MQE: add support for or

[charleskorn]

What this PR does

This PR adds support for the or binary operation to MQE.

In comparison to Prometheus' engine, MQE runs up to 70% faster and with up to 40% lower peak memory utilisation in our benchmarks:

goos: darwin
goarch: arm64
pkg: github.com/grafana/mimir/pkg/streamingpromql/benchmarks
cpu: Apple M1 Pro
                                                                      │  Prometheus  │               Mimir                │
                                                                      │    sec/op    │   sec/op     vs base               │
Query/a_1_or_b_1{l=~'.*[0-4]$'},_instant_query-10                        724.9µ ± 1%   728.9µ ± 2%        ~ (p=0.937 n=6)
Query/a_1_or_b_1{l=~'.*[0-4]$'},_range_query_with_100_steps-10           750.8µ ± 1%   727.2µ ± 1%   -3.15% (p=0.002 n=6)
Query/a_1_or_b_1{l=~'.*[0-4]$'},_range_query_with_1000_steps-10          1.328m ± 3%   1.188m ± 1%  -10.54% (p=0.002 n=6)
Query/a_100_or_b_100{l=~'.*[0-4]$'},_instant_query-10                    1.633m ± 1%   1.622m ± 2%        ~ (p=0.093 n=6)
Query/a_100_or_b_100{l=~'.*[0-4]$'},_range_query_with_100_steps-10       3.600m ± 1%   2.334m ± 1%  -35.16% (p=0.002 n=6)
Query/a_100_or_b_100{l=~'.*[0-4]$'},_range_query_with_1000_steps-10     20.976m ± 1%   8.572m ± 0%  -59.13% (p=0.002 n=6)
Query/a_2000_or_b_2000{l=~'.*[0-4]$'},_instant_query-10                  17.38m ± 1%   17.01m ± 7%        ~ (p=0.065 n=6)
Query/a_2000_or_b_2000{l=~'.*[0-4]$'},_range_query_with_100_steps-10     61.91m ± 2%   30.69m ± 2%  -50.43% (p=0.002 n=6)
Query/a_2000_or_b_2000{l=~'.*[0-4]$'},_range_query_with_1000_steps-10    487.5m ± 1%   145.3m ± 1%  -70.19% (p=0.002 n=6)
geomean                                                                  7.116m        4.874m       -31.51%

                                                                      │  Prometheus  │                Mimir                │
                                                                      │      B       │      B        vs base               │
Query/a_1_or_b_1{l=~'.*[0-4]$'},_instant_query-10                       69.93Mi ± 1%   69.32Mi ± 3%        ~ (p=0.394 n=6)
Query/a_1_or_b_1{l=~'.*[0-4]$'},_range_query_with_100_steps-10          69.87Mi ± 0%   69.73Mi ± 1%        ~ (p=0.240 n=6)
Query/a_1_or_b_1{l=~'.*[0-4]$'},_range_query_with_1000_steps-10         67.76Mi ± 2%   68.26Mi ± 3%        ~ (p=0.818 n=6)
Query/a_100_or_b_100{l=~'.*[0-4]$'},_instant_query-10                   65.95Mi ± 1%   66.36Mi ± 1%        ~ (p=0.132 n=6)
Query/a_100_or_b_100{l=~'.*[0-4]$'},_range_query_with_100_steps-10      67.23Mi ± 0%   66.59Mi ± 1%   -0.95% (p=0.009 n=6)
Query/a_100_or_b_100{l=~'.*[0-4]$'},_range_query_with_1000_steps-10     73.69Mi ± 1%   69.51Mi ± 1%   -5.67% (p=0.002 n=6)
Query/a_2000_or_b_2000{l=~'.*[0-4]$'},_instant_query-10                 68.48Mi ± 1%   69.93Mi ± 1%   +2.12% (p=0.009 n=6)
Query/a_2000_or_b_2000{l=~'.*[0-4]$'},_range_query_with_100_steps-10    87.44Mi ± 1%   78.52Mi ± 1%  -10.19% (p=0.002 n=6)
Query/a_2000_or_b_2000{l=~'.*[0-4]$'},_range_query_with_1000_steps-10   227.8Mi ± 4%   134.0Mi ± 3%  -41.16% (p=0.002 n=6)
geomean                                                                 80.85Mi        74.94Mi        -7.30%

Which issue(s) this PR fixes or relates to

(none)

Checklist

• Tests updated.
• [n/a] Documentation added.
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX].
• [n/a] about-versioning.md updated with experimental features.

[jhesketh]

Approach looks good to me :-). Nothing in particular I'd change, other than the extra tests you're working on.

[jhesketh]

lgtm, thanks!