API¶
Import the package as:
import elpigraph
Principal graphs¶
computeElasticPrincipalCurve(X, NumNodes[, …]) |
Construct a principal elastic curve |
computeElasticPrincipalCircle(X, NumNodes[, …]) |
Construct a principal elastic circle |
computeElasticPrincipalTree(X, NumNodes[, …]) |
Construct a principal elastic tree |
generateInitialConfiguration(X, Nodes[, …]) |
Produce an initial graph with a given structure |
Graph editing¶
addPath(X, PG, source, target[, n_nodes, …]) |
Add path (principal curve) between two nodes in the principal graph |
delPath(X, PG, source, target[, …]) |
Delete path between two nodes in the principal graph |
findPaths(X, PG[, min_path_len, nnodes, …]) |
This function tries to add extra paths to the graph by computing a series of principal curves connecting two nodes and retaining plausible ones using heuristic parameters |
ExtendLeaves(X, PG[, Mode, ControlPar, …]) |
Extend leaves with additional nodes |
CollapseBranches(X, PG[, Mode, ControlPar, …]) |
Filter “small” branches |
ShiftBranching(X, PG[, SelectionMode, …]) |
Move branching nodes to areas with higher point density |
fineTuneBR(X, NumNodes[, NumEdges, …]) |
Expand the nodes around a branching point. |
Utils¶
utils.getProjection(X, PG) |
Compute graph projection from principal graph dict (result stored in PG dict under ‘projection’ key) |
utils.getPseudotime(X, PG, source[, target, …]) |
Compute pseudotime given source: int source node target: int optional target node nodes_to_include: list optional nodes to include in the path (useful for complex topologies with loops, where multiple paths are possible between 2 nodes) project: bool if False, will save computation time by using the projection already stored in PG dict (computed using elpigraph.utils.getProjection) |
utils.getWeights(X[, bandwidth, griddelta, …]) |
Get point weights as the inverse density of data X: np.array, (n_sample x n_dims) bandwidth: sklearn KernelDensity bandwidth if method == ‘sklearn’ griddelta: FFTKDE grid step size if method ==’fft’ exponent: density values are raised to the power of exponent |
Plotting¶
plot.PlotPG(X, PG[, X_color, Node_color, …]) |