sqrt Runtimewarning error

[davidh-]

Hi,

I have the code working with my T3 Search. Just fixing errors that are popping up and one of them is the following:

/home/pi/ht301_hacklib/ht301_hacklib.py:94: RuntimeWarning: invalid value encountered in sqrt
np_Ttot = np_v8**0.5 - l_flt_1000337C - ABSOLUTE_ZERO_CELSIUS

I added some print statements to see what's the issue and found np_v8 is a numpy array instantiated to have negative values for the first ~16384 entries. The sqrt throws an error when it encounters these. Would you agree is is fine to set all of the negative values in the numpy array to 0? Or do you have a better fix for this?

testing device nr: 0
width: 384.0 height: 292.0
width: 384.0 height: 292.0
device_info: ['2.01.220420', '', '', '', '', 'LF1452']
m3: [  0   0   0   0   0   0 200  65   0   0 200  65 102 102 230  62  72 225
 122  63   1   0 128  63   0   0   0  65   0   0]
Fix_ 0.0
refltmp_ 25.0
airtmp_ 25.0
Humi_ 0.44999998807907104
Emiss_ 0.9800000190734863
Distance_ 1

v5 6000
coretmp_ 36.35000000000002 3095

flt_10003360 0.20020000636577606
flt_1000335C 26.92799949645996
flt_1000339C -5.999999848427251e-05
flt_10003398 0.008899999782443047
flt_10003394 0.670199990272522
water vapour content coefficient: 10.296338751668296
transmittance of atmosphere:      0.993438244014494

np_v8: [-16370.89846022 -16366.28886379 -16361.67926736 ...  59138.90062188
  59143.51021831  59148.11981474] np_v5: [-5880. -5879. -5878. ... 10501. 10502. 10503.] v22: 0.9228412341685726 v23: 1243.3615546075637 flt_10003360 0.20020000636577606 l_flt_1000337C_2: 4522.931556256735
/home/pi/ht301_hacklib/ht301_hacklib.py:94: RuntimeWarning: invalid value encountered in sqrt
  np_Ttot = np_v8**0.5 - l_flt_1000337C - ABSOLUTE_ZERO_CELSIUS
[         nan          nan          nan ... 449.08216649 449.09164386
 449.10112086]

cx: 6000 v2: 120
v5: -5880
flt_1000339C -5.999999848427251e-05 flt_10003398 0.008899999782443047 flt_10003394 0.670199990272522 fpatmp_ 38.25
v22: 0.9228412341685726
v23: 1243.3615546075637
np1: [nan, nan, nan, nan, nan, nan, nan, nan, nan, nan]

[stawel]

nice finding!

Would you agree is is fine to set all of the negative values in the numpy array to 0?

yes, adding:

np_v8[np_v8 < 0] = 0 

definitely won't hurt.

Or do you have a better fix for this?

unfortunately not, for now at least