The core of ADAMS is the workflow engine, which follows the philosophy of less is more. Instead of letting the user place operators (or actors in ADAMS terms) on a canvas and then manually connect inputs and outputs, ADAMS uses a tree-like structure. This structure and the control actors define how the data is flowing in the workflow, no explicit connections necessary. The tree-like structure stems from the internal object representation and the nesting of sub-actors within actor-handlers.
|Machine learning/data mining||WEKA, WEKA webservice, MOA, MEKA, deeplearning4j, parameter optimization, experiment generation on-the-fly, setup generators, time series|
|Data processing||WEKA, R-Project, XML, XSLT, XPath, HTML, JSON|
|Streaming||MOA, Twitter (record/replay)|
|Spreadsheets||MS Excel (r/w), ODF (r/w), CSV (r/w), Gnumeric (r/w), fixed column (r/w)|
|Databases||MS Access, MySQL, SQLite, PostgreSQL, HSQL, MSSQL, Sybase, JDBC|
|Imaging||ImageJ, JAI, BoofCV, OpenIMAJ, ImageMagick, Gnuplot, LIRE, OCR (tesseract), Barcodes (Zxing)|
|Graphics output||BMP, JPG, PNG, TIF, PDF, RAW (dcraw, ufraw)|
|Visualization||Scatter and line plots, Control charts, Images, GIS (OpenStreetMap)|
|Web||HTTP, FTP, SFTP, SSH, Email, Webservices|
|Other||de/-compression (tar, zip, bzip2, gzip, lzma), Java code generation|
- Horizontal and vertical indicators show how the data flows
- Each actor has its own icon, since the name can be changed arbitrarily
- Four types of actors: standalone (red; no input, no output), source (orange; only output), transformer (green; input and output) and sink (greyl; only input)
- Control actors, i.e., actors that determine the flow of data or flow execution are blue
- Actors with parameters usually show a so-called quick info on the right-hand side of the name, to give the user an idea how the actor is parametrized without having to open up the dialog with the options.