Struct DictionaryBuilder 

pub struct DictionaryBuilder { /* private fields */ }

Implementations

impl DictionaryBuilder

pub fn new(metadata: Metadata) -> Self

pub fn with_context_id_freq(self, path: impl Into<PathBuf>) -> Self

Attach a bundled context-ID frequency histogram used for the connection-cost remap ranking.

Arguments

• path - Histogram file (as produced by the ctxfreq instrumentation).

Returns

The builder with the frequency source attached.

pub fn build_dictionary( &self, input_dir: &Path, output_dir: &Path, ) -> LinderaResult<()>

Build all dictionary artifacts from input_dir into output_dir.

The independent stages run concurrently on non-wasm targets and sequentially on wasm (which has no OS threads). The output files are identical regardless of the path taken.

Arguments

• input_dir - Directory containing the source dictionary files.
• output_dir - Directory to write the built artifacts into.

Returns

Ok(()) on success, or the first stage error in stage order.

pub fn build_metadata( &self, output_dir: &Path, remap: Option<&Arc<ContextIdMap>>, ) -> LinderaResult<()>

Write metadata.json, stamping the dictionary format version and embedding the context-ID permutation when one was applied.

The format version is taken from DICTIONARY_FORMAT_VERSION rather than from the source metadata: it describes the artifacts this builder just wrote, so the builder is the only thing that can state it truthfully. A source metadata.json carrying a stale (or invented) version must not be able to mislabel a freshly built dictionary.

Persisting the permutation is what lets a user dictionary compiled later be relabeled into the same ID space (see crate::dictionary::UserDictionary::remap_context_ids).

Arguments

• output_dir - Directory to write metadata.json into.
• remap - The permutation applied by this build, if any.

pub fn build_character_definition( &self, input_dir: &Path, output_dir: &Path, ) -> LinderaResult<CharacterDefinition>

pub fn build_unknown_dictionary( &self, input_dir: &Path, output_dir: &Path, chardef: &CharacterDefinition, remap: Option<Arc<ContextIdMap>>, ) -> LinderaResult<()>

pub fn build_prefix_dictionary( &self, input_dir: &Path, output_dir: &Path, remap: Option<Arc<ContextIdMap>>, ) -> LinderaResult<()>

pub fn build_connection_cost_matrix( &self, input_dir: &Path, output_dir: &Path, remap: Option<Arc<ContextIdMap>>, ) -> LinderaResult<()>

pub fn build_user_dictionary( &self, input_file: &Path, output_file: &Path, ) -> LinderaResult<()>

pub fn build_user_dict( &self, input_file: &Path, ) -> LinderaResult<UserDictionary>

Trait Implementations

impl Clone for DictionaryBuilder

fn clone(&self) -> DictionaryBuilder

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.