signal-pipes 0.1.16

Stream operation for processing of (physiological) signals

Signal pipes (aka sigpipes) are Python library for processing of digital representation of signals.

The library is focused on processing physiological signal (e.g. EKG or EMG) but is basic design supports processing of any multichannel signals.

The library is based on two mechanism:

1. flexible containers for signals and their metadata in the form of hierarchical dictionary with a set of "well known" keys forming solid skeleton of data representation.

2. pipelines that are formed by operators for processing, storing and visualization og signals. The pipelines are not limited to simple linear chains. The more complex forms are supported (branching, alternatives, parallel processing).

Current version support these tools in form of pipeline operators.

• readers of physionet.org databases
• readers of matlab files from Megawin application
• exporter to pandas dataframes
• configurable plotting using matplotlib
• selection of channels or subsamples
• extraction of signal features
• signal adjustment (any unary ufunc, convolution, cross correlation, polyphase resampling) based on scipy.signals and numpy
• serialization to HDF5