APIStreamer

APIStreamer is a class that reads directly from buffers on the hardware without waiting for it to go to the tank. Because it is essentially competing with Synapse for access to the hardware, it has lower bandwidth than SynapseLive, but it has much faster access.

You can download the latest MATLAB SDK files here. The MATLAB APIStreamer class installs into:

C:\TDT\TDTMatlabSDK\TDTSDK\APIStreamer

Examples install into:

C:\TDT\TDTMatlabSDK\Examples\OnlineAnalysis

The Python APIStreamer installs with pypi:

pip install tdt

Release v96 of Synapse and above includes the custom gizmos that APIStreamer can read from, in the Gizmo List under Custom > TDT > API.

1. Attach one of the APIStreamer gizmos (e.g. APIStreamMC for multi-channel signals) to your data source.

2. Use one of the examples to read from the buffer inside the APIStream gizmo during the recording.

The data rate limit for APIStreamer is ~8 channels of 32-bit floats at ~25 kHz or the equivalent data rate (e.g. 32 channels at ~6 kHz). Beyond that we recommend a PO8e Streaming Interface card, which is a direct connection from the RZ to a separate PCI card, bypassing Synapse altogether.

Matlab Examples

	Online Streaming Data Read a stream of multi-channel data directly into Matlab using APIStreamer
	Online Spectrum Analyzer Read a single channel and plot the power spectrum over time Good for visualization and online analysis

Python Examples

Input Parameter	Description
host	IP Address of Synapse computer (default 'localhost')
gizmo	Gizmo name in Synapse (default 'APIStreamerMC1')
history_seconds	How much data to keep in Python memory (default 1 second)
callback	Function handle to run every time new data is collected
verbose	Print debugging information to the console (default True)

Online Data Processing

APIStreamer is collecting data in the background. Reads the data buffer every 1 second.

import time
import tdt

s = tdt.APIStreamer(gizmo='APIStreamerMC1', history_seconds=10)

while 1:
    print(s.get_data())
    time.sleep(1)

Online Data Processing Callback

Runs the do_something function every time new data is read from the hardware

import tdt

def do_something(result):
    print('process data', result.data.shape)

s = tdt.APIStreamer(gizmo='APIStreamerMC1', history_seconds=10, callback=do_something)

Online Streaming Data

Plots the data from the APIStreamer

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.animation as animation

import tdt

s = tdt.APIStreamer(gizmo='APIStreamerMC1', history_seconds=10)

fig = plt.figure(figsize=(10,5))
ax1 = fig.add_subplot(111)

def animate(i):

    result = s.get_data()
    if result is None:
        return

    xs, ys = result
    if len(xs) == len(ys) == 1:
        return

    ax1.clear()
    ax1.plot(xs, ys.T + np.arange(ys.shape[0]).T)
    ax1.set_xlabel('time, s')
    ax1.set_ylabel('V')

ani = animation.FuncAnimation(fig, animate, interval=100, cache_frame_data=False)

plt.show()

Plotting Multiple Streams

Reads from two APIStreamer gizmos and plots them

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.animation as animation

import tdt

s = tdt.APIStreamer(gizmo='APIStreamer1Ch1', history_seconds=10)
s2 = tdt.APIStreamer(gizmo='APIStreamer2Ch1', history_seconds=10)

fig = plt.figure(figsize=(10,5))
ax1 = fig.add_subplot(211)
ax2 = fig.add_subplot(212)

def animate(i):

    result = s.get_data()
    if result is None:
        return

    result2 = s2.get_data()
    if result2 is None:
        return

    xs, ys = result
    ax1.clear()
    ax1.plot(xs, ys.T + 1*np.arange(ys.shape[0]).T)
    ax1.set_xlabel('time, s')
    ax1.set_ylabel('V')

    xs2, ys2 = result2
    ax2.clear()
    ax2.plot(xs2, ys2.T + 1*np.arange(ys2.shape[0]).T)
    ax2.set_xlabel('time, s')
    ax2.set_ylabel('V')

ani = animation.FuncAnimation(fig, animate, interval=100, cache_frame_data=False)
plt.show()