Skip to content

v1.0.1

Choose a tag to compare

View all tags
@gustavo-meilus gustavo-meilus released this 19 May 18:03
· 2 commits to main since this release
v1.0.1
340bc5a
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Removed

  • Vector subsystem (src/vector/, vector_search, vector_db_status, logseq-sync CLI, [vector] dep group) — retraction; the v1.0.0 implementation shipped without an embedder and was non-functional. See MIGRATION.md.
  • get_page_links tool alias — use get_page_backlinks.
  • pathlib>=1.0.1 and load-dotenv runtime dependencies.

Added

  • LogseqClient.excluded_page_names(ttl_seconds) cached privacy helper.
  • LogseqNotFoundError raised on HTTP 404; LogseqConnectionError reserved for network failures only.
  • pyyaml runtime dependency for proper Markdown frontmatter parsing.

Changed

  • register_all_tools(mcp, client, config) — shared LogseqClient lifetime across the process.
  • LogseqConfig.verify_ssl default flipped to True.
  • LogseqClient.create_page now passes properties as the second positional arg per the official IEditorProxy signature.
  • LogseqClient.append_block_in_page forwards optional options arg.
  • LogseqClient.resolve_property_ident query now binds ?ident (was incorrectly ?e) and escapes input.
  • LogseqClient.resolve_page_uuids batches via a single datascript query.
  • extract_tags is case-insensitive and supports the DB-mode [{name: "x"}] shape.
  • update_page replace mode uses insert_batch_block to preserve block hierarchy.

Fixed

  • src/tools/__init__.py no longer writes to stdout (fixes MCP stdio JSON-RPC corruption on tool-import failure).
  • find_pages_by_property accepts namespaced property names; escapes \ and "; caps values at 256 chars.
  • DB-mode tools resolve UUID → name before calling getPageBlocksTree (workaround for logseq/logseq#4920).
Assets 2
Loading