[O] ipython integration fails

[Prof. Dr. Johanna May]

Hi everybody,

I am creating lecture notes including some python snippets. Now, in
order to teach jupyter notebook use I would like to include just the
same code as in a python snippet in an ipython snippet.

However ipython blocks don't work in my set up and I could not figure
out which dependency or config line I might need to add in order to do
this properly (and scimax also creates errors, but since I have my own
config running I thought ipython might  already do the job).

The ipython snippet I used is:
#+BEGIN_SRC ipython :session :results raw drawer
  %matplotlib inline
  import matplotlib.pyplot as plt
  import numpy as np
#+END_SRC

The error message is:
Traceback (most recent call last):
  File "/home/username/.emacs.d/elpa/ob-ipython-20180224.953/client.py", line 
60, in <module>
    c = create_client(args.conn_file)
  File "/home/username/.emacs.d/elpa/ob-ipython-20180224.953/client.py", line 
43, in create_client
    cf = find_connection_file('emacs-' + name)
  File "/usr/lib/python3.6/site-packages/jupyter_client/connect.py", line 218, 
in find_connection_file
    raise IOError("Could not find %r in %r" % (filename, path))
OSError: Could not find 'emacs-default' in ['.', '/run/user/1000/jupyter']

Maybe I do not need ipython anyway. What works are python code snippets
like the following one. Minted exports both, code and results and they
are both displayed in the pdf slides/manuscript (using lualatex,
beamerarticle etc.)

#+begin_src python :results output :session :exports both
import time
time.sleep(3)
#+end_src

However, then, what does not work properly is several code blocks in one
session, e.g.

(1)
#+begin_src python :results output :session :exports both
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import datetime as dt
import seaborn as sns
sns.set(style="darkgrid")
#+end_src

(2)
#+begin_src python :results output :session :exports both
df = pd.read_csv('./household_data_1min_singleindex.csv')
#+end_src

(3)
#+begin_src python :results output :session :exports both
zeit=df.utc_timestamp.apply(lambda x: dt.datetime.strptime(x, 
"%Y-%m-%dT%H:%M:%SZ"))
#+end_src

(4)
#+begin_src python :results file :session :var 
matplot_lib_filename=(org-babel-temp-file "figu" ".png"),fontsize=fs :exports 
both
plt.rcParams.update({'font.size':fontsize})
rcParams.update({'figure.autolayout': True})
plt.plot(zeit,df.DE_KN_residential2_washing_machine)
plt.savefig(matplot_lib_filename)
matplot_lib_filename
#+end_src

I do not like the fact that plt.rcParams.update and other lines that
jupyter does not need get exported - is there a simple way to exclude
certain lines from export?

And I do get in serious trouble when I add a second data set (another csv) and 
do
further analysis steps - all graphics then look the same and bad and do
not represent the actual result of the code block. It seems that all get
the same wrong name and then at each graphics location this same png
gets displayed. 

So, maybe there is a good tutorial on using sessions in ob-python or
some other hint that I could use?

Is there a way to restart the (python) kernel with some header message
in the src block?

Thank you very much,

Cheers,

J. May