Fix Rule Type

[willtebbutt]

Reverts a change that was made in #339 . In that PR, I changed the implementation of rule_type in such a way that it didn't agree with what type inference produced when I actually derived a rule. This causes some problems, so I've reverted it.

A regression test has been added to catch the offending case, found by @penelopeysm in #345 . I've checked that this fixes the problem locally.

[codecov]

Codecov Report

Attention: Patch coverage is 90.90909% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/test_resources.jl	0.00%	1 Missing ⚠️

Files with missing lines	Coverage Δ
src/interpreter/s2s_reverse_mode_ad.jl	91.60% <100.00%> (+0.09%)	⬆️
src/test_resources.jl	89.04% <0.00%> (-0.26%)	⬇️

[github-actions]

Performance Ratio:
Ratio of time to compute gradient and time to compute function.
Warning: results are very approximate! See here for more context.

┌────────────────────────────┬──────────┬──────────┬─────────────┬─────────┐
│                      Label │ Mooncake │   Zygote │ ReverseDiff │  Enzyme │
│                     String │   String │   String │      String │  String │
├────────────────────────────┼──────────┼──────────┼─────────────┼─────────┤
│                   sum_1000 │     69.8 │      1.0 │        5.51 │ missing │
│                  _sum_1000 │     6.69 │ 337000.0 │        34.4 │ missing │
│               sum_sin_1000 │     2.29 │     1.71 │        10.6 │ missing │
│              _sum_sin_1000 │     2.67 │    218.0 │        13.8 │ missing │
│                   kron_sum │     74.6 │     6.87 │       268.0 │ missing │
│              kron_view_sum │     53.5 │     8.61 │       208.0 │ missing │
│      naive_map_sin_cos_exp │     2.49 │  missing │        7.08 │ missing │
│            map_sin_cos_exp │      2.7 │     1.45 │        6.18 │ missing │
│      broadcast_sin_cos_exp │     2.57 │     2.26 │        1.46 │ missing │
│                 simple_mlp │     2.88 │     2.84 │        6.17 │ missing │
│                     gp_lml │      5.0 │     2.35 │     missing │ missing │
│ turing_broadcast_benchmark │     3.14 │  missing │        23.8 │ missing │
│         large_single_block │     4.13 │   4250.0 │        30.7 │ missing │
└────────────────────────────┴──────────┴──────────┴─────────────┴─────────┘