Types¶

All data types returned by the Vault API are importable from the markdown_vault_mcp.types module.

from markdown_vault_mcp.types import NoteContent, GroupedResult, SectionHit, NoteContext

Document Types¶

`NoteContent(path, title, folder, content, frontmatter, modified_at, etag=None)` `dataclass` ¶

Full content of a document, returned by :meth:~markdown_vault_mcp.facets.reader.ReaderFacet.read.

Attributes:

Name	Type	Description
`path`	`str`	Relative path from the vault root (e.g. `Journal/note.md`).
`title`	`str`	Document title derived from the first H1 heading or filename.
`folder`	`str`	Parent folder path (empty string for root-level documents).
`content`	`str`	The note's raw file text, frontmatter block included. A caller rewriting this through `write(body, frontmatter=...)` takes the block off first, or the writer stacks a second one above it (#1391).
`frontmatter`	`dict[str, Any]`	Parsed YAML frontmatter as a dict.
`modified_at`	`float`	Last-modified time as a Unix timestamp float.
`etag`	`str \| None`	Opaque hash of file content for optimistic concurrency checks.

`RevisionContent(path, historical_path, revision, content)` `dataclass` ¶

A note's content at a git revision, returned by :meth:~markdown_vault_mcp.facets.reader.ReaderFacet.read_revision.

Deliberately carries neither etag nor modified_at: both describe the note as it is now, and an etag taken from a revision read and fed back as if_match would assert a read of the current file that the caller never made.

Attributes:

Name	Type	Description
`path`	`str`	Relative path the note has today — the one the caller asked for.
`historical_path`	`str`	Relative path the note carried at revision. Differs from `path` when the note has been renamed since.
`revision`	`str`	The commit the content was read from.
`content`	`str`	Raw markdown at that revision, frontmatter included — or, when a section was requested, just that section's body, matching what a section read of the note on disk returns. Either way this is the committed blob, so a repository with `core.autocrlf` or an `eol` attribute returns the normalised form rather than whatever line endings the file had on disk.

`NoteInfo(path, title, folder, frontmatter, modified_at, kind='note', content_chars=0)` `dataclass` ¶

Summary info for a document, returned by :meth:~markdown_vault_mcp.facets.reader.ReaderFacet.list_documents.

Attributes:

Name	Type	Description
`path`	`str`	Relative path from the vault root.
`title`	`str`	Document title.
`folder`	`str`	Parent folder path.
`frontmatter`	`dict[str, Any]`	Parsed YAML frontmatter.
`modified_at`	`float`	Last-modified time as a Unix timestamp float.
`kind`	`str`	Always `"note"` for markdown documents; distinguishes from :class:`AttachmentInfo`.
`content_chars`	`int`	Character count of the note body, frontmatter excluded, so a caller can budget batches without reading the note (#1039). `0` for a row indexed before this field existed, and for rows from queries that select a narrower column set.

`ParsedNote(path, frontmatter, title, chunks, content_hash, modified_at, links=list(), content_chars=0)` `dataclass` ¶

A parsed markdown document with extracted structure.

Attributes:

Name	Type	Description
`path`	`str`	Relative path from the vault root.
`frontmatter`	`dict[str, Any]`	Parsed YAML frontmatter as a dict.
`title`	`str`	Document title derived from the first H1 heading or filename.
`chunks`	`list[Chunk]`	Ordered list of content chunks split by heading.
`content_hash`	`str`	SHA-256 hash of the raw file content for change detection.
`modified_at`	`float`	Last-modified time as a Unix timestamp float.
`links`	`list[LinkInfo]`	All links extracted from the document body.
`content_chars`	`int`	Character count of the document body as parsed: frontmatter removed and surrounding whitespace normalised by the frontmatter parser. Measured before chunking, because `chunk_overlap_words` duplicates text across chunks and would inflate any count summed from :attr:`chunks`.

`Chunk(heading, heading_level, content, start_line)` `dataclass` ¶

A chunk of a document, typically a section under a heading.

Attributes:

Name	Type	Description
`heading`	`str \| None`	Section heading text, or `None` for the document intro.
`heading_level`	`int`	Markdown heading level (1-6); 0 for intro chunks.
`content`	`str`	Plain text content of this chunk.
`start_line`	`int`	1-based line number where this chunk begins in the source.

Search & Link Types¶

`SearchResult(path, title, folder, heading, content, score, search_type, frontmatter)` `dataclass` ¶

A search result from :meth:~markdown_vault_mcp.facets.reader.ReaderFacet.search.

Attributes:

Name	Type	Description
`path`	`str`	Relative path of the document containing this chunk.
`title`	`str`	Document title.
`folder`	`str`	Parent folder path.
`heading`	`str \| None`	Section heading this chunk falls under, or `None` for the intro.
`content`	`str`	Matched chunk text — a query-relevant snippet by default (approximately `snippet_words` words plus optional leading/trailing ellipsis markers, centred on matched terms). Pass `snippet_words=0` to `search` for the full chunk verbatim, or recover the full chunk after seeing a snippet via `read(path, section=heading)`.
`score`	`float`	Relevance score. Higher is better; not comparable across search types.
`search_type`	`Literal['keyword', 'semantic', 'hybrid']`	`"keyword"` (BM25), `"semantic"` (cosine similarity), or `"hybrid"` (chunk appeared in both keyword and semantic channels).
`frontmatter`	`dict[str, Any]`	Parsed YAML frontmatter of the parent document.

`SectionHit(heading, content, score)` `dataclass` ¶

One section's contribution to a :class:GroupedResult.

Attributes:

Name	Type	Description
`heading`	`str \| None`	Section heading text, or `None` for the document intro.
`content`	`str`	Matched snippet — query-relevant window by default, or full chunk if `snippet_words=0` was passed.
`score`	`float`	Chunk-level relevance score after length-downweight. Not comparable across search modes.

`GroupedResult(path, title, folder, score, search_type, frontmatter, sections)` `dataclass` ¶

A file-grouped search result.

Replaces the flat per-chunk :class:SearchResult across search, get_similar, and get_context.similar. See issue #469.

Attributes:

Name	Type	Description
`path`	`str`	Relative path of the document.
`title`	`str`	Document title.
`folder`	`str`	Parent folder path.
`score`	`float`	File-level score = `max(section.score for section in sections)`.
`search_type`	`Literal['keyword', 'semantic', 'hybrid']`	`"keyword"`, `"semantic"`, or `"hybrid"`.
`frontmatter`	`dict[str, Any]`	Parsed YAML frontmatter.
`sections`	`list[SectionHit]`	Up to the per-file cap best-matching sections, sorted by `(score DESC, start_line ASC, section_id ASC)` so ties surface in document order — the `section_id` key gives a fully deterministic order even when chunks share a `start_line` (e.g. word-split fragments of one oversize source line).

`FTSResult(path, title, folder, heading, content, score, chunk_count=1, start_line=0, section_id=0)` `dataclass` ¶

A raw search result from the FTS5 index layer.

Attributes:

Name	Type	Description
`path`	`str`	Relative path of the document containing this chunk.
`title`	`str`	Document title.
`folder`	`str`	Parent folder path.
`heading`	`str \| None`	Section heading this chunk falls under, or `None`.
`content`	`str`	Matched chunk text — full chunk by default; truncated to a tokenizer-aware snippet when `snippet_words` is passed to the search call.
`score`	`float`	BM25 relevance score (higher is better).
`chunk_count`	`int`	Total number of chunks belonging to the parent document.
`start_line`	`int`	Line number of the chunk's first line in the source document. Defaults to `0` for the document intro chunk and as a fallback when the underlying section row cannot be resolved.
`section_id`	`int`	`sections` table rowid of the matched chunk, used as the final deterministic tie-break when chunks share both `score` and `start_line` (e.g. word-split fragments of one oversize source line). Defaults to `0` when the section row cannot be resolved (legacy index) or for non-keyword channels.

`BacklinkInfo(source_path, source_title, link_text, link_type, fragment=None, raw_target='')` `dataclass` ¶

A document that links to a given path, returned by :meth:~markdown_vault_mcp.facets.graph.GraphFacet.get_backlinks.

Attributes:

Name	Type	Description
`source_path`	`str`	Relative path of the document containing the link.
`source_title`	`str`	Title of the linking document.
`link_text`	`str`	Display text of the link.
`link_type`	`Literal['markdown', 'wikilink', 'reference']`	Link syntax: `"markdown"` (`[text](path)`), `"wikilink"` (`[[path]]`), or `"reference"` (`[text][ref]`).
`fragment`	`str \| None`	Heading anchor (`#section`) if present in the link target.
`raw_target`	`str`	The unresolved link target exactly as written in the source.

`OutlinkInfo(target_path, link_text, link_type, fragment=None, raw_target='', exists=False)` `dataclass` ¶

A link from a document to another path, returned by :meth:~markdown_vault_mcp.facets.graph.GraphFacet.get_outlinks.

Attributes:

Name	Type	Description
`target_path`	`str`	Resolved relative path of the link target.
`link_text`	`str`	Display text of the link.
`link_type`	`Literal['markdown', 'wikilink', 'reference']`	Link syntax: `"markdown"`, `"wikilink"`, or `"reference"`.
`fragment`	`str \| None`	Heading anchor if present.
`raw_target`	`str`	The unresolved link target exactly as written.
`exists`	`bool`	`True` if the target document exists in the vault.

`BrokenLinkInfo(source_path, source_title, target_path, link_text, link_type, fragment=None, raw_target='')` `dataclass` ¶

A link whose target does not exist, returned by :meth:~markdown_vault_mcp.facets.graph.GraphFacet.get_broken_links.

Attributes:

Name	Type	Description
`source_path`	`str`	Relative path of the document containing the broken link.
`source_title`	`str`	Title of the linking document.
`target_path`	`str`	Resolved path the link points to (does not exist).
`link_text`	`str`	Display text of the link.
`link_type`	`Literal['markdown', 'wikilink', 'reference']`	Link syntax: `"markdown"`, `"wikilink"`, or `"reference"`.
`fragment`	`str \| None`	Heading anchor if present.
`raw_target`	`str`	The unresolved link target exactly as written.

`LinkInfo(target_path, link_text, link_type, fragment=None, raw_target='')` `dataclass` ¶

A link extracted from a markdown document.

Attributes:

Name	Type	Description
`target_path`	`str`	Resolved relative path of the link target.
`link_text`	`str`	Display text of the link.
`link_type`	`Literal['markdown', 'wikilink', 'reference']`	Link syntax: `"markdown"` (`[text](path)`), `"wikilink"` (`[[path]]`), or `"reference"` (`[text][ref]`).
`fragment`	`str \| None`	Heading anchor (`#section`) if present in the link target.
`raw_target`	`str`	The unresolved link target exactly as written in the source.

`NoteContext(path, title, folder, frontmatter, modified_at, backlinks, outlinks, similar, folder_notes, tags)` `dataclass` ¶

Consolidated context for a document, returned by :meth:~markdown_vault_mcp.facets.reader.ReaderFacet.get_context.

Attributes:

Name	Type	Description
`path`	`str`	Relative path from the vault root.
`title`	`str`	Document title.
`folder`	`str`	Parent folder path.
`frontmatter`	`dict[str, Any]`	Parsed YAML frontmatter.
`modified_at`	`float`	Last-modified time as a Unix timestamp float.
`backlinks`	`list[BacklinkInfo]`	Documents that link to this document.
`outlinks`	`list[OutlinkInfo]`	Links from this document with existence flags.
`similar`	`list[GroupedResult]`	Up to `similar_limit` semantically similar notes, field-collapsed. Each entry is a :class:`GroupedResult` with exactly one section (chunks_per_file=1 by default).
`folder_notes`	`list[str]`	Paths of other notes in the same folder (up to 20).
`tags`	`dict[str, list[str]]`	Tag values for each indexed frontmatter field.

`MostLinkedNote(path, title, folder, backlink_count)` `dataclass` ¶

A document with its inbound backlink count, returned by :meth:~markdown_vault_mcp.facets.graph.GraphFacet.get_most_linked.

Attributes:

Name	Type	Description
`path`	`str`	Relative path from the vault root.
`title`	`str`	Document title.
`folder`	`str`	Parent folder path (empty string for root-level documents).
`backlink_count`	`int`	Number of other documents that link to this document.

Operation Results¶

`WriteResult(path, created, previous_revision=None)` `dataclass` ¶

Result of a write operation.

Attributes:

Name	Type	Description
`path`	`str`	Relative path of the document that was written.
`created`	`bool`	`True` if the document was newly created; `False` if overwritten.
`previous_revision`	`str \| None`	Set only by an overwriting :meth:`~markdown_vault_mcp.facets.writer.WriterFacet.write`; every other producer of this result — `append`, `write_attachment`, and direct `DocumentManager` writes — leaves it `None`. On a git-backed vault it names the commit holding the content the write replaced, the route back to it via :meth:`~markdown_vault_mcp.facets.reader.ReaderFacet.read_revision`. `None` whenever no commit provably holds it: a create, a vault without git, a note not committed yet, or a working copy that had already moved on from its newest commit. Content that never reached a commit is not recoverable from git, and a breadcrumb is withheld rather than pointing at an older revision that would look like the replaced content without being it. The probe and the write share the vault's write lock, so no other vault writer can change which note occupies the path in between.

`EditResult(path, replacements, match_type='exact')` `dataclass` ¶

Result of an edit operation.

Attributes:

Name	Type	Description
`path`	`str`	Relative path of the document that was edited.
`replacements`	`int`	Number of text replacements made (always 1 for exact match).
`match_type`	`str`	How the replacement was found: `"exact"` (verbatim match) or `"normalized"` (whitespace-normalised match).

`DeleteResult(path)` `dataclass` ¶

Result of a delete operation.

Attributes:

Name	Type	Description
`path`	`str`	Relative path of the document that was deleted.

`RenameResult(old_path, new_path, updated_links=0, hint=None)` `dataclass` ¶

Result of a rename operation.

Attributes:

Name	Type	Description
`old_path`	`str`	Original relative path.
`new_path`	`str`	New relative path after the rename.
`updated_links`	`int`	Number of backlinks in other documents that were rewritten.
`hint`	`str \| None`	Why no links were rewritten although `update_links` was requested — set when the renamed file is an attachment, whose references are not tracked as links (#1338). `None` otherwise.

`IndexStats(documents_indexed, chunks_indexed, skipped)` `dataclass` ¶

Statistics from :meth:~markdown_vault_mcp.facets.index.IndexFacet.build_index.

Attributes:

Name	Type	Description
`documents_indexed`	`int`	Number of documents successfully indexed.
`chunks_indexed`	`int`	Total number of chunks indexed.
`skipped`	`int`	Number of documents skipped due to parse errors.

`ReindexResult(added, modified, deleted, unchanged, skipped=0)` `dataclass` ¶

Result of :meth:~markdown_vault_mcp.facets.index.IndexFacet.reindex.

Attributes:

Name	Type	Description
`added`	`int`	Documents added since the last index.
`modified`	`int`	Documents that changed since the last index.
`deleted`	`int`	Documents removed since the last index.
`unchanged`	`int`	Documents with no changes.
`skipped`	`int`	Files present on disk that were deliberately not indexed (missing required frontmatter, matching an exclude pattern, or unparseable), whether newly skipped this scan or unchanged since they were last skipped (#665).

`VaultStats(document_count, chunk_count, folder_count, semantic_search_available, indexed_frontmatter_fields=list(), attachment_extensions=list(), link_count=0, broken_link_count=0, orphan_count=0)` `dataclass` ¶

Vault-wide statistics, returned by :meth:~markdown_vault_mcp.facets.reader.ReaderFacet.stats.

Attributes:

Name	Type	Description
`document_count`	`int`	Number of indexed markdown documents.
`chunk_count`	`int`	Total number of indexed sections (chunks).
`folder_count`	`int`	Number of distinct folder paths.
`semantic_search_available`	`bool`	`True` if a vector index is loaded and ready.
`indexed_frontmatter_fields`	`list[str]`	Frontmatter fields configured for tag indexing.
`attachment_extensions`	`list[str]`	File extensions recognised as attachments.
`link_count`	`int`	Total number of links extracted from all documents.
`broken_link_count`	`int`	Number of links whose target does not exist.
`orphan_count`	`int`	Number of documents with no inbound or outbound links.

`ChangeSet(added, modified, deleted, unchanged, skipped_unchanged=0)` `dataclass` ¶

Documents that changed since the last index build.

Attributes:

Name	Type	Description
`added`	`list[str]`	Paths of newly discovered documents.
`modified`	`list[str]`	Paths of documents whose content changed.
`deleted`	`list[str]`	Paths of documents that no longer exist on disk.
`unchanged`	`int`	Count of documents with no changes (not listed individually).
`skipped_unchanged`	`int`	Count of files previously recorded as skipped (never indexed) whose content has not changed since; they appear in no other bucket and need no re-evaluation (#665).

Attachment Types¶

`AttachmentContent(path, mime_type, size_bytes, content_base64, modified_at, etag=None)` `dataclass` ¶

Full content of an attachment, returned by :meth:~markdown_vault_mcp.facets.reader.ReaderFacet.read_attachment for non-.md files.

Attributes:

Name	Type	Description
`path`	`str`	Relative path from the vault root.
`mime_type`	`str \| None`	Detected MIME type, or `None` if unknown.
`size_bytes`	`int`	File size in bytes.
`content_base64`	`str`	Base64-encoded file content.
`modified_at`	`float`	Last-modified time as a Unix timestamp float.
`etag`	`str \| None`	Opaque hash for optimistic concurrency checks.

`AttachmentInfo(path, folder, mime_type, size_bytes, modified_at, kind='attachment')` `dataclass` ¶

Summary info for an attachment, returned by :meth:~markdown_vault_mcp.facets.reader.ReaderFacet.list_documents when include_attachments=True.

Attributes:

Name	Type	Description
`path`	`str`	Relative path from the vault root.
`folder`	`str`	Parent folder path.
`mime_type`	`str \| None`	Detected MIME type, or `None` if unknown.
`size_bytes`	`int`	File size in bytes.
`modified_at`	`float`	Last-modified time as a Unix timestamp float.
`kind`	`str`	Always `"attachment"`; distinguishes from :class:`NoteInfo`.

Git Types¶

`HistoryEntry(sha, short_sha, timestamp, author, message, paths_changed)` `dataclass` ¶

A commit that touched a note or the vault, returned by :meth:~markdown_vault_mcp.facets.reader.ReaderFacet.get_history.

Attributes:

Name	Type	Description
`sha`	`str`	Full commit SHA, in the repository's hash algorithm (40 hex digits under SHA-1, 64 under SHA-256).
`short_sha`	`str`	Abbreviated 7-character SHA.
`timestamp`	`str`	ISO 8601 commit timestamp.
`author`	`str`	Commit author name and email.
`message`	`str`	First line of the commit message.
`paths_changed`	`list[str]`	Files touched by the commit. Populated for vault-wide queries (`path=None`). Always empty for single-note queries, since the path is already determined by the query arguments — callers know which file the commit touched without needing it echoed back.

`CommitDiff(sha, short_sha, timestamp, message, diff)` `dataclass` ¶

A per-commit diff entry, returned by :meth:~markdown_vault_mcp.facets.reader.ReaderFacet.get_diff when per_commit=True.

Attributes:

Name	Type	Description
`sha`	`str`	Full commit SHA, in the repository's hash algorithm (40 hex digits under SHA-1, 64 under SHA-256).
`short_sha`	`str`	Abbreviated 7-character SHA.
`timestamp`	`str`	ISO 8601 commit timestamp.
`message`	`str`	First line of the commit message.
`diff`	`str`	Unified diff text for the commit.

Callbacks¶

WriteOperation

Type alias for the kind of write operation reported to callbacks. WriteCallback and the op argument below both reference it.

WriteOperation = Literal["write", "edit", "delete", "rename"]

WriteCallback

Type alias for the on_write callback passed to Vault. Called after each successful write operation.

WriteCallback = Callable[[Path, str, WriteOperation], None]

Arguments received by the callback:

Argument	Type	Description
`path`	`Path`	Absolute path of the modified file
`content`	`str`	New file content (empty string for binary attachments and deletes)
`op`	`WriteOperation`	Which write operation fired (see `WriteOperation` above)

Types¶

Document Types¶

NoteContent(path, title, folder, content, frontmatter, modified_at, etag=None) dataclass ¶

RevisionContent(path, historical_path, revision, content) dataclass ¶

NoteInfo(path, title, folder, frontmatter, modified_at, kind='note', content_chars=0) dataclass ¶

ParsedNote(path, frontmatter, title, chunks, content_hash, modified_at, links=list(), content_chars=0) dataclass ¶

Chunk(heading, heading_level, content, start_line) dataclass ¶

Search & Link Types¶

SearchResult(path, title, folder, heading, content, score, search_type, frontmatter) dataclass ¶

SectionHit(heading, content, score) dataclass ¶

GroupedResult(path, title, folder, score, search_type, frontmatter, sections) dataclass ¶

FTSResult(path, title, folder, heading, content, score, chunk_count=1, start_line=0, section_id=0) dataclass ¶

BacklinkInfo(source_path, source_title, link_text, link_type, fragment=None, raw_target='') dataclass ¶

OutlinkInfo(target_path, link_text, link_type, fragment=None, raw_target='', exists=False) dataclass ¶

BrokenLinkInfo(source_path, source_title, target_path, link_text, link_type, fragment=None, raw_target='') dataclass ¶

LinkInfo(target_path, link_text, link_type, fragment=None, raw_target='') dataclass ¶

NoteContext(path, title, folder, frontmatter, modified_at, backlinks, outlinks, similar, folder_notes, tags) dataclass ¶

MostLinkedNote(path, title, folder, backlink_count) dataclass ¶

Operation Results¶

WriteResult(path, created, previous_revision=None) dataclass ¶

EditResult(path, replacements, match_type='exact') dataclass ¶

DeleteResult(path) dataclass ¶

RenameResult(old_path, new_path, updated_links=0, hint=None) dataclass ¶

IndexStats(documents_indexed, chunks_indexed, skipped) dataclass ¶

ReindexResult(added, modified, deleted, unchanged, skipped=0) dataclass ¶

VaultStats(document_count, chunk_count, folder_count, semantic_search_available, indexed_frontmatter_fields=list(), attachment_extensions=list(), link_count=0, broken_link_count=0, orphan_count=0) dataclass ¶

ChangeSet(added, modified, deleted, unchanged, skipped_unchanged=0) dataclass ¶

Attachment Types¶

AttachmentContent(path, mime_type, size_bytes, content_base64, modified_at, etag=None) dataclass ¶

AttachmentInfo(path, folder, mime_type, size_bytes, modified_at, kind='attachment') dataclass ¶

Git Types¶

HistoryEntry(sha, short_sha, timestamp, author, message, paths_changed) dataclass ¶

CommitDiff(sha, short_sha, timestamp, message, diff) dataclass ¶

Callbacks¶

`NoteContent(path, title, folder, content, frontmatter, modified_at, etag=None)` `dataclass` ¶

`RevisionContent(path, historical_path, revision, content)` `dataclass` ¶

`NoteInfo(path, title, folder, frontmatter, modified_at, kind='note', content_chars=0)` `dataclass` ¶

`ParsedNote(path, frontmatter, title, chunks, content_hash, modified_at, links=list(), content_chars=0)` `dataclass` ¶

`Chunk(heading, heading_level, content, start_line)` `dataclass` ¶

`SearchResult(path, title, folder, heading, content, score, search_type, frontmatter)` `dataclass` ¶

`SectionHit(heading, content, score)` `dataclass` ¶

`GroupedResult(path, title, folder, score, search_type, frontmatter, sections)` `dataclass` ¶

`FTSResult(path, title, folder, heading, content, score, chunk_count=1, start_line=0, section_id=0)` `dataclass` ¶

`BacklinkInfo(source_path, source_title, link_text, link_type, fragment=None, raw_target='')` `dataclass` ¶

`OutlinkInfo(target_path, link_text, link_type, fragment=None, raw_target='', exists=False)` `dataclass` ¶

`BrokenLinkInfo(source_path, source_title, target_path, link_text, link_type, fragment=None, raw_target='')` `dataclass` ¶

`LinkInfo(target_path, link_text, link_type, fragment=None, raw_target='')` `dataclass` ¶

`NoteContext(path, title, folder, frontmatter, modified_at, backlinks, outlinks, similar, folder_notes, tags)` `dataclass` ¶

`MostLinkedNote(path, title, folder, backlink_count)` `dataclass` ¶

`WriteResult(path, created, previous_revision=None)` `dataclass` ¶

`EditResult(path, replacements, match_type='exact')` `dataclass` ¶

`DeleteResult(path)` `dataclass` ¶

`RenameResult(old_path, new_path, updated_links=0, hint=None)` `dataclass` ¶

`IndexStats(documents_indexed, chunks_indexed, skipped)` `dataclass` ¶

`ReindexResult(added, modified, deleted, unchanged, skipped=0)` `dataclass` ¶

`VaultStats(document_count, chunk_count, folder_count, semantic_search_available, indexed_frontmatter_fields=list(), attachment_extensions=list(), link_count=0, broken_link_count=0, orphan_count=0)` `dataclass` ¶

`ChangeSet(added, modified, deleted, unchanged, skipped_unchanged=0)` `dataclass` ¶

`AttachmentContent(path, mime_type, size_bytes, content_base64, modified_at, etag=None)` `dataclass` ¶

`AttachmentInfo(path, folder, mime_type, size_bytes, modified_at, kind='attachment')` `dataclass` ¶

`HistoryEntry(sha, short_sha, timestamp, author, message, paths_changed)` `dataclass` ¶

`CommitDiff(sha, short_sha, timestamp, message, diff)` `dataclass` ¶