expression like '/^name$' for exact matches.

• POST /{project_id}/skeletons/import:
  The new parameters source_id and source_url allow to associate a source
  data reference with the imported skeleton. The new list parameter
  "annotations" can have annotation names, which are added to the import
  skeleton. By default only the annotation "Import" is set.

• POST /{project_id}/skeleton/join:
  The new parameter lose_sampler_handling controls how to deal with samplers
  defined for the merged in skeleton. Possible values are 'delete-sampler' and
  'keep-sampler', the latter being the default. If samplers should be kept, it
  is required that the sampler domains are defined on parts of the skeleton that
  don't need local rerooting. Otherwise an error is raised.

Deprecations

None.

Removals

None.