[][src]Crate pointcloud

Point Cloud

Abstracts data access over several files and glues metadata files to vector data files

Modules

data_sources

Some data sources and a trait to dimension and uniformly reference the data contained. The only currently supported are memmaps and ram blobs.

glued_data_cloud

Simple gluing structs that abstracts away multi cloud access

label_sources

Some label sets to modularly glue together with the data sources.

loaders

Loaders for datasets. Just opens them up and returns a point cloud.

pc_errors

The errors that can occur when a point cloud is loading, working or saving

summaries

Summaries for some label types

Structs

AdjMatrix

A sparse adjacency matrix.

CosineSim

Not a norm! Still, helpful for document clouds and the like

DenseIter
L1

L1 norm, the sum of absolute values

L2

L2 norm, the square root of the sum of squares

Linfty

L infity norm, the max of the absolute values of the elements

SimpleLabeledCloud

Simply shoves together a point cloud and a label set, for a modular label system

SummaryCounter

Simply shoves together a point cloud and a label set, for a modular label system

Enums

Point

An actual point, self contained with it's own objects on the heap.

PointRef

Reference to a point inside of a dataset.

Traits

LabelSet

A trait for a container that just holds labels. Meant to be used in conjunction with SimpleLabeledCloud to be and easy label or metadata object.

LabeledCloud

A point cloud that is labeled

MetaCloud

Allows for expensive metadata, this is identical to the label trait, but enables slower update

Metric

The trait that enables a metric

NamedCloud

Enables the points in the underlying cloud to be named with strings.

PointCloud

Base trait for a point cloud

Summary

A summary for labels and metadata. You can make this an empty zero sized type for when you don't need it.

Type Definitions

DefaultCloud

A sensible default for an unlabeled cloud

DefaultLabeledCloud

A sensible default for an labeled cloud

PointIndex

To make things more obvious, we type the point index. This is abstracted over the files that were used to build the point cloud

PointName

To make things more obvious, we type the point name that we pull from the label CSV