pydistsim.utils.localization.basestitcher.BaseStitcher

class BaseStitcher(selector=None)

Bases: object

Base class for stitching two clusters.

Subclasses should implement own methods for subcluster stitching named:

stitch_subclusters

These methods should take two subclusters as dictionary {node: position} where position is array([x,y,theta]) with at least one common node and return rotation matrix R, scaling factor s and translation vector t.

If stitch_subclusters method can’t be sure that all parameters are correct i.e. when common nodes are not in generic formation then it should return None for all parameters.

Methods

__init__
align	Align (modify) src w.r.t.
intrastitch
stitch	Stitch src cluster to dst cluster based on selector and its criteria.
transform	Transform node position.

_get_rotation_matrix_horn(commonNodes, dstSubPos, srcSubPos, p_d, p_s)

Horn method for calculating rotation matrix.

_stitch(dst, src, is_intra=False)

Iteratively:

• selects subclusters to stitch using self.selector

• stitch - find out R, s and t

• append transformed node positions from src to dst subcluster

Return stitched dictionary where:

keys - are pairs of indexes of all stitched subclusters values - are their R, s and t transformation parameters

or all of them are None if can’t be calculated reliably or () if src subcluster is subset of dst

align(dst, src)

Align (modify) src w.r.t. dst.

stitch(dst, src, do_intra=True)

Stitch src cluster to dst cluster based on selector and its criteria.

1. Stitch all src subclusters to dst.

2. Append unstitched source subclusters to dst cluster.

3. If do_intra stitch subclusters internally.

src and dst clusters are in format defined in pydistsim.utils.memory.positions.Positions.subclusters: [{node1: array(x1,y1), node2: array(x2,y2), …}, …]

transform(R, s, t, pos, ori=nan)

Transform node position.