Skip to content

v0.4.0 — Category & connection details

Choose a tag to compare

@impactjo impactjo released this 13 Jun 09:32
· 37 commits to main since this release

New tool: get_category_details

Returns the full details of a category — path, parent, item counts,
purpose, note, description, and user manual. Rounds out the read side to
match get_element_details and get_connection_details.

New

Richer metadata on categories and connections:

  • Categories now store purpose, note, and user_manual in addition to
    description. All four are settable on create/update and shown in
    get_category_details.
  • Connections now store status and user_manual. Both are settable on
    create/update and shown in get_connection_details. get_connections
    gains a status filter, and searchAllFields now also searches the
    user manual.
  • list_categories shows compact flags for which of these fields are
    populated.
  • update_category now guides the assistant to check existing content
    before overwriting a text field, matching update_element and
    update_connection.

Improvements

  • find_element orders results by their position in the tree, then by name,
    so siblings stay grouped under their parent.
  • delete_status reports connection usage in addition to element usage when
    a status is still assigned.

Fixes

  • The Existing status filter in find_element and get_connections now
    also matches records that have no status set — the normal case for things
    that are simply present.
  • Path filters (under) escape % and _ so they are matched literally
    instead of as wildcards.
  • update_element and update_connection reject calls that provide no
    fields to update instead of silently bumping the change timestamp.
  • An update_status call that only changes the note no longer rewrites the
    status name with the casing used in the lookup.
  • An empty short name on update is rejected instead of being stored blank;
    use the explicit clear option to remove it.
  • get_connections no longer prints a duplicate source header when two
    sources share a name prefix.
  • Connection and category lookups fold case in the database instead of in
    the application, so case-insensitive matching behaves the same across all
    tools (relevant for non-ASCII names).

Updated

  • Server instructions describe the element hierarchy and clarify the scope
    of connections.
  • README and tool descriptions: clearer guidance on field choice and element
    naming.