Skip to content

Latest commit

 

History

History
47 lines (39 loc) · 2.36 KB

CHANGELOG.md

File metadata and controls

47 lines (39 loc) · 2.36 KB

17-Dec-2023: Pat Welch, [email protected]

  • Processing for CASPER East and West and gluing up and down profiles together

8-Dec-2023: Pat Welch, [email protected]

  • Restore path in examples

1-Dec-2023: Pat Welch, [email protected]

  • Point and click dissipation inspector

28-Nov-2023: Pat Welch, [email protected]

  • Fix timezone issue with date_created... for NetCDF files

27-Nov-2023: Pat Welch, [email protected]

  • Implement bottom crash detection for downcast VMPs. This was tested using data from 15-40m casts into the bottom of the Gulf of Mexico off Lousiana/Texas. i.e. a nice soft bottom.
  • Implement bottom trimming and bottom upwards dissipation estimates for VMP downcasts with bottom detection.
  • Implement binning by elevation above the bottom.

17-Nov-2023: Pat Welch, [email protected]

  • Split up profile and dissipation calculations.
  • Add support for binning profiles by time for use in analyzing AUV data.
  • Documenation updates.

19-Oct-2023: Pat Welch, [email protected]

Refactor the code to handle the following:

  • Merged P files are kept in the output_root directory in a merged_p_files directory. If the user changes theh p_file_merge setting from true to false, the user won't be using a merged P file, like before.
  • Changed from using Matlab's hashKey function to SHA-1. The hashKey had a random seed, so between Matlab sessions, the generated hash of a configuration was not stable. With SHA-1 it now is stable.
  • Changed the output directory filenaming from a prefix plus the hash key of the configuration, to a sequential numbering as the configuration changes. There is now a JSON file, with the configuration parameters for that section in it.
  • Renamed process_VMP_files to process_P_files to reflect it is dealing with both MR and VMP P files.
  • Changed processing from all the odas_p2mat operations, then generating all the profiles, then binning, then to combining everything, into a pipeline for each P file. This reduced the amount of disk reads and decreases the execution time by ~20%.
  • Changed many of the operations to be parallelized using parfor. This reduces the execution time by roughly 50% of the number of cores. i.e. a 10 core system reduces the execution time to roughly 20% of the single process execution time.
  • Changed all the example scripts for the refactored code.