Child pages
  • IM Calculation
Skip to end of metadata
Go to start of metadata

Please read the readme @ for instructions on how to run the code.


  • IM calculations have been separated from the "old post-processing" repository, extracting the relevant functions and classes.
  • IM values validated on Hypocentre and Kupe against "old post-processing" on the same data.
  • Two types of workflows: text based (most likely observations) and binary based (simulations once the binary workflow is in place).
    • If binary workflow does not happen soon, the text based option will be used in both cases.
  • Outputs match the formats requested in File Formats Used On GM and should therefore be usable on the upcoming Non-ergodic codes.
  • Tested on very simple multi-process on Kupe with good speed-up using 40 and 80 cores. For the sample, 2228 stations were used



With command : python ../BB.bin b -o /home/yzh231/ -i Albury_666_999 -r Albury -t s -v 18p3 -n 112A -m PGV pSA -p 0.02 0.03 -e -c geom -np 2

  • input file path: ../BB.bin
  • b: input file type is binary
  • -o: output result csvs location is /home/yzh231, default is /home/$user
  • -i: unique identifier/runname of the simrun and output folder name are Albury_666_999, default is 'all_station_ims'. This attribute will be stored in the meta data file.
  • -r: rupture name is Albury, default is unknown. This attribute will be stored in the meta data file.
  • -t: type of simrun is  simulated, default is unknown. This attribute will be stored in the meta data file.
  • -v: version of simrun is v18p3, default is XXpY. This attribute will be stored in the meta data file.
  • -n: station names used to perform im claculation are 112A, default is all the stations in the binary file
  • -m: measures used to perform im calculation are PGV  and pSA, default is all the measures
  • -p: period of pSA used to perform im calculation are 0.02 0.03, default is Karim's 15 periods
  • -e: In addition to the period specified by -p option, use extended 100 period of pSA, default not using
  • -c: component of waveform acceleration used to perform im calculation is geom, default is '090, 000, ver'
  • -np: number of processors used to perform im calculation is 2, default is 2


The result is outputted to the following location, where:

  • 'Albury_666_999' is the folder that contains all outputs. The folder name 'Albury_666_999' is made of the string specified by the '-i' argument. Default is 'all_station_ims' if  not specified.
  • 'Albury_666_999.csv' is the summary csv file that contains all stations' im calculations. The summary file name is made of the string specified by the '-i' argument.
  • '' is the meta data file. The meta data file name is made of the string specified by the '-i' argument.
  • 'station' is the folder that contains all individual station's im_calculations. The folder name is defaulted and cannnot be specified by the user.
  • '112A_geom.csv' is the individual csv file that contains geom component im calculation for station 112A. Each name of the individual station csv file name is made of  station_name + component 


All the steps below are to be carried out in hypocentre

1.Generate summary benchmark:

The following steps should only be performed once for each selected binary file

  1. Select a source binary file: /nesi/transit/nesi00213/RunFolder/daniel.lagrava/Kelly_VMSI_Kelly-h0p4_EMODv3p0p4_180531/BB/Cant1D_v2-midQ_leer_hfnp2mm+_rvf0p8_sd50_k0p045/Kelly_HYP01-03_S1244/Acc/BB_with_siteamp.bin
  2. Identify corresponding databse for the selected source binary file: /home/nesi00213/RunFolder/wdl16/database_old_pp/database.db
  3. Find the script to extract benchmark im value files from the database in step 2: /nesi/projects/nesi00213/dev/impp_datasets/extract_ims.sql
  4. Create a folder to store benchmark files. eg benchmark_im_sims
  5. Execute extract_ims.sql in database.db 4 times with specified components. eg: 'ver'
  6. Export results to benchmark_im_sims/benchmark_im_sim_ver.csv. Clik OK and don't change anything when 'Export data as csv' window prompts

  7. Repeat step 4 and 5 with different components: '090', '000', 'geom'
  8. Now you have 4 summary benchmark files benchmark_im_sim_090/000/ver/geom.csv

2.Generate test input files

  1. Follow the instruction in Binary Workflow FAQ, we can generate single waveform files. These waveforms are intended for the testing of ascii functionality of Open a python cell

    Now we have all the waveforms. 

3. Create Test Folder

  1. Create The test folder structure follows Testing Standards for ucgmsim Git repositories
  2. Select 10 stations you want to test and cp corresponding waveforms files to the singel_files directory as below
  3. Copy the source binary file 'BB_with_siteamp.bin' to the input folder
  4. Run 'write_benchmark_csv(sample_bench_path)' function inside to generate 'new_im_sim_benchmark.csv', where 'sample_bench_path' is the folder we created in 1.4 Generate summary_benchmark: benchmark_im_sims. This function should only be run once for each binary file.

NOW you have all the input files ready

4. Run Pytest

Make sure you are currently under the test_calculate_ims folder, run:  




Checkpointing is needed for IM_calculation due to large job size and limited running time on Kupe. Therefore, we implemented checkpointing to track the current progress of an im_calculation job, and carry on from where the job was interrupted by slurm.

To run checkpointing, first git clone the IM_calculaiton repo, then check out the checkpoint branch

Open im_calc_sl.template, change the IMPATH variable (line 22) to where you have cloned the git repository

Note, the checkpointing code relies on the input/output directory structure specified in the im_calc_al.template in the checkpoint branch. Failure to match the dir structure will result in runtime error. A quick fix would be modifying the template to suit your own dir structure.


(1) Simulation

Input/output structure defined in im_calc_al.template

Actual input data structure:

The input binary file is under:

The output IM_calc folder is under:

(2) Observed

Input/output structure defined in im_calc_al.template

Actual input data structure:

The output IM_calc folder is under:

Splitting a big slurm

Splitting a big slurm script into several smaller slurms is needed due to the maximum number of lines allowed in a slurm script on Kupe.

Still in the checkpointing branch, Run the script that uses both checkpointing and splitting. The -ml argument specifies the maximum number of lines of python call to Header and footer like  '#SBATCH --time=15:30:00', 'date' etc are NOT included.

Say if the max number of lines allowed in a slurm script is 1000, and your (header + footer) is 30 lines, then the number n that you pass to -ml should be 0 < n <=967. eg. -ml 967.


We have 250 simulation dirs to run, by specifying -ml 100 (100 python calls to per slurm script), we expect 3 sim slurm scripts to be outputted.(1-100, 100-200,  200-250)

We have 3 observed dirs to run, by specifying -ml 100 (100 python calls to per slurm script), we expect 1 sim slurm scripts to be outputted.

We have 61 rrup files to run, by specifying -ml 100 (100 python calls to per slurm script), we expect 1 sim slurm scripts to be outputted.

Command to run checkpointing and splitting:


To submit the slurm script:

The reason that we have to run '' under  '/nesi/nobackup/nesi00213/tmp/auto_preproc' is otherwise slurm cannot find machine.env specified by the script:


  • Creation of semi-automatic slurm generation that will have all the calls to produce the results as needed.
  • Progress printing statements
  • Rrup calculation on a smaller station list - currently when generating the slurm script it does the full grid even for stations outside the domain


  • Extensive re-writing of code needs to have smaller deliverables in the future, as this simplifies the integration.
  • No labels