900820_MOOS_Upper_Canyon_Experiment : About Vertical Profiler Software

This is the short-form explanation of the status of Vertical Profiler data processing software. If there are any questions you are welcome to contact me.

SeaHorse Vertical Profiler

In 2007 I wrote software for the SeaHorse, based on some rough Matlab code by Andy Hamilton, that did the following:

This software produced some plots that John Ryan used to consider the data from the initial SeaHorse deployment, but never was deployed in a real-time operational scenario, although that was the original goal. (To the best of my knowledge the SeaHorse VP has not been redeployed since then.)

Work for this project was saved in the CVS repository, in DPforSSDS/mse/profilers/seahorse/batchscripts. (Auxiliary files folder in there also has the useful configuration files for the SeaBird software.) Quoting from the README file therein,

These scripts are simply a quick first version that does not copy any data or metadata
back to SSDS.  Instead data and images are copied to the ssdsdata directory, in a temporary
location, for use until the permanent software is released.

To execute the scripts, run PlotVPData from Matlab.

You will need the following local files and directories along with these scripts:

For more information, see that README file, which leads to the scripts that are used.

If you are eager to try to run with the original development environment, it is on MBARI 877, SeaFarer, a PC being left in my office. It is deployed under SeaHorseProcessing\ in my Documents\ directory on that machine.

Original Priorities

Incidentally, the following answers to questions when I worked on that project may be a useful starting point for someone else.

Who are the customers for this data?
 Francisco, John Ryan 
 

Priorities:
 A) 1d: Get an interim 'production' system putting SeaHorse information up on some web pages.
 B) Replicate MMP processing.
 C) Interim production system for MMP.
 D) Work on SSDS-based processing, submission, and plotting.
 
> This order is good.

How important will it be to regularly re-process the raw data that's in a HEX file?
> Probably not so much. More important to have the cnv files.

What data is worth moving forward into analyzed/binned data?
 - nominal bin depth
 - median value
 - mean depth of samples
 - mean time of samples
 - stdev of values of samples
 - time range
 - bin #
 - # samples
 
How can we identify the actual sample time for each sample (interval=seconds:0.25)?
> yes, use that information.

What other data and metadata is in the .HEX file that we haven't captured?
> No telling, go take a look once this other stuff is done.
(Note: There is some useful stuff there, it's worth looking at this 
if you're going to rev up a service.--JBG)

McLane Mooring Profiler

In 2009 I was asked to write software to post-process the McLane Mooring Profiler data. Toward this end, I collected all the information that Andy Hamilton had. He told me he had run the software only once, and never automatically.

Automating post-processing

To run in a real-time operational scenario, we had to either use the provided 3rd-party software (which only ran on a PC), or convert the data ourselves. Favoring the latter option, I contacted the vendor, who provided 2 attachments under the following emails. I have put those attachments into CVS under DPforSSDS/mse/mmp/.

Here's some additional descriptive information on the MMP files of interest. This, together with the structures sent last week, should do the trick. (Note: the TT8 uses Big Endian ordering.)

Here are the structure definitions for MMPX3_14. Hope these prove helpful
for you.

I looked at these only long enough to see that they looked non-trivial. (Actually, now that I just looked again, they aren't so bad – just the Engineering data has a variable format. And we have the C headers.)

Necessary software

Andy pointed me to, or delivered to me on CD actually, a collection of software that he had accumulated. The README file for the collection says:

This directory contains:

- the raw binary data retrieved from the MMP (Data\MMPFileSystemImage)
- the unpacked and processed data (Data\UnpackedData and Data\ProcessedData)
- the McLane tool used to convert the binary data to ascii files 
  (McLaneSoftware\3_04Unpacker\UnPackerV304.exe)
- the matlab tools used to process the data (DataProcTools)
- some documentation for all of this (Doc)

The software includes both manufacturer and science analysis software (the latter has a how_to.html file that looks pretty helpful at explaining things).

I have uploaded a copy of this software to the old 600027.MOOS.Mooring folder on the old ProjectLibrary; it's in "MMP Profiler Store". (It's a 0.4GB Zip file, apparently too big for Alfresco to upload.) To start you'll probably want to grab this file and unzip it and start poking around.