Python framework that provides the basic building blocks to implement data processing graphs.
Hydra is a framework (actually, currently more of a collection of classes) that provides the basic building blocks to implement data processing software as a graph of nodes while making heavy use of the publisher/subscriber design pattern.
There are four kinds of nodes in each graph:
Sources: Nodes which are only publishing data items to subscribing nodes in the graph (e.g. file readers, sockets to read data from, REST endpoints etc).
Sinks: Nodes that are only consuming data from publishing nodes. These are the end-points of the graph (e.g. loggers, display widgets, sockets used to send data, REST endpoints, websockets etc).
SourceSinks: Nodes that are both consuming and producing data and are graph end-points at the same time (e.g. sockets used to send and receive data).
Intermediate nodes: Nodes which are both consuming (subscribers) and publishing data at the same time. Such nodes can be used for various purposes such as buffering, examining, transforming, filtering data etc.
The available classes are broken down to the following modules:
hydra_core: Contains the basic classes from which graphs are built.hydra_common: Classes for commonly used node types (e.g. queues).hydra_net: TCP and UDP SourceSinks.hydra_rest: REST API endpoint node (uses Flask and its built-in
development server).Install with pip:
pip install hydra-framework
See example.py.
hydra_websocket: Module that provides REST and websocket sinks and sources.hydra_gui_qt5: Module that provides GUI components for QT5 (Python for QT).requests module.MIT license