[Repo Assist] fix: use string key for column lookup in conditional_MI (fixes KeyError with multi-char column names)

[github-actions]

🤖 This is an automated fix from Repo Assist.

Closes #949

Root Cause

In dowhy/utils/cit.py, conditional_MI used data[list(x)] and data[list(y)] to access columns. When x and y are string column names (as they always are when called from GraphRefuter.conditional_mutual_information), list(x) iterates over individual characters:

list('Foo')  # → ['F', 'o', 'o']
data[['F', 'o', 'o']]  # → KeyError: "None of [Index(['F', 'o', 'o'], ...)] are in the [columns]"

This means any column name longer than one character would trigger the error.

Fix

Change data[list(x)] → data[x] and data[list(y)] → data[y], which performs standard single-column Series lookup. This is correct since x and y are scalar string column names.

Trade-offs

• The fix is minimal and surgical — no behaviour change for the entropy calculation (iterating a Series yields values, which is what the downstream zip / format-string logic expects).
• Single-character column names continue to work correctly.

Test Status

New tests added in tests/causal_refuters/test_graph_refuter.py:

• test_conditional_mi_multi_char_column_names — regression test confirming multi-char names no longer raise KeyError
• test_conditional_mi_single_char_column_names — ensures single-char names still work
• test_graph_refuter_with_multi_char_columns — end-to-end refute_model call with multi-char columns

All new tests pass locally. Formatting (black, isort) and linting (flake8) pass on changed files.

Generated by 🌈 Repo Assist, see workflow run. Learn more.

To install this agentic workflow, run

gh aw add githubnext/agentics/workflows/repo-assist.md@11c9a2c442e519ff2b427bf58679f5a525353f76