Data Modification with Multiple Labels

[joshuatbland]

Hello everyone! I am new to crappy and I am trying to learn more about on-the-fly data modification capabilities. It looks like modifiers are the best way to do real time calculations with single data labels. What would be the best way to do on-the-fly calculations with multiple labels (like taking the average or mean of several labels at every time point)? Any insight into this would be greatly appreciated. Thanks.

[WeisLeDocto]

Hi !

It all depends if the labels come from a same Block, or from different ones.

• If they come from a same Block, it is possible to perform any operation on all the labels at once. Basic example:

  class MyModifier(crappy.modifier.Modifier):
      def __call__(self, data):
          data['avg'] = 0.5 * (data['label_1'] + data['label_2'])
          return data

• If they come from different Blocks, things get much trickier and there is no universal answer. What's for sure is that a Modifier alone cannot do the job, a Block needs to be used. In the general case, the labels will not be generated in a synchronized way by the several Blocks, so it's impossible to define operations to compute at every time point. One can define time-based operations (e.g. average all input labels together over a given period of time), but they won't run at every time point.

  That being said, Crappy comes with two Blocks that can synchronize signals from several input Blocks into a single output on a common time base. This output can then be processed by a regular Modifier, as described in my first point. The first Block is the MeanBlock, that averages the received labels over a given delay (1 average per label), and returns all the averages. It's recommended when a low data rate is acceptable (<10 SPS). The second Block is the Multiplexer, that performs linear interpolation on the received labels to output data at a target frequency. It's indicated when a higher data rate is needed.

  Note that with both options the original data is lost when averaging or interpolating. You'll find examples for both Blocks in the examples folder on GitHub.

[joshuatbland]

Ok, great - thank you for the response. This was very helpful!