Closed parevalo closed 8 years ago
I just realized that the NameError happens only because I was a little confused about the order in which pdb is working. The code IS actually working and the values are being assigned correctly. The least squares code is actually producing correct data, but the second element, which is the one that the code is trying to get, is an empty array: array([], dtype=float64)
.
So the empty array is being produced because, according to the numpy.linalg.lstsq documentation, that's the expected behavior when M <= N. For this case, I'm getting that they are 4 and 4, so the condition is being met. But why is that the case, if my minimum number of observations is set to 6?
Skipping test for two models when either model has n <= k
(number of observations fewer than or equal to the number of parameters estimated).
I'm getting the following error when trying to run yatsm line with commission test enabled:
Location of the log file for that specific line: /projectnb/landsat/projects/Colombia/logs/859/FIT2/y859_1.o6460045 My conda environment: /usr3/graduate/parevalo/miniconda2/envs/conda_env Yatsm location: /projectnb/landsat/users/parevalo/yatsm Script I'm using: /projectnb/landsat/projects/Colombia/workflow/multi_scene/yatsm.sh
The script I'm using is being run for a single line out of 400, because I was trying to debug the code. The config file is in the same folder as that script. My postprocess.py script has been slightly modified to call the debugger once yatsm.px = 2362, because that's when it seems to fail for that specific line I'm running. What seems to be happening is that
m_1
andm_2
(and their corresponding start and end variables) are not being assigned any values (I get errors like*** NameError: name 'm_1' is not defined
). So when it reaches the least squares code, it complains becausem_2_start:m_2_end
is actually a string and not a numeric value. However, while debugging, if I manually type the code and run it, the variables are assigned their correct values.