bamCompare: compute --operation first/second/add/mean and orient reciprocal_ratio as documented

[cindykrafft]

Fixes #1457.

• Does the PR contain new feature? No
• Does the PR contain bugfix? Yes — fixes bamCompare (4.0.0 Rust backend): --operation first/second/add/mean write the log2 ratio, and reciprocal_ratio is inverted #1457
• Does the PR contain documentation changes? A CHANGES.txt bullet only
• Does the PR contain changes to the galaxy wrapper? No

calc_ratio (src/calc.rs) matched only log2, ratio, reciprocal_ratio and subtract; its catch-all arm computed the log2 ratio, so the four remaining --operation choices (first, second, add, mean) wrote the log2 track. The reciprocal_ratio arm returned b/a for a/b ≥ 1 and −a/b otherwise, the inverse of the documented "a/b if a/b ≥ 1 else −b/a" that 3.5.x's getRatio implemented (2/3 became −0.67 instead of −1.5). This adds four explicit arms returning the scaled signals without a pseudocount (as before, and as the --pseudocount help says), swaps the two reciprocal_ratio branches, and makes the catch-all a panic so an unknown operation cannot produce a plausible track. Behaviour change to flag in review: the existing test_calc_ratio pinned the inverted reciprocal value (-0.27 for 6/22); it now expects −22/6 = −3.67. Not changed here, but worth a look: the subtract arm adds both pseudocounts before subtracting (invisible at the default --pseudocount 1, a shift of p1 − p2 with two different pseudocounts; 3.5.x subtracted the scaled signals only).

Tests: test_calc_ratio corrected, test_calc_ratio_reciprocal_ratio_sign_and_orientation and test_calc_ratio_first_second_add_mean added (src/tests/test_calc.rs); test_bam_compare_operations_first_second_add_mean_reciprocal checks all five operations on the shipped testA.bam/testB.bam. On master the pytest fails (first writes 3R 0 50 0 | 50 100 -1 | …) and the cargo tests fail (left: -0.27, right: -3.67; left: 0.5, right: 2.0); with the patch cargo test 50 passed, test_bamCoverage_and_bamCompare.py 22 passed, full pytest suite 233 passed.

Found in Mytochondria, a volunteer project that checks the numerical core of research software and verifies every finding by execution (methods and harnesses: https://github.com/cindykrafft/mytochondria/tree/main/audits/deeptools)

---

Generated by Claude Code

---