failed to calculate eigen rays with newer version of arlpy (ubuntu 20.04)

[patel999jay]

Newer version is failed to calculate ray tracing.

Here is the output of my pip freeze - pip_freeze.txt

Here is the error while computing eigen rays. I had also attached copy of example notebook for reference.

bellhop-Copy1.pdf

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-25-1af3b30f110e> in <module>
----> 1 rays = pm.compute_eigenrays(env)
      2 pm.plot_rays(rays, env=env, width=900)

~/.local/lib/python3.8/site-packages/arlpy/uwapm.py in compute_eigenrays(env, tx_depth_ndx, rx_depth_ndx, rx_range_ndx, model, debug)
    352     if debug:
    353         print('[DEBUG] Model: '+model_name)
--> 354     return model.run(env, eigenrays, debug)
    355 
    356 def compute_rays(env, tx_depth_ndx=0, model=None, debug=False):

~/.local/lib/python3.8/site-packages/arlpy/uwapm.py in run(self, env, task, debug)
    605             else:
    606                 try:
--> 607                     results = taskmap[task][1](fname_base)
    608                 except FileNotFoundError:
    609                     print('[WARN] Bellhop did not generate expected output file')

~/.local/lib/python3.8/site-packages/arlpy/uwapm.py in _load_rays(self, fname_base)
    819                     'ray': [ray]
    820                 }))
--> 821         return _pd.concat(rays)
    822 
    823     def _load_shd(self, fname_base):

~/.local/lib/python3.8/site-packages/pandas/core/reshape/concat.py in concat(objs, axis, join, ignore_index, keys, levels, names, verify_integrity, sort, copy)
    269     ValueError: Indexes have overlapping values: ['a']
    270     """
--> 271     op = _Concatenator(
    272         objs,
    273         axis=axis,

~/.local/lib/python3.8/site-packages/pandas/core/reshape/concat.py in __init__(self, objs, axis, join, keys, levels, names, ignore_index, verify_integrity, copy, sort)
    327 
    328         if len(objs) == 0:
--> 329             raise ValueError("No objects to concatenate")
    330 
    331         if keys is None:

ValueError: No objects to concatenate

[mchitre]

@patel999jay thanks for the detailed report. I'm unable to reproduce the error, as when I run the code you have in the notebook, I get correct results.

To debug further, I suggest we look at the Bellhop files you get. Run this:

import arlpy.uwapm as pm
env = pm.create_env2d()
rays = pm.compute_eigenrays(env, debug=True)

You should get an output like this (with different filenames):

[DEBUG] Model: bellhop
[DEBUG] Bellhop working files: /var/folders/d5/wwnqgqpj13s503lrdb4pp0680000gn/T/tmpvj7gmy3o.*

In the specified folder, you should see a few relevant files, e.g.,

tmpvj7gmy3o.env
tmpvj7gmy3o.prt
tmpvj7gmy3o.ray

Zip those files and upload here for me to take a look. Thanks.

[patel999jay]

Hi Dr Mandar, I would suspect that this is the debian related issue (I just upgrade to Ubuntu 20.04 LTS), something might not looks right to me. It still the same, I am not able to compute eigenrays with arlpy somehow. I had tested manually run munk tests with bellhop and it did worked, although i found unusual print files for them. MunkB_ray.zip

[mchitre]

@patel999jay the files in the zip are your manually created ones. In fact if I try to run it, I get a fatal error "Unknown option for SSP approximation". I'm guessing we have a very different version of Bellhop? Are you on the latest one, compiled from sources?

I'm testing against the version here.

[patel999jay]

@mchitre I had tested again your version as well. If you look at the manual bellhop run in the zip file it has some extra errors generated by program itself. Here :

Tracing beam      41     20.00

 *** WARNING ***
 Generated by program or subroutine: TraceRay2D
 Insufficient storage for ray trajectory

 CPU Time =        21.3    s

I am using ubuntu 20.04 LTS (I had tested the same in virtual system LTS 18.04, it works fine. I mean bellhop without any errors.)

[mchitre]

So do I understand correctly that this is a Bellhop problem on your system, and not an issue with arlpy?

[patel999jay]

Your version of bellhop compiled correctly on newer version of ubuntu. It should perfectly run the arlpy right as i can see bellhop in my path via arlpy. Does arlpy have any dependencies on any python 2 packages ?

[mchitre]

No, Python 3 only.

Great, I'm marking this issue as resolved. If you encounter any further problems, re-open or create new issue. Thanks.

[vinz-uts]

Hi, I'm trying to use Bellhop model and arlpy library and I get a similar problem. If I try to run

>>> import arlpy.uwapm as pm
>>> env = pm.create_env2d()
>>> rays = pm.compute_eigenrays(env, debug=True)
[DEBUG] Model: bellhop
[WARN] Bellhop did not generate expected output file
[DEBUG] Bellhop working files: /tmp/tmp6cp_l3tx.*

a Warning arise and only the *.env file is generate in tmp folder. I've just installed the Acoustics Toolbox on Ubuntu 20.04.6 and tested it on Matlab, but I would try to use it also in Python 3. I tried to follow your previous instruction, but the link about your tested version of at now dosen't work. Do you have some suggestion about this issue, or someone had the same problem? Thanks in advance!

[mchitre]

No other files generated in the /tmp/ folder? If so, check if the bellhop.exe is in path and can be run from Python.

[vinz-uts]

Hi @mchitre, thanks for your reply! I created a lib folder in at directory and I copied there all the *.exe files, included the bellhop.exe file. Then I add this folder in my path by typing export PATH="~/...../at/lib:$PATH". Please, can you better explain to me how I can check if I'm able to run bellhop.exe from Python on Ubuntu? Thanks for you support!

[mchitre]

You should be able to run this:

>>> import arlpy.uwapm as pm
>>> pm.models()
['bellhop']

to check that the bellhop model executable is found and runnable.

[vinz-uts]

Thank you @mchitre. So I already test it and it works, I have your same results: ['bellhop']. But pm.compute_eigenrays function don't generate other files in tmp folder...

[mchitre]

Then there should be at least a prt file telling you that there is an error, if you run with debug set to True

[vinz-uts]

No *.prt file is generated, only the *.env file was generated when I execute the code posted in my first question (that as debug=True). I just try to execute it now again.

[vinz-uts]

I'm doing some tests more... Watching inside the uwapm.py file a error occurs in _Bellhop.run() function, more precisely adding some line of code in the try-except statement at lines 606-609, I discovered that the error was a [Errno 2] No such file or directory: '/tmp/tmpoq4ejr80.ray'. I don't if this information can be useful to you to helping me to understand what is the problem... @update: I followed the execution thread of the code and the problem seems to be the execution of the _load_rays function in uwapm.py file. Indeed fname_base+'.ray' can't be loaded because is not present neither in /tmp folder nor in the local folder of execution. Who has to create it? @update2 [fixed]: Continuing to investigate, I see the problem is in the call of function _proc.run(f'bellhop.exe {....}...)' and, at the end I found that the problem was, as you said in the begining, in the environmental variable $PATH which don't understand ~ as /home/user path! So now I tested the compute_eigenray() function and it properly generate an output. I probably get ['bellhop'] as output of pm.models() because at the end of uwapm.py file there is the line _models.append(('bellhop', _Bellhop)). So, @mchitre, thank you for your support!

[mchitre]

Glad it's resolved! And thanks for posting the details of what the issue was.

[jrmckenna]

@mchitre i am having the same exact problem as @vinz-uts, but trying to run on windows 10. I've added the path variables (to point to the bellhop.exe, but [bellhop] is working from python, but only the *.env file is being created. My paths are here .

here is how they show up;

Any idea? Thank you so much for your time.

[mchitre]

Make sure your PATH is set to point to bellhop for the environment in which you run Python. You should be able to check the environment variable from your Python REPL/notebook.