Optimize select() statements by removing redundant conditions

[MaxGhenis]

Summary

Optimizes all np.select() statements in the codebase by removing conditions that return the same value as the default. This is the numpy-efficient pattern where the default handles the most common case(s).

Key insight: When using np.select(), each condition requires an array comparison. By setting default= to the most common value and removing explicit conditions for that value, we reduce evaluations from N to N-k.

Example optimization (taxsim_mstat.py)

Before:

return select(
    [
        filing_status == fstatus.SINGLE,
        filing_status == fstatus.HEAD_OF_HOUSEHOLD,
        filing_status == fstatus.JOINT,
        filing_status == fstatus.SEPARATE,
        filing_status == fstatus.SURVIVING_SPOUSE,
    ],
    [1, 1, 2, 6, 8],
)

After:

return select(
    [
        filing_status == fstatus.JOINT,
        filing_status == fstatus.SEPARATE,
        filing_status == fstatus.SURVIVING_SPOUSE,
    ],
    [2, 6, 8],
    default=1,  # SINGLE, HEAD_OF_HOUSEHOLD
)

Changes

• 74 files modified across federal and state tax/benefit calculations
• Net -10 lines of code (fewer conditions = cleaner code)
• Added clarifying comments documenting what cases the default covers
• Common pattern: SINGLE filing status as default for state income tax calculations

Performance benefit

Each removed condition eliminates one boolean array comparison per element. For microsimulations with millions of tax units, this reduces memory allocations and CPU cycles.

Test plan

• All 74 modified files pass Python syntax validation
• Package imports successfully
• CI tests pass

Supersedes #7242 (which only added default= without removing redundant conditions)

🤖 Generated with Claude Code