CPW eclabs file

[mathcodemaster]

Hello,
quite new to the community and i hope that i state this issue in a helpful manner.
I am using yadg to extract data from mpr files which works for a lot of them,
but now i have gotten a format that is not covered yet,
it is CPW used to establish Ragone plots,
it would be great to implement this mode to,
i am happy to provide a solution if i can help but i think i would need slightly better understanding of the data structure

from hacking around i have figured out a way to read it by changing 3 lines in the yadg repo,
this might be helpful to see what is a good minimal solution for the professionals,

1. would be: in tequniques.py line nr. 1388 (in the technique_params_dtypes dictionary):
   just add one line:

    0x88: ("BCD", _bcd_params_dtypes),\b
    0x24: ("CPW", _ca_params_dtypes),   #ca works for cpw!<-----------added Ox24 is otherwise unrecognised
}

2. would be in mpr.py line 275:
   from:
   pardict[k] = param_from_key(k, v, to_str=True)
   to:

  try:
      pardict[k] = param_from_key(k, v, to_str=True)
  except:
      pardict[k]=str(k)        

which is a dirty hack but still delivers reasanobale results, problaby people with better understanding of the data can produce a cleaner solution.

3. would be in mpr.py line 599:
   from
   Eranges = [a - b for a, b in zip(E_range_max, E_range_min)]
   to:
   Eranges = [a - (b if not np.isnan(float(b)) else 0.0) for a, b in zip(E_range_max, E_range_min)]
   which just replaces a string 'nan' with a default 0, which is probably needed due to my dirty hacking in step 2)

Please let me know what you think about this solution,
is it easy to imporve in terms of cleanliness?
do i make a mistake that i just dont get?
the data i get looks quite promising (i mainly look at voltage and current curves over time at teh moment,

best regards

Franz

[PeterKraus]

Hello Franz,
Feel free to submit a PR. My suggestion would be to copy the _ca_params_dtypes into a new _cpw_data_dtypes and keep the technique_params_dtypes dictionary sorted (i.e. put the new entry between GEIS and CV).

Then you can run it in verbose mode. We are quite strict on checking the parameters, and a hard check (i.e. without a try -- except as you suggest) is intentional, so that it's obvious when one is trying to parse an unsupported file.

I suggest exporting the mpr file into a mpt file using EC-Lab, implementing support for both, and then cross-checking that the header is parsed correctly / consistently.

If you'd prefer me to have a look at it instead, feel free to attach a mpr file with this technique (ideally a small one) that I can use for testing.

Cheers,
Peter