pst_from.add_parameters - MF6 list parameters with auxiliary variables

[rhugman]

Hi again,

I am trying to add list file parameterization for a single column in a file with multiple columns and auxiliary variable names (for example below, a GHB in a 2D x-section variable density model).

  1 1 151       0.00000000       0.10000000       0.70000000  riverBC
  1 1 152       0.00000000       0.10000000       0.70000000  riverBC

From the documentation, I assume that assigning index_cols=[0,1,2], use_cols=[5] informs pst_from to use the first three columns as [row, lay, col], and column 5 as the column to be parameterised (am I misinterpreting?).

However, this returns the error below. Am I setting this up incorrectly? Can you provide any insight on how to structure the code in such a case? I haven't found a similar case in the example notebooks(edit1: nvm, I see the Freyburg example does).

edit2: using par_style='multiplier', does not result in this error.

---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
<ipython-input-17-b3817055a53a> in <module>
     10                     pargp="ghb_gr",
     11                     upper_bound = 1e8, lower_bound=1e-8,
---> 12                     index_cols=[0,1,2], use_cols=[5])

..\pyemu\pyemu\utils\pst_from.py in add_parameters(self, filenames, par_type, zone_array, dist_type, sigma_range, upper_bound, lower_bound, transform, par_name_base, index_cols, use_cols, pargp, pp_space, use_pp_zones, num_eig_kl, spatial_reference, geostruct, datetime, mfile_fmt, mfile_skip, ult_ubound, ult_lbound, rebuild_pst, alt_inst_str, comment_char, par_style)
   1478                 zero_based=self.zero_based,
   1479                 input_filename=in_fileabs,
-> 1480                 par_style=par_style,
   1481             )
   1482             assert (

..\pyemu\pyemu\utils\pst_from.py in write_list_tpl(filenames, dfs, name, tpl_filename, index_cols, par_type, use_cols, suffix, zone_array, gpname, longnames, get_xy, ij_in_idx, xy_in_idx, zero_based, input_filename, par_style)
   2129             ij_in_idx=ij_in_idx,
   2130             xy_in_idx=xy_in_idx,
-> 2131             zero_based=zero_based,
   2132         )
   2133     else:

..\pyemu\pyemu\utils\pst_from.py in _write_direct_df_tpl(in_filename, tpl_filename, df, name, index_cols, typ, use_cols, suffix, zone_array, longnames, get_xy, ij_in_idx, xy_in_idx, zero_based, gpname)
   2322         df_ti.loc[:, "sidx"] = df_ti.sidx.apply(lambda x: tuple(xx - 1 for xx in x))
   2323     df_ti.loc[:, "idx_strs"] = df_ti.sidx.apply(
-> 2324         lambda x: j.join([fmt.format(iname, xx) for xx, iname in zip(x, inames)])
   2325     ).str.replace(" ", "")
   2326 

C:\ProgramData\Anaconda3\lib\site-packages\pandas\core\series.py in apply(self, func, convert_dtype, args, **kwds)
   4198             else:
   4199                 values = self.astype(object)._values
-> 4200                 mapped = lib.map_infer(values, f, convert=convert_dtype)
   4201 
   4202         if len(mapped) and isinstance(mapped[0], Series):

pandas\_libs\lib.pyx in pandas._libs.lib.map_infer()

..\pyemu\pyemu\utils\pst_from.py in <lambda>(x)
   2322         df_ti.loc[:, "sidx"] = df_ti.sidx.apply(lambda x: tuple(xx - 1 for xx in x))
   2323     df_ti.loc[:, "idx_strs"] = df_ti.sidx.apply(
-> 2324         lambda x: j.join([fmt.format(iname, xx) for xx, iname in zip(x, inames)])
   2325     ).str.replace(" ", "")
   2326 

..\pyemu\pyemu\utils\pst_from.py in <listcomp>(.0)
   2322         df_ti.loc[:, "sidx"] = df_ti.sidx.apply(lambda x: tuple(xx - 1 for xx in x))
   2323     df_ti.loc[:, "idx_strs"] = df_ti.sidx.apply(
-> 2324         lambda x: j.join([fmt.format(iname, xx) for xx, iname in zip(x, inames)])
   2325     ).str.replace(" ", "")
   2326 

KeyError: '1|3'

[briochh]

Hi @rhugman, great that you are stress testing this PstFrom functionality. Thanks for your patience! It sounds like you are setting it up right, could you send a snippet of your add_parameters() line? And maybe a copy of the file you are trying to set up.

[rhugman]

Sure thing. See attached.
gwf.ghb_stress_period_data_1.txt

Code snippets:
This returns the error above:

pf.add_parameters(filenames=i,
                    par_style='direct',
                    par_type="constant",
                    par_name_base="ghb_gr",
                    pargp="ghb_gr", 
                    upper_bound = 10., lower_bound=0.1,
                    index_cols=[0,1,2], use_cols=[4])

This works (as in, pyEMU carries on and pestpp-ies runs):

pf.add_parameters(filenames=i,
                    par_style='multiplier',
                    par_type="constant",
                    par_name_base="ghb_gr",
                    pargp="ghb_gr", 
                    upper_bound = 10., lower_bound=0.1,
                    index_cols=[0,1,2], use_cols=[4])

This also works (unique parameter per cell):

pf.add_parameters(filenames=i,
                    par_style='direct',
                    par_type="grid",
                    par_name_base="ghb_gr",
                    pargp="ghb_gr", 
                    upper_bound = 10., lower_bound=0.1,
                    index_cols=[0,1,2], use_cols=[4])

[jtwhite79]

I wonder if the issue is using direct constant pars? I think only grid-type pars are support for the direct-style pars on list-type files since grid-type is the only one that can represent potential variability in the use_cols

[briochh]

@jtwhite79 that would make sense. Do we want to support constant direct par setup? Seems like there might be use case such as this. Else we probs do need to warn earlier.

[jtwhite79]

I don't see why we couldn't support it. We just need to check that all the use cols have the same value.

[briochh]

Hi @rhugman, do the changes on #187 help with this issue?

[rhugman]

Thanks @briochh, seems to have resolved it. Cheers