Module Documentation

Core

imagepypelines Package

Functions

add_plugin(plugin_name, plugin_module[, …])

adds the given plugin to ImagePypelines

arrsummary(arr)

returns a Summarizer object for the given array

blockify([kwargs, batch_type, types, …])

decorator which converts a normal function into a un-trainable block which can be added to a pipeline.

connect_to_dash(name, host, port)

Connects every pipeline in this session to

get_logger(name[, pipeline, parent])

Creates a new child logging adapter from the given parent (root logger by default)

get_master_logger()

get_plugin_by_name(plugin_name)

fetches the plugin module using its name

load_plugins()

Load all installed plugins to the imagepypelines namespace

make_numbered_prefix(file_number[, …])

returns a number string designed to be used in the prefix of systematic outputs.

n_dashboards()

returns the number of connected dashboards

passgen(passwd[, salt])

generate a hashed key from a password

prevent_overwrite(filename[, create_file])

checks to see if a file or directory already exists and creates a new filename if it does.

require(plugin_name)

check to make sure the given plugin is loaded and raise an error if it is not in the imagepypelines namespace

set_log_level(log_level)

sets the global master logger level

timer(func)

Decorator to time how long a func takes to run in milliseconds

timer_ms(func)

Decorator to time how long a func takes to run in milliseconds

Classes

Block([name, batch_type, types, shapes, …])

a contained algorithmic element used to construct pipelines.

BlockError

Error raised within a Block

DashboardComm

Object to send messages from the pipelines to dashboard(s)

DashboardWarning

Warning for connection failure to dashboard

Data(data)

Object to batch lists or arrays as block processable data

FuncBlock(func[, preset_kwargs])

Block that will run any function you give it, either unfettered through the __call__ function, or with optional hardcoded parameters for use in a pipeline.

ImagepypelinesLogger(name[, level])

subclass of logging.Logger that can be pickled, also adds colored logging outputs if desired.

Input([index])

An object to inject data into the graph

Leaf(var_name)

a block to act as a leaf node in the Pipeline Graph.

Pipeline([tasks, name])

processing algorithm manager for simple pipeline construction

PipelineBlock(pipeline, fetch)

Block which runs a pipeline internally (used for nesting pipelines within pipelines)

PipelineError

Error raised within a Pipeline

Timer()

Timer which can be used to time processes

Class Inheritance Diagram

Inheritance diagram of imagepypelines.core.Block.Block, imagepypelines.core.Exceptions.BlockError, imagepypelines.core.DashboardComm.DashboardComm, imagepypelines.core.Exceptions.DashboardWarning, imagepypelines.core.Data.Data, imagepypelines.core.block_subclasses.FuncBlock, imagepypelines.Logger.ImagepypelinesLogger, imagepypelines.core.block_subclasses.Input, imagepypelines.core.block_subclasses.Leaf, imagepypelines.core.Pipeline.Pipeline, imagepypelines.core.block_subclasses.PipelineBlock, imagepypelines.core.Exceptions.PipelineError, imagepypelines.core.util.Timer