Iterate over all provenance graph hashes on disk.

Parses all tracked source files and writes `CALLS` edges between entity nodes. For every tracked file, this function: 1. Extracts function/method ent

Enrich the KG with change data from the current view's history. For each change: - Creates a `change:{short_hash}` node with message and timestamp -

Parses all supported source files in the working copy and creates `entity:{file}:{name}:{line}` nodes with `DEFINES` edges from the corresponding file

Enrich the KG with tracked file nodes. Creates a `file:{path}` node for each file tracked in the repository.

Enrich the KG with AST entities from tracked files. Enrich the KG with `INCLUDES` edges derived from import entities. Re-parses tracked source files

Enrich the KG with module (directory) nodes and `PART_OF` edges. For each directory that directly contains at least one tracked file, creates a `modu

Get the 1-hop neighborhood of a node (outgoing + incoming edges + connected nodes).

Check if this change knows about another change. A change "knows" another if it's either a dependency or extra_known.

Short label for display.

Build a single layer from an in-memory set of `(relative_path, bytes)` files plus optional whiteouts. Whiteouts mark deletions relative to a lower la

Returns the leaf operations (new content for Insert/Modify, old content for Delete). Returns an empty slice for Restore operations.

Get the number of lines in this range.

List tags in the current view.

Load the provenance accumulator for a session (read-only, locked). Best-effort: returns `None` on failure (logged, never fatal).

Load all content chunks for a change, in order. Reads from CHANGE_CHUNKS (manifest) + CONTENT_CHUNKS (data). Decompresses each chunk and returns the

Load global ignore rules from ~/.config/atomic/ignore

Load a local identity by name, falling back to a case-insensitive match. Tenant slugs are lowercase (e.g. `aaron`), but locally stored identities are

Load repository-local ignore rules from .atomicignore

Load a provenance graph from the repository by hash.

Load the unhashed data for a change (if any). Returns `None` if no unhashed data is stored.

Get forward history log from a view. Returns an iterator over history entries starting from the given sequence number and proceeding forward (oldest

Helper: make a GraphNode<Option<Hash>> with a specific hash.

The stamp to write onto sessions born under the managed run.

The view declared by the governing managed run, if any.

Mark an edge as deleted.

Marks a trunk as deleted. Does not remove the trunk, just updates its state.

Materialize the working copy using parallel file processing. This is an optimized version of `materialize` that: 1. Buffers each file's content in me

Set the maximum vertices per file.

Optionally switch to the new view and print hint.

Merge statistics from another recording.

Merges another stats instance into this one.

Merge stats from another operation.

Merge consecutive Equal operations in the result.

Merge adjacent hunks of the same kind.

Get the message or a default placeholder.

message_or_default Get the message or a default placeholder. Returns the change message if one is set, or "(no message)" as a fallback for display pu

Add metadata key-value pair.

Set the minimum width for this column. # Arguments `width` - The minimum width in characters # Returns Self for method chaining.

Get the number of missing context conflicts.

Set all model usage entries at once.

Get the modification time of a file.

Marks a file for move/rename.

Returns a human-readable name for this section type. Used in error messages and debug output. # Examples ```rust use atomic_core::change::format_v3

Get only the changed hunks from the new line.

Return `max_id + 1`, or error if the ID space is exhausted.

Find all nodes of a given kind.

Normalize a path for storage, optionally stripping a repo root prefix. This handles the case where absolute paths are accidentally passed in. On macO

Set optional notes.

Set the number of worker threads.

Get only the changed hunks from the old line.

Insert only specific changes.

Get the number of order conflicts.

Build the org-scoped base URL. Given `url = "https://atomic.storage"` and `org = "alice"`, returns `"https://alice.atomic.storage"`. Given `url = "h

Reconstruct a file's bytes by walking the CRDT layer. This is the alternative to `output::repo::output_file_with_filter`: it derives line order from

Set whether to output name conflicts.

Set output tokens.

Iterate through line pairs.

Enable parallel output.

Parse a change reference string. Supported formats: - `@` or empty: Last change - `@~1`, `@~2`, etc.: Relative to last - Anything else: Treated as ha

Parse "Atomic-Changes: HASH1, HASH2, ..." from a commit message.

Parse a change hash from a string (full or abbreviated).

Parse minimal identity info from an identity.toml file.

Parse an identity type from a string.

Parse a PR/MR number from a commit message. Supports multiple forge formats: - GitHub: `(#42)` or `Merge pull request #42 from ...` - GitLab:

Parse the reference string into a ChangeRef.

Get the path as a PathBuf.

Get the parent path of a file path. Returns `None` if the path has no parent (is a top-level file).