Open klemenjak opened 4 years ago
Good point @klemenjak
@Rithwikksvr @PMeira Let us use sample_interval
consistently across NILMTK and NILMTK-contrib? Thoughts?
Yes @nipunbatra we need to use sample_interval. Also, we need to specify if user wishes to downsample the data.
(Moving to NILMTK)
Hi,
it's me again. I discovered a "naming conflict" in the source code. Let me explain:
The class API contains a variable sample_period. I guess this variable holds the time between two samples in seconds? A little down the code you fetch params['sample_rate'] from the definition of the experiment. As I looked up in the Jupyter notebooks, sample_rate is supposed to hand over the sample period.
This might be confusing for many people, including myself, because sample period and sample rate are two different things. A period is measured in seconds, whereas a rate gives information on some quantity per second. In our case:
This may result in people understanding that in the following experiment, we have 60 samples per second:
I'm pretty sure that 60 stands for 60 seconds and not for 60 Hertz.
I would suggest replacing sample_period by sample_period or sample interval so that it's clear that users have to provide values in seconds.
Feel free to ignore this in case you disagree. There are many opinions on this matter.
Best, Christoph