Python Driver GPI synchronisation

When I use the python Driver class provided by the pypi library, I can only seem to set the sampling frequency of the current measurement (driver.publish(“h/fs”, fs) which is undocumented it seems). but it doesn’t affect the GPI readings which happen with full resolution (2MS/s). As a results, I have mismatching data (from a 0.5s recording):

u/js220/002627.current:      0.06291326085940004 A | 500000 samples
u/js220/002627.gpi[0]:                 0.199584   | 1000000 samples
u/js220/002627.gpi[1]:                 0.163816   | 1000000 samples
u/js220/002627.gpi[2]:                 0.163817   | 1000000 samples

I am trying to align data in a pandas dataframe using all kinds of metadata (sample_id_range, sample_rate, utc_range) but nothing seems to be in place. It actually looks like the sampling rate in wrongly reported for the GPIs.
What is the best practice to do that?
I am using a jupyter notebook / WSL / Win11 stack, which might cause problems, but the current readings are actually looking very good and I seldomly get sample dropouts.

Hi @MLehmann and welcome to the Joulescope forum!

Yes, the h/fs topic only changes the output sample frequency for current, voltage, and power. The current_range and the general purpose inputs always update at 2 Msps.

However, it is easy to correlate data from a single JS220. Simply use the sample_id field, which always updates at the instrument’s 2 Msps rate. You will notice that the example always divides by decimate_factor which adjusts sample_id to be single incrementing based on the h/fs value.

We have not done a great job independently documenting the available topics. However, the topics and their metadata are available:

python -m pyjoulescope_driver info -v *

I don’t know if the latest commit in the fixed something, but the alignment now works by using sample_id_range.
BTW: in a linux shell, the command for getting topics and metadata must be:
python -m pyjoulescope_driver info -v "*"

The sample_id used in sample_id_range in the example is already divided by the decimate factor:

sample_id = value['sample_id'] // decimate_factor

Using this value will align correctly for signals with the same decimate factor. However, this will not help align current with a GPI0, for example. You can modify the code to also store the original sample_id values, like this:

signal_state['sample_id_raw_range'] = [value['sample_id'], value['sample_id']]
signal_state['sample_id_raw_range'][-1] = value['sample_id'] + sample_count * decimate_factor