Fix Container.as_dict for case of flatten=True and add_prefix=True

[maxnoe]

This change has the main goal of having
ImageParametersContainer.as_dict(recursive=True, flatten=True, add_prefix=True) produce a dict with the same keys as the column names
in the dl1 parameters table in the hdf5 file.

Before:

{'params_hillas_hillas_intensity': nan,
 'params_hillas_hillas_skewness': nan,
 'params_hillas_hillas_kurtosis': nan,
 'params_hillas_hillas_fov_lon': <Quantity nan deg>,
 'params_hillas_hillas_fov_lat': <Quantity nan deg>,
 'params_hillas_hillas_r': <Quantity nan deg>,
 'params_hillas_hillas_phi': <Quantity nan deg>,
 'params_hillas_hillas_length': <Quantity nan deg>,
 'params_hillas_hillas_length_uncertainty': <Quantity nan deg>,
 'params_hillas_hillas_width': <Quantity nan deg>,
 'params_hillas_hillas_width_uncertainty': <Quantity nan deg>,
 'params_hillas_hillas_psi': <Quantity nan deg>,
 'params_timing_timing_intercept': nan,
 'params_timing_timing_deviation': nan,
 'params_timing_timing_slope': <Quantity nan 1 / deg>,
 'params_leakage_leakage_pixels_width_1': nan,
 'params_leakage_leakage_pixels_width_2': nan,
 'params_leakage_leakage_intensity_width_1': nan,
 'params_leakage_leakage_intensity_width_2': nan,
 'params_concentration_concentration_cog': nan,
 'params_concentration_concentration_core': nan,
 'params_concentration_concentration_pixel': nan,
 'params_morphology_morphology_num_pixels': -1,
 'params_morphology_morphology_num_islands': -1,
 'params_morphology_morphology_num_small_islands': -1,
 'params_morphology_morphology_num_medium_islands': -1,
 'params_morphology_morphology_num_large_islands': -1,
 'params_intensity_statistics_intensity_max': nan,
 'params_intensity_statistics_intensity_min': nan,
 'params_intensity_statistics_intensity_mean': nan,
 'params_intensity_statistics_intensity_std': nan,
 'params_intensity_statistics_intensity_skewness': nan,
 'params_intensity_statistics_intensity_kurtosis': nan,
 'params_peak_time_statistics_peak_time_max': nan,
 'params_peak_time_statistics_peak_time_min': nan,
 'params_peak_time_statistics_peak_time_mean': nan,
 'params_peak_time_statistics_peak_time_std': nan,
 'params_peak_time_statistics_peak_time_skewness': nan,
 'params_peak_time_statistics_peak_time_kurtosis': nan,
 'params_core_core_psi': <Quantity nan deg>}

After:

In [4]: c.as_dict(recursive=True, flatten=True, add_prefix=True)
Out[4]: 
{'hillas_intensity': nan,
 'hillas_skewness': nan,
 'hillas_kurtosis': nan,
 'hillas_fov_lon': <Quantity nan deg>,
 'hillas_fov_lat': <Quantity nan deg>,
 'hillas_r': <Quantity nan deg>,
 'hillas_phi': <Quantity nan deg>,
 'hillas_length': <Quantity nan deg>,
 'hillas_length_uncertainty': <Quantity nan deg>,
 'hillas_width': <Quantity nan deg>,
 'hillas_width_uncertainty': <Quantity nan deg>,
 'hillas_psi': <Quantity nan deg>,
 'timing_intercept': nan,
 'timing_deviation': nan,
 'timing_slope': <Quantity nan 1 / deg>,
 'leakage_pixels_width_1': nan,
 'leakage_pixels_width_2': nan,
 'leakage_intensity_width_1': nan,
 'leakage_intensity_width_2': nan,
 'concentration_cog': nan,
 'concentration_core': nan,
 'concentration_pixel': nan,
 'morphology_num_pixels': -1,
 'morphology_num_islands': -1,
 'morphology_num_small_islands': -1,
 'morphology_num_medium_islands': -1,
 'morphology_num_large_islands': -1,
 'intensity_max': nan,
 'intensity_min': nan,
 'intensity_mean': nan,
 'intensity_std': nan,
 'intensity_skewness': nan,
 'intensity_kurtosis': nan,
 'peak_time_max': nan,
 'peak_time_min': nan,
 'peak_time_mean': nan,
 'peak_time_std': nan,
 'peak_time_skewness': nan,
 'peak_time_kurtosis': nan,
 'core_psi': <Quantity nan deg>}

[codecov]

Codecov Report

Merging #1887 (ce67f31) into master (82a8809) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #1887   +/-   ##
=======================================
  Coverage   92.05%   92.05%           
=======================================
  Files         189      189           
  Lines       14862    14868    +6     
=======================================
+ Hits        13681    13687    +6     
  Misses       1181     1181           

Impacted Files	Coverage Δ
ctapipe/core/container.py	88.46% <100.00%> (ø)
ctapipe/core/tests/test_container.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 82a8809...ce67f31. Read the comment docs.

[LukasNickel]

Does this work with multiple algorithms predicting the same quantity (dl2, ml)?
Havent tested it yet, just my first random thought

[LukasNickel]

is_valid from the reconstructed shower container subcontainers for example

[maxnoe]

We should revisit this after #1949 is merged, changing to draft for now

[nbiederbeck]

The use-case for this change is still valid: testing with current master and this branch yields the same results as in the PR description. Therefore, I would suggest merging this, as there are no conflicts and already approved.

Additionally, this fixes a failing test I have on the ml branch.