Solution to the problem that scikit-sparse required for GARPOS does not fit into Apple Silicon (M1, etc.) macs even if the dependent package is installed as per HP.

• Problem with installation

Find the correct path to the library as follows;

1. Open Finder.
2. Press Cmd + Shift + G.
3. Type /opt/homebrew/Cellar/suite-sparse and press Enter.
4. Find the directory, as shown in the attached screenshots.

Now, you can pip install as follows, specifying the correct path to the library. Ensure the paths match.
SUITESPARSE_INCLUDE_DIR=/opt/homebrew/Cellar/suite-sparse/7.7.0/include/suitesparse SUITESPARSE_LIBRARY_DIR=/opt/homebrew/Cellar/ suite-sparse/7.7.0/lib pip install scikit-sparse

• Problem with import
  In my Anaconda environment, I was able to install, but an error occurred on import. I was able to deactivate Anaconda and install and import successfully with another python.

The 2nd derivative matrix of the B-spline basis should be (delta.T @ w^(-1)) @ delta, not (delta.T @ w) @ delta. The formula can be fond in the book "Nonparametric Regression and Generalized Linear Models A roughness penalty approach" by P.J. Green, Bernard. W. Silverman, page 25

Dear mainteners,

I have tried to run GARPOS with estimating seafloor coordinates only (no gradient parameters) by setting the parameter inversiontype of the config file to 0.
However, this returns an error message from setup_model.py due to a division by zero.

I managed to make a quick and dirty fix that allows to make the program work. But further modification should be made for consistency.

Best regards,

EL

Modification :

mp_estimation.py
if invtyp == 0:
#knotint0 = 0.
#knotint1 = 0.
#knotint2 = 0.
pass

Error message :

suffix lamb0 lgrad mu_t mu_m
0 0.1 0.1 0.0 0.5
bin/garpos_v101/setup_model.py:135: RuntimeWarning: divide by zero encountered in scalar divide
nknots = [ int(obsdur/knint) for knint in knotintervals ]
multiprocessing.pool.RemoteTraceback:
"""
Traceback (most recent call last):
File "multiprocessing/pool.py", line 125, in worker
result = (True, func(*args, **kwds))
File "multiprocessing/pool.py", line 51, in starmapstar
return list(itertools.starmap(args[0], args[1]))
File "bin/garpos_v101/mp_estimation.py", line 157, in MPestimate
knots = make_knots(shots, spdeg, knotintervals)
File "bin/garpos_v101/setup_model.py", line 135, in make_knots
nknots = [ int(obsdur/knint) for knint in knotintervals ]
File "bin/garpos_v101/setup_model.py", line 135, in
nknots = [ int(obsdur/knint) for knint in knotintervals ]
OverflowError: cannot convert float infinity to integer
"""

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "bin//solveSingleEpoch.py", line 47, in
rf = drive_garpos(options.cfgfile, options.invcfg, odir, options.suf, mc)
File "bin/garpos_v101/garpos_main.py", line 169, in drive_garpos
outputs = parallelrun(inputs, maxcore)
File "bin/garpos_v101/garpos_main.py", line 65, in parallelrun
reslist = p.starmap(MPestimate, inp)
File "multiprocessing/pool.py", line 375, in starmap
return self._map_async(func, iterable, starmapstar, chunksize).get()
File "multiprocessing/pool.py", line 774, in get
raise self._value
OverflowError: cannot convert float infinity to integer

I am using data from two different surveys at one site. The initial procedure works well using Log_Lambda0 = -1 and mu_t = 0.0. For the fix step, higher mu_t values (in this case 0.5) leads to a program failure due too much memory usage. I have tracked this down to the data correlation step, specifically with lu.solve in line 263 of setup_model.py. I am uncertain if there is a more efficient/less memory intensive way of addressing this issue. Here is the information for matrix E:

<66864x66864 sparse matrix of type '<class 'numpy.float64'>'
with 4331926 stored elements in List of Lists format>

Currently, my RAM is limited to 64 gb. I would be happy to pack the data into a zip file and provide it through Google Drive, if you would like to use it for a test case. I understand if this an impossible issue at the moment, and I greatly appreciate your time.

Dear author, hello. I am a student who has just started learning about underwater positioning. Currently, I am not sure about the specific purpose of this software, how to use the files it generates, and what the software does. I would like to ask if you have any relevant learning documents and how to use this software.