fix: correct tuple index in ProForma.__getitem__ slicing

[steps-re]

Bug

Slicing a ProForma sequence that has a grouped modification (e.g. #g1) raises a TypeError.

In ProForma.__getitem__, when handling a slice, the code loops over tag_hits from find_tags_by_id(group_id, include_position=True), which returns a list of (position, tag) tuples. The loop variable i gets reassigned to tag_hits[0], the whole tuple, and is then used directly as subseq[i]. __getitem__ only accepts int or slice, so indexing with a tuple fails.

Repro:

from pyteomics import proforma

seq = "EMEVT[#g1]S[#g1]ES[#g1]PEK"
obj = proforma.ProForma.parse(seq)
obj[2:9]
# TypeError: list indices must be integers or slices, not tuple

Fix

Use tag_hits[0][0] (the position) instead of the full tuple.

Testing

• Added test_slice_grouped_modification in tests/test_proforma.py, which fails with the TypeError above on the old code and passes with the fix.
• Ran the full tests/test_proforma.py suite: 62 passed, 0 failed, no regressions.

[mobiusklein]

Thank you for reporting this. I believe this is amongst the many historical bugs being fixed in #212. Manually testing your example with that branch worked as expected. Could you please test that branch for your use-case?

[steps-re]

Confirmed, thanks @mobiusklein. I checked out fix/proforma-coverage-fixes (#212) and ran my exact repro against it:

from pyteomics import proforma
sub = proforma.ProForma.parse("EMEVT[#g1]S[#g1]ES[#g1]PEK")[2:9]
# -> "EVT[#g1]S[#g1]ES[#g1]P", group_ids == ['#g1']

No more TypeError and the slice is correct, so #212 covers this. Closing this in favor of it.

If it's useful, the regression test I added here (test_slice_grouped_modification) pins this specific grouped-modification slice case, and #212 is all about backfilling tests for historically untested paths, so you're welcome to lift it in. Happy to open a tiny test-only PR against your branch if you'd prefer. Thanks for the fast turnaround.