Building blocks

Blocks

Block is a complex object that bind user interface with major experiment steps. There is no predefined experiment sctructure, user choose appropriate block from available pallet. By defining Internally each block consist of multiply action, states and user input processing. Blocks can be splitted in a few groups by its purpose.

Main Block types:

• Data input - file upload, gene set picker, GEO fetch, ...
• Data conversion - dealing with missing data, normalisatioon, unit conversions
• Classifier - creates model from annotated data set, possible also compute classification error
• Analizer - apllies classifier model on data to find precision or other properties
• Visializer - render result in human friendly form, e.g.: graphics, tables, schemes.
• Container/Meta-plugin - encapsulate other block inside to implement more complex opertions. Currently only Cross validation block has this type.

Base functionality

Each block has:

• Title - Human readable title
• Code - machine identifier
• UUID - unique identifier after generated for each added block
• Name - unique to experiment human readable identifier, can be altered by user
• Type - semantic group, from BlockTypes definition
• Scope - defines visibility scope, top most on is root
• Finite state machine (FSM) definition + state - hidden from user, FSM control block behavior and interaction with other experiments entities.
• User actions - shown to user when applicable

Each block could have:

• Exported variables - what block to generate during experiment. Could be added dynamically upon user input.
• Input ports - defines data flow by binding block the inputs to exported variables of visible blocks in the current scope.
• Block parameters - specific settings for block. Can have default values, be requiered or not
• Block specific elements