Solid DMFT error: Key Aimp_maxent_0 does not exist

[reducedpc]

After successfully completing Step 3 of the solid_dmft 3.3.2 tutorial on SrVO3 , we come to Step 4 where plot the spectral function.

First we run max entropy method:

mpirun -n 30 python3 <path_to_solid_dmft>/python/solid_dmft/postprocessing/maxent_gf_imp.py ./J0.0/U4/out/svo.h5

I obtain the following output.

Starting run with 64 MPI rank(s) at : 2025-06-08 16:10:00.038016
Continuing impurities with blocks:
- Imp 0: ['down_0', 'up_0']
--------------------------------------------------

Rank 0: solving impurity 1 block 'down_0'

Rank 1: solving impurity 1 block 'up_0'
appending
appending
appending
appending
appending
appending
Optimal alphas , Imp 1 block 'down_0':
--- Real part ---
 [[34 -1 -1]
 [-1 31 -1]
 [-1 -1 32]]
(a -1 indicates that maxent did not run for this block due to symmetry)
Optimal alphas , Imp 1 block 'up_0':
--- Real part ---
 [[34 -1 -1]
 [-1 31 -1]
 [-1 -1 32]]
(a -1 indicates that maxent did not run for this block due to symmetry)
--------------------------------------------------------------------------------
DONE
Total run time: 247 s.

In my case, I used mpirun -n 64 python3 /home/barman/triqs_tools/triqs3_install/lib/python3.10/site-packages/solid_dmft/postprocessing/maxent_gf_imp.py J0.0/U4/svo.h5.

However, when I run python3 read_spectral_function.py, I get this:

Traceback (most recent call last): File "/home/barman/CODES_SOLID_DMFT/SrVO3/read_spectral_function.py", line 24, in <module> A_imp = np.array(read_h5(filename, f'{internal_prefix}/Aimp_w_line_fit/up_0')) File "/home/barman/CODES_SOLID_DMFT/SrVO3/read_spectral_function.py", line 12, in read_h5 read_arr = archive[h5_internal_path] File "/home/barman/triqs_tools/triqs3_install/lib/python3.10/site-packages/h5/archive.py", line 203, in __getitem__ for s in l[:-1]: a = a.get_raw(s) File "/home/barman/triqs_tools/triqs3_install/lib/python3.10/site-packages/h5/archive.py", line 194, in get_raw return self.__getitem1__(key,False) File "/home/barman/triqs_tools/triqs3_install/lib/python3.10/site-packages/h5/archive.py", line 211, in __getitem1__ raise KeyError("Key %s does not exist."%key) KeyError: 'Key Aimp_maxent_0 does not exist.'

Seems Aimp_maxent_0 is not available in the hdf5 file.

Is this a familiar bug or am I missing any important file?

[the-hampel]

Hi @reducedpc ,

no that should work I think. I just checked and maxent_gf_imp.py should store results into the archive in that group, i.e. h5ls ./J0.0/U4/out/svo.h5/DMFT_results/last_iter should list all members of that group including Aimp_maxent_0 . Can you check this via h5ls? You can check

/home/barman/triqs_tools/triqs3_install/lib/python3.10/site-packages/solid_dmft/postprocessing/maxent_gf_imp.py

there is a function called:

_write_spectral_function_to_h5(unpacked_results, external_path, iteration):       

that stores the maxent results. It points to archive[h5_internal_path][f'Aimp_maxent_{i}'] = res . So that should be correct. Let me know what h5ls tells you and we will start from there. If you found Aimp_maxent_0 in the h5. Try to print also it contents:

h5ls ./J0.0/U4/out/svo.h5/DMFT_results/last_iter/Aimp_maxent_0/Aimp_w_line_fit

Best,
Alex

[the-hampel]

okay thanks for checking. Can you check again that

mpirun -n 30 python3 <path_to_solid_dmft>/python/solid_dmft/postprocessing/maxent_gf_imp.py ./J0.0/U4/out/svo.h5

runs without problems / errors and you are pointing both scripts to the same h5? You could even add a little print in

 <path_to_solid_dmft>/python/solid_dmft/postprocessing/maxent_gf_imp.py

where it stores the spectral function to see if it is called. It should store the results in the h5 it read from.

Best,
Alex

[reducedpc]

Hi Alex,

How can I check specifically check if the input and output hd5 are in the same location? We have _read_h5(external_path, iteration) and _write_spectral_function_to_h5(unpacked_results, external_path, iteration). So, shouldn't both by construction, the code use the same hd5 file?

Btw I added a print command for the iteration below:

 with HDFArchive(external_path, 'a') as archive:
        for i, res in enumerate(unpacked_results):
            print ('i=',i,' res=',res)
            archive[h5_internal_path][f'Aimp_maxent_{i}'] = res

and I get

...
Optimal alphas , Imp 1 block 'down_0':
--- Real part ---
[[34 -1 -1]
[-1 34 -1]
[-1 -1 35]]
(a -1 indicates that maxent did not run for this block due to symmetry)
Optimal alphas , Imp 1 block 'up_0':
--- Real part ---
[[34 -1 -1]
[-1 34 -1]
[-1 -1 35]]
(a -1 indicates that maxent did not run for this block due to symmetry)
i= 0  res= {'mesh': array([-2.00000000e+01, -1.96577295e+01, -1.93172095e+01, -1.89784666e+01,
     -1.86415278e+01, -1.83064205e+01, -1.79731724e+01, -1.76418119e+01,
     -1.73123674e+01, -1.69848680e+01, -1.66593431e+01, -1.63358226e+01,
     -1.60143367e+01, -1.56949161e+01, -1.53775918e+01, -1.50623954e+01,
     -1.47493588e+01, -1.44385144e+01, -1.41298948e+01, -1.38235333e+01,
     -1.35194634e+01, -1.32177191e+01, -1.29183348e+01, -1.26213454e+01,
     -1.23267859e+01, -1.20346922e+01, -1.17451000e+01, -1.14580458e+01,
     -1.11735664e+01, -1.08916988e+01, -1.06124806e+01, -1.03359497e+01,
     -1.00621443e+01, -9.79110281e+00, -9.52286423e+00, -9.25746773e+00,
     -8.99495281e+00, -8.73535928e+00, -8.47872721e+00, -8.22509696e+00,
     -7.97450912e+00, -7.72700452e+00, -7.48262422e+00, -7.24140948e+00,
     -7.00340171e+00, -6.76864253e+00, -6.53717365e+00, -6.30903694e+00,
     -6.08427433e+00, -5.86292785e+00, -5.64503956e+00, -5.43065155e+00,
     -5.21980588e+00, -5.01254460e+00, -4.80890969e+00, -4.60894303e+00,
     -4.41268639e+00, -4.22018135e+00, -4.03146935e+00, -3.84659156e+00,
     -3.66558893e+00, -3.48850210e+00, -3.31537138e+00, -3.14623673e+00,
     -2.98113770e+00, -2.82011341e+00, -2.66320248e+00, -2.51044304e+00,
     -2.36187265e+00, -2.21752828e+00, -2.07744628e+00, -1.94166230e+00,
     -1.81021130e+00, -1.68312747e+00, -1.56044421e+00, -1.44219411e+00,
     -1.32840884e+00, -1.21911921e+00, -1.11435505e+00, -1.01414522e+00,
     -9.18517531e-01, -8.27498761e-01, -7.41114578e-01, -6.59389520e-01,
     -5.82346964e-01, -5.10009090e-01, -4.42396852e-01, -3.79529945e-01,
     -3.21426784e-01, -2.68104470e-01, -2.19578772e-01, -1.75864097e-01,
     -1.36973477e-01, -1.02918544e-01, -7.37095123e-02, -4.93551692e-02,
     -2.98628554e-02, -1.52384577e-02, -5.48639875e-03, -6.09630646e-04,
      6.09630646e-04,  5.48639875e-03,  1.52384577e-02,  2.98628554e-02,
      4.93551692e-02,  7.37095123e-02,  1.02918544e-01,  1.36973477e-01,
      1.75864097e-01,  2.19578772e-01,  2.68104470e-01,  3.21426784e-01,
      3.79529945e-01,  4.42396852e-01,  5.10009090e-01,  5.82346964e-01,
      6.59389520e-01,  7.41114578e-01,  8.27498761e-01,  9.18517531e-01,
      1.01414522e+00,  1.11435505e+00,  1.21911921e+00,  1.32840884e+00,
      1.44219411e+00,  1.56044421e+00,  1.68312747e+00,  1.81021130e+00,
      1.94166230e+00,  2.07744628e+00,  2.21752828e+00,  2.36187265e+00,
      2.51044304e+00,  2.66320248e+00,  2.82011341e+00,  2.98113770e+00,
      3.14623673e+00,  3.31537138e+00,  3.48850210e+00,  3.66558893e+00,
      3.84659156e+00,  4.03146935e+00,  4.22018135e+00,  4.41268639e+00,
      4.60894303e+00,  4.80890969e+00,  5.01254460e+00,  5.21980588e+00,
      5.43065155e+00,  5.64503956e+00,  5.86292785e+00,  6.08427433e+00,
      6.30903694e+00,  6.53717365e+00,  6.76864253e+00,  7.00340171e+00,
      7.24140948e+00,  7.48262422e+00,  7.72700452e+00,  7.97450912e+00,
      8.22509696e+00,  8.47872721e+00,  8.73535928e+00,  8.99495281e+00,
      9.25746773e+00,  9.52286423e+00,  9.79110281e+00,  1.00621443e+01,
      1.03359497e+01,  1.06124806e+01,  1.08916988e+01,  1.11735664e+01,
      1.14580458e+01,  1.17451000e+01,  1.20346922e+01,  1.23267859e+01,
      1.26213454e+01,  1.29183348e+01,  1.32177191e+01,  1.35194634e+01,
      1.38235333e+01,  1.41298948e+01,  1.44385144e+01,  1.47493588e+01,
      1.50623954e+01,  1.53775918e+01,  1.56949161e+01,  1.60143367e+01,
      1.63358226e+01,  1.66593431e+01,  1.69848680e+01,  1.73123674e+01,
      1.76418119e+01,  1.79731724e+01,  1.83064205e+01,  1.86415278e+01,
      1.89784666e+01,  1.93172095e+01,  1.96577295e+01,  2.00000000e+01]), 'Aimp_w_line_fit': {'down_0': array([[[1.89218171e-09 +0.j, 1.77947265e-09 +0.j, 1.67833191e-09 +0.j,
         ..., 6.45819486e-05 +0.j, 6.38234152e-05 +0.j,
       6.31921194e-05 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j]],

     [[           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [1.56959199e-09 +0.j, 1.53324484e-09 +0.j, 1.50404473e-09 +0.j,
       ..., 2.03698185e-04 +0.j, 2.05121091e-04 +0.j,
       2.06808945e-04 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j]],

     [[           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [3.91175745e-04 +0.j, 3.85344407e-04 +0.j, 3.79674038e-04 +0.j,
       ..., 2.10712891e-06 +0.j, 2.02981512e-06 +0.j,
       1.96227379e-06 +0.j]]]), 'up_0': array([[[1.89218171e-09 +0.j, 1.77947265e-09 +0.j, 1.67833191e-09 +0.j,
       ..., 6.45819486e-05 +0.j, 6.38234152e-05 +0.j,
       6.31921194e-05 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j]],

     [[           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [1.56959199e-09 +0.j, 1.53324484e-09 +0.j, 1.50404473e-09 +0.j,
       ..., 2.03698185e-04 +0.j, 2.05121091e-04 +0.j,
       2.06808945e-04 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j]],

     [[           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [3.91175745e-04 +0.j, 3.85344407e-04 +0.j, 3.79674038e-04 +0.j,
       ..., 2.10712891e-06 +0.j, 2.02981512e-06 +0.j,
          1.96227379e-06 +0.j]]])}, 'Aimp_w_chi2_curvature': {'down_0': array([[[4.04604759e-12 +0.j, 3.73832724e-12 +0.j, 3.48926090e-12 +0.j,
       ..., 6.11981566e-05 +0.j, 6.04623217e-05 +0.j,
       5.98425592e-05 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j]],

     [[           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [2.01935470e-14 +0.j, 1.27962032e-14 +0.j, 1.58729096e-14 +0.j,
       ..., 3.18437770e-04 +0.j, 3.24288905e-04 +0.j,
       3.30464728e-04 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j]],

     [[           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [3.46390319e-04 +0.j, 3.42828368e-04 +0.j, 3.39330037e-04 +0.j,
       ..., 2.07875017e-07 +0.j, 1.95293406e-07 +0.j,
       1.84302079e-07 +0.j]]]), 'up_0': array([[[4.04604759e-12 +0.j, 3.73832724e-12 +0.j, 3.48926090e-12 +0.j,
       ..., 6.11981566e-05 +0.j, 6.04623217e-05 +0.j,
       5.98425592e-05 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j]],

     [[           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [2.01935470e-14 +0.j, 1.27962032e-14 +0.j, 1.58729096e-14 +0.j,
       ..., 3.18437770e-04 +0.j, 3.24288905e-04 +0.j,
       3.30464728e-04 +0.j],
      [0.00000000e+00 +0.j, 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       ..., 0.00000000e+00 +0.j, 0.00000000e+00 +0.j,
       0.00000000e+00 +0.j]],

     [[           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
      [           nan+nanj,            nan+nanj,            nan+nanj,
       ...,            nan+nanj,            nan+nanj,
                  nan+nanj],
[3.46390319e-04 +0.j, 3.42828368e-04 +0.j, 3.39330037e-04 +0.j,
       ..., 2.07875017e-07 +0.j, 1.95293406e-07 +0.j,
       1.84302079e-07 +0.j]]])}}
--------------------------------------------------------------------------------
DONE
Total run time: 387 s.

[the-hampel]

Yes that should write to the same archive you read from. This is very confusing. Can you maybe execute the script from the folder and change the path in the script so that it finds the h5 file. Then check after it runs that the timestamp changed via 'ls -al' of the hdf5 file. I don't see an obvious reason why this fails here.

If you can you can maybe also copy the files to another computer and try there.

[reducedpc]

@the-hampel Sorry for the late response. Actually, the bug was pretty trivial. The read_spectral_function.py in the SrVO3 example has J=0.75 while following the instruction, it has to be J=0.0. The read_spectral_function_transition.py has correctly mentioned J=0.0, so no issues while running that file.

I suggest that you update the file read_spectral_function.py in the GitHub example, or mention to check/modify the values of U and J parameters accordingly in that file before running it.

You may close this ticket, or I shall mark it answered after you finally confirm.