search

OxWearables / biobankAccelerometerAnalysis

Extracting meaningful health information from large accelerometer datasets
https://biobankaccanalysis.readthedocs.io/en/latest/index.html
Other
179 stars 61 forks source link

can't accProcess .csv directly #258

Closed tbonewmy closed 1 week ago

tbonewmy commented 4 weeks ago

Hi there, I tried to use accProcess to process csv file directly as this link suggested https://biobankaccanalysis.readthedocs.io/en/latest/usage.html#basic-usage. However, when I give either a csv file or a csv.gz file, the call() function in device.py give a non-zero exit code. The process then fail and exist.

the command line I use is (myenv) PS C:\Users\my> accProcess SomeDiractory/sample-small.csv.gz the csv data is downloaded from the place mentioned in above link in the format of time,x,y,z. My machine is windows. Python 3.9 accelerometer 7.1.2

I feel like it's the csv format doesn't compatible with the java code, since everything seems normal until call the java functions. Please help me.

chanshing commented 4 weeks ago

Hi @tbonewmy Can you try unzipping it first?

tbonewmy commented 4 weeks ago

Hi @tbonewmy Can you try unzipping it first?

Yes, I tried both zipped and upzipped files. sample-small.csv.gz and sample-small.csv. Neither of them works. Below is the error message from the command line end. image

And as I am writing, I think I found the cause to my problem. I tried it. It works. The issue is the example file only have time,x,y,z columns. But the default csvTimeXYZTempColsIndex parameter has values 0,1,2,3,4. So the indices don't match. I gave it 0,1,2,3 and problem solved.

chanshing commented 4 weeks ago

Thanks @tbonewmy I'll fix the documentation soon