Streaming

Task Streaming

Data streaming task is defined at embci.apps.streaming. It provides a pylsl data stream available on this machine. Multiple clients can connect to this stream and fetch realtime data from it simultaneously. The default data input of this stream is ESP32 MCU on EmBCI Hardware (embedded system), while fake data generator is used on PC platform. Because each clients may need to apply different processing algorithums on their fetched data, such as denoising or component analysis. This stream provides raw data without any pre-processing from amplifier and ADC.

Communication

Two interfaces are available for communicating with this task:

• JSON-RPC

• ZMQ + python-argparse

JSON Remote Procedure Call (JSON-RPC) is an excellent protocal for interprocess communication both locally and through network. However, the second method is more suggested. In this task ZMQ and argparse are used cooperatively to construct a command-recieve-and-execute server. Comparing to JSON-RPC, ZMQ also support local or remote TCP request (and many more mechanisms like PGM, IPC and In-proc). Argparse can provide a neat help output and error handle interface, as well as support to subcommands.

ZMQ is preferred if you want to interact with this task from C++/Python/Java or even command line. JSON-RPC is suitable for web users in case someone want to control the task from webpage using JavaScript.

ZMQ interface is listening on tcp://0.0.0.0:9997, from wich users can set parameters of data stream at runtime. RPC server binds to http://0.0.0.0:9996.

ZMQ Example

>>> import zmq
>>> c = zmq.Context()
>>> q = c.socket(zmq.REQ)
>>> q.connect('tcp://0.0.0.0:9997')
>>> while 1:
...     q.send(raw_input('console@E01:$ '))
...     print(q.recv())
console@E01:$ bias_output
True
console@E01:$ bias_output False # Choose one from ON|off|False|true|1|0
console@E01:$ bias_output
False

Or you can use function streaming.send_message_streaming(), which accepts a string as command and return a string as result:

>>> from embci.apps.streaming import send_message_streaming
>>> while 1:
...     print(send_message_streaming(raw_input('$ ')))
$ summary
Status:
sample_rate:    499/500 Hz
bias_output:    enabled
input_source:   normal
stream_control: paused
impedance:      disabled
$ stream_control resume

RPC Example

You can use RPC service in JavaScript like:

var request = {
    method: 'system.describe',  // cascaded function name
    params: ['all'],            // parameters are optional
    id: '0xdeadbeef',           // id may be omitted
    jsonrpc: 2.0,               // see JSON-RPC Specification for version
};
jQuery.ajax({
    url: 'http://0.0.0.0:9996',
    method: 'POST',                 // or type: 'POST'
    crossDomain: true,              // enable CORS ajax request
    data: JSON.stringify(request),  // must pass data in string
    dataType: 'json',               // response format
    success: obj => console.log('Result', obj.result || obj.error),
    error: (xhr, status, error) => console.error(error)
});

In Python, you can use client class provided by embci.utils.jsonrpc.

>>> from embci.utils import jsonrpc
>>> client = jsonrpc.JSONRPCClient('http://0.0.0.0:9996')
>>> client.stream_control('close')
>>> client.measure_impedance('true')
>>> print(client.summary())
Status:
sample_rate:    500/500 Hz
bias_output:    enabled
input_source:   normal
stream_control: closed
impedance:      enabled
>>> print(jsonrpc.__doc__)  # for more help message on JSON-RPC