Re: [Health-dev] MyGNUHealth: Import recordings - first test results

[Luis Falcon]

Hi, Mat!

Awesome! Very happy about the results. That is a nice dataset.
The awkward looking chart I believe comes from sending the non-sorted
data to matplotlib, or repeating the import. 

I will work on this in the coming days. 

Once we have this sort out no matter how many times we
accidentally import the results, the plot should be the same.

Enjoy vacation!
Luis

> Mat Knoll commented:
> 
> 
> Hi Luis,
> Sorry for the delay......
> 
> I updated the app and copied the modified script.
> 
> Now my first test results:
> - import of 1532 rows went smoothly! Great!!
> - typos in the csv caused syntax errors (as expected)
> - the last row intentionally did not contain any values, just
> delimiters (","). The script recognized an error in date and skipped
> the line
> 
> The chart, however, looks strange. I'll have to take another closer
> look. It now starts at the year 1900....however, the csv doesn't
> contain any values earlier then 2020..... The imgur link to the
> screenshot: https://imgur.com/OD0uH80
> 
> After my vacation i will test some special cases and report again.
> 
> Preliminary conclusion: The import script works correctly!
> 
> Best,
> Mat
> 



On Wed, 28 Jul 2021 14:45:13 +0200
Mat Knoll <research@aeyon.de> wrote:

> Hi Luis,
> 
> great! Thanks! I'll report.
> 
> Best,
> Mat
> 
> On 28.07.21 14:13, Luis Falcon wrote:
> > Hi, Mat
> >
> > Some updates...
> >
> > On Wed, 21 Jul 2021 13:36:19 +0200
> > Mat Knoll <research@aeyon.de> wrote:  
> >> - When (accidentally) inserting an empty row (even at the end of
> >> the csv) an error occurred:
> >> Traceback (most recent call last):
> >> line 100, in parse_values
> >>       vdate = line[0]
> >> IndexError: list index out of range  
> > I think now we should handle both scenarios, a true empty line
> > (which triggered the traceback you mentioned), and a line full of
> > empty columns (usually from spreadsheets)
> >
> > 2021-03-12,14:10,91,178,115,78,73,99,85,14500,1800,
> > ,,,,,,,,,,, <--- EMPTY row from Libreoffice Calc
> > 2021-03-17,07:10,92.3,178,113,75,75,,,,,
> >
> > 2021-03-12,14:10,91,178,115,78,73,99,85,14500,1800,
> > <-- Empty line
> > 2021-03-17,07:10,92.3,178,113,75,75,,,,,
> >
> > Both scenarios are now handled
> >  
> >>
> >> One more little thing:
> >> Typo found in header of sample_reading.csv: date(yyymmdd) - y is
> >> missing  
> > Fixed.
> >  
> >> Suggestion:
> >> I have a lot of test data that records the values individually per
> >> column (>1500). I know this is a fundamental design aspect and
> >> certainly there are very good reasons for the current design,
> >> however would it be conceivable to separate the two values
> >> systolic and diastolic by comma? Otherwise I need to write a small
> >> macro to merge the two columns using the "/" separator. Possibly
> >> not only my import data is stored in separate columns?  
> > Now systolic and diastolic have their own column
> >  
> >> When a decision is made on how to proceed, I prepare the test data
> >> accordingly and test further cases (invalid values etc.)
> >>
> >> I hope I was able to support. If there are any questions, please do
> >> not hesitate to contact me.  
> > Thank you for the testing and let me know how this second round
> > goes :)
> >
> > You can view / download the latest changesets here:
> >
> > https://hg.savannah.gnu.org/hgweb/health-mygnuhealth/file/2aa85a52e7f0/scripts
> >
> > All the best
> > Luis
> >  
>