Re: pyserial and end-of-line specification

2017-07-18 Thread FS
Thank you for your response Andre. I had tried some code like that in the 
document but it did not seem to work. However ever leaving my terminal for a 
time the code eventually wrote out the records so apparently there is some very 
deep buffering going on here. A little more searching on the web revealed the 
following:

https://stackoverflow.com/questions/10222788/line-buffered-serial-input

It is apparent that pySerial, or at least the documentation is falling short of 
my needs. It is very unclear what module in the layer is handling the buffering 
and newlines and so forth. Also unclear is whether the coupled python and OS is 
reading FIFO or LIFO--something important in quasi realtime scientific 
applications.
  This is problematic since the serial port is still so ubiquitous to a lot of 
scientific instrumentation. I probably will patch up some byte oriented code 
for this or perhaps write the module in C.

Thanks again
Fritz
-- 
https://mail.python.org/mailman/listinfo/python-list


Re: pyserial and end-of-line specification

2017-07-21 Thread FS
Thanks Rob.
  Yes I ended up with a read(1) and use a field count and a few other checks to 
make sure I don't get a partial record. Serial is the "best of times and worst 
of times". Sure beats dealing with USB enumeration, power hungry ethernet 
processors and a lot of other stuff. I can still "see" serial on my o'scope 
which is always nice and I don't see it going away any time soon at least in 
the laboratory.

Python has been a bit of a chore-seems like a lot of verison/rev 
inconsistencies. At any rate I am going to stick with it. I used PERL in the 
past but I covet the stats packages and a few other things I lost when I left 
matlab and  I want to try for chemometrics work.

cheers
fritz
-- 
https://mail.python.org/mailman/listinfo/python-list


Installing matplotlib on python3

2017-07-27 Thread FS
I just installed matplotlib on debian and I tried to import it on python3. It 
cannot be found however it can be found on python 2.x. No surprise:
  A 'find -name matplotliib' reveals:
/usr/share/matplotlib
/usr/lib/python2.7/dist-packages/matplotlib

I am not sure how the apt-get elected to place matplotlib in the python2.7 
directory but I want to "properly" install it so it can import under python3. 

>There are probably commands from python3 to point its import to the 2.7 
>directory, but I expect that is just a workaround and I am uneasy about 
>whether I have somehow installed a 2.7 compatible version only of matplotlib.
>Possibly there is some recommended way to re-install matplotlib for python3 
>sending it to the appropriate directories.

Any advice here? 
thanks
fritz

-- 
https://mail.python.org/mailman/listinfo/python-list


Re: Installing matplotlib on python3

2017-07-27 Thread FS
Thank you Jussi. I didn't realize there was a separate version--I have it 
installed now

cheers
fritz
-- 
https://mail.python.org/mailman/listinfo/python-list