[dvvrtmshra]

Pull Request the OpenBB Platform

Description

• Summary of the change/ bug fix.

  df_to_basemodel followed by basemodel_to_df raised a KeyError when the DataFrame had an unnamed or partially named MultiIndex. Unnamed levels are stored in multiindex_names as None, but reset_index() turns them into columns named level_0, level_1, etc., so set_index could not find them.

  This PR records the reset-index column names in a new multiindex_columns field, separately from the original level names. basemodel_to_df rebuilds the index from those columns and then restores the original level names (including None). If multiindex_columns is absent, it falls back to the previous behavior, so existing payloads and the fully named case are unaffected.

• Link # issue, if applicable.

  Fixes #ISSUE_NUMBER

• Screenshot of the feature or the bug before/after fix, if applicable.

  N/A. Before/after is shown by the reproduction below.

• Relevant motivation and context.

  Round-tripping a DataFrame with an unnamed MultiIndex, which pandas produces by default from MultiIndex.from_tuples, groupby on multiple keys, etc., failed outright. Fully named MultiIndexes already worked.

• List any dependencies that are required for this change.

  None.

How has this been tested?

• Added a parametrized regression test, test_multiindex_round_trip_unnamed_levels, in openbb_platform/core/tests/app/test_utils.py, covering [None, None], ["group", None], and [None, "id"]. Each case asserts the round-tripped DataFrame equals the original via pd.testing.assert_frame_equal.
• The new tests fail on the current develop code with the reported KeyError and pass with this change. All existing tests in test_utils.py still pass.

Reproduction:

import pandas as pd
from openbb_core.app.utils import basemodel_to_df, df_to_basemodel

frame = pd.DataFrame(
    {"value": [10, 20]},
    index=pd.MultiIndex.from_tuples([("A", 1), ("B", 2)]),
)
basemodel_to_df(df_to_basemodel(frame))
# Before: KeyError: 'None of [None, None] are in the columns'
# After: returns the DataFrame with its unnamed MultiIndex restored

Test configuration: Windows, Python 3.11.9, OpenBB installed from source via dev_install.py.

• Ensure all unit and integration tests pass.
• If you modified/added command(s):
  • Ensure the command(s) execute with the expected output.
    • API.
    • Python Interface.
  • If applicable, please add new tests for the command (see CONTRIBUTING.md to leverage semi-automated testing).
• If a new provider was introduced or a new fetcher was added to an existing provider:
  • Ensure the existing tests pass.
  • Ensure the new provider and/or fetcher is stable and usable.
  • If applicable, please add new tests for the provider and/or fetcher (see CONTRIBUTING.md to leverage semi-automated testing).
• If a new provider or extension was added:
  • Update the list of Extensions.
  • Update the list of Providers.
  • If it's a community extension or provider, update the integration tests GitHub Action workflow.

Checklist

• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have adhered to the GitFlow naming convention and my branch name is in the format of feature/feature-name or hotfix/hotfix-name.
• I ensure that I am following the CONTRIBUTING guidelines.
  • (If applicable) I have updated tests following these guidelines.

[CLAassistant]

All committers have signed the CLA.

[dlee-bq]

Hi @dvvrtmshra, thanks for the PR, however this PR is superseded by #7762 because of the V5 file move. Closing as a duplicate.