TAResearch 2019-03-18
Categories : CHPC, Weather, ML Data Preprocessing,
Re-vectorize data on CHPC
Fixing Run Errors on CHPC
Traceback (most recent call last):
File "fd_pedestal_data_vectorization.py", line 285, in <module>
main()
File "fd_pedestal_data_vectorization.py", line 110, in main
print('Loading and Padding FD Pedestal Data part: {0} {1} ({2}/{3})'.format(ymds, part, i, n_all_parts))
NameError: global name 'n_all_parts' is not defined
Fixed and transferred to kingspeak
. Missing psutils
module in the header of the python script.
Time Out on CHPC
Looks like everything was running fine. Just ran out of time.
SLURM Job_id=6860606 Name=vect_rnn Failed, Run time 03:00:15, TIMEOUT, ExitCode 0
and looking at the log file :
$ tail out/fd_pedestal_data_vectorization.out
Saving Padded Vectorized FD Pedestal Data as Numpy Arrays with shape (35, 32, 96) to /scratch/local/u0949991/Data/fd_ped_vect/y2016m12d04s0p36_ped_fluct_vectorized_padded.npy...
Found 25 Frames for y2016m12d04s0 part 35
Saving Vectorized FD Pedestal Data as Numpy Arrays with shape (25, 32, 96) to /scratch/local/u0949991/Data/fd_ped_vect/y2016m12d04s0p35_ped_fluct_vectorized_padded.npy...
Saving Padded Vectorized FD Pedestal Data as Numpy Arrays with shape (25, 32, 96) to /scratch/local/u0949991/Data/fd_ped_vect/y2016m12d04s0p35_ped_fluct_vectorized_padded.npy...
Found 36 Frames for y2016m12d04s0 part 31
Saving Vectorized FD Pedestal Data as Numpy Arrays with shape (36, 32, 96) to /scratch/local/u0949991/Data/fd_ped_vect/y2016m12d04s0p31_ped_fluct_vectorized_padded.npy...
Saving Padded Vectorized FD Pedestal Data as Numpy Arrays with shape (36, 32, 96) to /scratch/local/u0949991/Data/fd_ped_vect/y2016m12d04s0p31_ped_fluct_vectorized_padded.npy...
Found 24 Frames for y2016m12d04s0 part 30
Saving Vectorized FD Pedestal Data as Numpy Arrays with shape (24, 32, 96) to /scratch/local/u0949991/Data/fd_ped_vect/y2016m12d04s0p30_ped_fluct_vectorized_padded.npy...
Looks to be working fine. Just need to edit the .slm
to use more time in line #SBATCH --time=10:00:00
Update stats.html
Updated post_stats.py
to parse the html in categories.html
to find all categories and find the associated hyper reference, date of entry, and entry title. Then it creates a time indexed dataframe and creates a calendar heatmap of the top 5 calendars and place them in /assets/calendars/
.
Edited stats.html
with post_stats.py
to write the html file for the page to add the top 5 categories calendar heatmpa images to the index file. Works great. I think this is all i really need for my digital journal.