bend-edit-distance

Search uses release 0.2.2.0 · 210 definitions

Releases (4)

type ModeEdlib sentinels are distinct constructors, never wrapped U32 values.
type Task
type Threshold
type EqualityPair
type CigarFormat
type AlignConfig
type AlignJobEach independent batch job keeps its own configuration and orientation.
type Status
type EditDistance
type Coordinate
type Operation
type CigarRunEncoded runs have positive counts and are merged after mapping to the selected alphabet: adjacent Match/Mismatch operations form one Standard M run.
type AlignResult
def default_config
def pair_matchesLiteral equality plus explicitly configured symmetric pairs.
def equal
def end_coordinate
def distance_show
def coordinate_show
def status_show
def operation_show
def operation_char
def cigar_run_step
def alignment_runs_reversedThe current run depends on the preceding symbol, so this scan is sequential.
def alignment_runs
def cigar_runs_show
def alignment_to_cigar
def coordinates_show
def starts_show
def alignment_show
def cigar_showCIGAR contains only decimal counts and MID=X, so needs no JSON escaping.
def result_show
def not_found
def within_threshold
def distance_result

def distancePublic CPU library entry point, independent of CLI and foreign transport.
def default_config
def align
def align_batchOne result per job, in input order, with independent CPU alignment leaves.
def alignment_to_cigarFormat any path, including an explicitly empty path, independently of align.

type LocationMyers owns nonempty distance and location searches.
type Best
def update_best
def ordered
def ends
def starts
def span
def path_result
def location_valid
def path_checked
def path_validate
def path_reconstruct
def result
def empty_resultPinned Edlib returns before applying k or calculating starts/path on empty inputs, including PATH.

type WeightedJob
type Group
type PlanBuild the entire plan before entering the alignment fork tree.
type IndexedResult
def axes
def extra_work
def estimate_axes
def estimate_sizesA scheduling heuristic, not a timing prediction: input scans, 256-symbol profiles, full Myers columns, worst-case HW start searches and PATH DP.
def estimate
def indexed
def heavier
def add
def choose
def place
def partitionLargest estimated jobs go to the lighter branch.
def group_jobs
def plan_step
def plan
def prepare
def single
def serial
def executeOnly ready, independent subplans enter parallel calls.
def earlier
def results
def ordered
def align_batch

def initial_row
def cellD[i,j] = min(D[i-1,j] + 1, D[i,j-1] + 1, D[i-1,j-1] + !equal(pairs, row_symbol, column_symbol)).
def row_cellsConsume the old row as the new row is built.
def next_row
def last_cell
def rows
def distance_shorter
def distance_with_equalitiesPublic entry point: importing this module does not import or call Myers.
def distance
def initial_semiglobal_rowDirectional DP keeps query rows and target columns in their original order.
def final_row
type Best
def collect_best
def collect
def semiglobal_endpoints
def semiglobal
def semiglobal_result.accept
def semiglobal_result
def alignThe independent worker supports configured DISTANCE in every mode, including Edlib's empty-input early return before threshold filtering.