Issues related to traceback print about pyasdf HOT 11 CLOSED

Haha :-) Oh well...yea i did not test that. I'll have a look.

What actually happens when you store the above mentioned data in an ASDF file? Is it stored as single bytes? As mentioned in the obspy issue: you should really fix the file - whatever ends up in the ASDF file is likely garbage.

from pyasdf.

In asdf file, it stores as byte:

In [4]: ds.waveforms.XA_SA71.raw_observed[0].data
Out[4]: 
array(['\x0c', '\r', '\x05', ..., '\x07', '\n', '\n'], 
      dtype='|S1')

My current plan for this is: just throw those file away during processing...So if pyasdf now can catch the error and run through, I feel OK.

from pyasdf.

You can probably fix them with this tool: https://seiscode.iris.washington.edu/projects/msmod

Just change the encoding to whatever it really is - likely STEIM1 or STEIM2.

from pyasdf.

I'll make pyasdf raise an exception if one tries to add non-numeric arrays - that does not really make sense for waveforms in ASDF - the auxiliary data should be used for that.

from pyasdf.

I agree.

So two things?

1. Check the data type for the array during conversion.
2. fix the UnicodeDecodeError in error handling.

from pyasdf.

I strip the code out to form a simple test, but couldn't really reproduce the error.

I have attached my code and data here.
test_read.py.txt
XA.SA72.mseed.txt

Not sure if I am doing the right thing...

from pyasdf.

The encoding issue should be fixed - can you check? You will likely no longer be able to test once the check for the data type is implemented.

from pyasdf.

I am going to test it now.

from pyasdf.

You may forget to delete this line.
https://github.com/SeismicData/pyasdf/blob/master/pyasdf/asdf_data_set.py#L1772

Otherwise, it is fixed :)

Thanks!

from pyasdf.

Nice catch.

from pyasdf.

Alright - should be all fixed. You are no longer allowed to use a datatype that is not valid ASDF. Is will raise a TypeError if you attempt to do it.

from pyasdf.