Documentation ¶
Index ¶
- func Cluster(piles []*pals.Pile, cfg ClusterConfig) (int, [][]*pals.Pile)
- func Group(clust [][]*pals.Pile, cfg GroupConfig) []graph.Nodes
- func Piles(in *gff.Reader, overlap int, pf pals.PairFilter) ([]*pals.Pile, error)
- func WriteJSON(cc []graph.Nodes, w io.Writer) error
- type ClusterConfig
- type GroupConfig
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func Cluster ¶
Cluster performs sub-pile clustering according to the config provided. The number of sub-piles and a collection of piles broken into sub-piles is returned.
func Group ¶
func Group(clust [][]*pals.Pile, cfg GroupConfig) []graph.Nodes
Group clusters the input pile collection based on the existence of satisfactory image alignments according to the provided config. Piles with nil locations are ignored. Checks are performed to ensure that the produced clusters can be unambiguously assigned to a DNA strand.
Types ¶
type ClusterConfig ¶
type ClusterConfig struct { // BandWidth specifies the maximum fractional distance between // endpoints of images being clustered into sub-piles. See // turner.Cluster for details. BandWidth float64 // RequiredCover specifies the target coverage fraction for // each input pile. If RequiredCover is greater than 1, all // all sub-piles are retained depending on the values of // KeepOverlaps and OverlapThresh. RequiredCover float64 // OverlapStrictness specifies the clustering behaviour. // If OverlapStrictness is zero, all clusters are passed // returned. If set to one, clusters containing clusters // with greater depth are rejected. If set to two, contained // features overlapping by more than OverlapThresh fraction // of the smaller pile are are discarded. OverlapStrictness byte OverlapThresh float64 // Procs specifies the number of independent clustering // instances to run in parallel. If zero, only single threaded // operation is performed. Procs int }
ClusterConfig specifies Cluster behaviour.
type GroupConfig ¶
type GroupConfig struct { // PileDiff specifies the acceptable fractional difference between // piles for assignment to the same group. PileDiff float64 // ImageDiff specifies the acceptable fractional difference between // an image and its containing pile when considering a pile for // assignment based on that image. ImageDiff float64 // When Classic is true, Group will run a reasonable approximation // of the original C PILER family grouping. Classic bool }
GroupConfig specifies Group behaviour
Click to show internal directories.
Click to hide internal directories.