"Fields" argument

[zahvib]

Hey,
I know this code has been published a while ago, but maybe someone will know how to solve my problem.
I am currently trying to extract some signals from a .dxd file, with python of course, and to put it in a pandas dataframe.
However, the args "midex_sample_rates" and "fields" from the read_dws function are kind of buggy. When I set mixed_sample_rates as True and fields as [] (to get all the fields), the result is only the four fields that have the same sampling rate (the largest one). And when I try to specify some fields in the blank list, Python tells me "not all fields read" (cf. line 439 in the read_dws function), but I dont get why its doing that. If someone has any idea on how to fix my problems, let me know !
Best,
Antoine

[JanKalin]

I've been updating SW on the local machine, but have not been uploading it to this repo. Try the latest version that has just been uploaded.

[zahvib]

Hi,
thanks for your answer, but I still cannot make it work. I am trying to read a dxd file, with some differents sample rates signals. However, when I am calling the function with an empty list (to read all the fields) and mixed_sample_rates=True, the function returns a pd dataframe with only the signals with the highest sample rate. Do you maybe know why it is doing that ?
I am trying to understand your code but I am kind of new with Python
Best,
Antoine

[zahvib]

Okay so I think I found the issue, it is at line 378, when you define srdiv.
As my sample rates are differents for my signals, and as those sample rates are not fully divisible, the value returned for the srdiv of those signals is nan.
And then when data.iloc[::srdiv[i] is called line 462, it doesnt work with those signals because of the nan value.
I am currently thinking on how to modify this to allow not perfectly divisible sample rates signals to be used as well, I will keep you updated.
Best,
Antoine

[JanKalin]

Ah, yes. It says so in line https://github.com/JanKalin/python_dewesoft/blob/master/dewesoft/DWDataReader.py#L259

What should work is to change the code at line https://github.com/JanKalin/python_dewesoft/blob/master/dewesoft/DWDataReader.py#L377 so that max_sample_cnt is not the maximum of individual counts but the least common multiple of individual counts.

I just did that, but I don't have data to test this. So please try it out and report back.

Cheers,
Jan

[JanKalin]

I was too hasty... There's some more work to be done. In the current code the assumption is that at least one (the "fastest") channel has all the values and timestamps defined. The timestamps of this channel define the final timestamps of the data. Now you would need to synthesise intermediate timestamps around line 414.

If you have some time, as in around a week or two, and provide me with a sample of data that you're trying to read, I should be able to fix the code so that it works

In any case, I've pushed the first necessary change to a difference branch and reverted the master branch.

[zahvib]

Hi, so I have been working on this issue. I made a "home made" version of your function where I just put the signals names and the signals themselves in a python dictionary, so I don't have problems with different size of signals and different sample rates. Let me know if you're interested in looking over the code, and thank you very much for your help.
Best,
Antoine