https://github.com/psbrandt/edf-parser

Forked from https://bitbucket.org/yehezkel/edf-parser
https://github.com/psbrandt/edf-parser

edf polysomnography psg

Last synced: about 1 month ago
JSON representation

Forked from https://bitbucket.org/yehezkel/edf-parser

• Host: GitHub
• URL: https://github.com/psbrandt/edf-parser
• Owner: psbrandt
• Created: 2019-12-25T18:18:58.000Z (about 5 years ago)
• Default Branch: master
• Last Pushed: 2023-01-05T03:40:16.000Z (about 2 years ago)
• Last Synced: 2024-04-14T20:04:24.489Z (9 months ago)
• Topics: edf, polysomnography, psg
• Language: JavaScript
• Size: 1.67 MB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 10
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Edf File parser

[![Build Status](https://travis-ci.org/psbrandt/edf-parser.svg?branch=master)](https://travis-ci.org/psbrandt/edf-parser)

Package to parse/process edf file as detailed on the [specification](http://www.edfplus.info/specs/edf.html) 

**Important:** This is not and edf+ parser

## Terminology

Below some terminology found in the source code that will help understanding the implementation faster:

**DataBlock**: The raw byte-encoded data samples as found in the file.

**Record**: The decoded samples found on a data block.

**SignalProcessor**: Interface for objects that can process/translate a raw set of bytes, representing the digital values,into the real physical values of a signal.

**RecordProcessor**: Interface for objects that can process/translate and entire datablock into its record.

## Sample

if there is no time for to much reading file **sample/sample.js** provides a quick usage intro.

## Some Implementation details

**edf.DataBlock**, implementing the stream.Transform interface, should be pipe into a stream reading the edf raw content, it

will parse the header, signals headers, triggering the events header and signals respectivily containing the models, right after

it will begin streaming each datarecord as a block

**edf.RecordsStream**, also implementing the stream.Transform interface, is design to take as input each datablock and translate

it into a record, which gets push out  the pipe (using object mode)for extra processing. the RecordStream does the transformation with the help

of an object implementing the  **edf.RecordProcessor**

**edf.RecordProcessor** object is a composition of multiple **edf.SignalProcessor**, and each signal processor is in charge of

processing/decoding the data of the signal it represents. The responsability of a RecordProcessor is simply coordinating between

each signal processor and keeping a time log to determine the start time of each data record.

The idea behind the composition is that different implementations of the SignalProcessor can be combine together.As sample i see the to-be-implemented DiscardSignal processor which basically discards all signal data from the record (usefull if only a subset of the signals needs to be use) or the logarithmic transformation described [here](http://www.edfplus.info/specs/edffloat.html).

the **edf.utils** namespace contains utility functions that facilitates the interconnection of the different parts of the parser.