Factor the exponential histogram into a re-usable package

[jmacd]

Description: The data structure is separated from the adapter used by the Lightstep metrics SDK. This will allow the entire structure package to be migrated into the upstream repository and referenced from the OTel Collector.

This causes minor churn:

1. there was a pattern across all aggregators having an extra/unused Storage generic parameter. It's removed in 4 places.
2. The SDK's use of JSON to parse instrument hints could not be preserved in the move upstream; I created a new config type and adapted the code to match style w/ the upstream repository; new JSON-specific types are used and converted into functional-option-generated configs.

Testing: Minor new tests added to preserve test coverage in both packages.

[jaronoff97]

Given this is mostly renaming and little code change it looks good to me. The structure package makes the import a bit longer which I don't love, but I do like the separation you introduce as it leads to the histogram to be much simpler. Are there any concerns with removing the storage parameter (I know it was unused, but was curious)

[jmacd]

I intend to do a bit more work here. @paivagustavo in a previous code review asked for an optimization in Move(). It requires a Swap() operation to do this right, and I'd like to take the opportunity now.

[jmacd]

(I know it was unused, but was curious)

(The Storage parameter was accidental redundancy, completely unnecessary and an artifact of me learning generics.)

[jmacd]

It's a long import name!

The plan is for github.com/lightstep/otel-launcher-go/lightstep/sdk/metric/aggregator/histogram/structure to become go.opentelemetry.io/otel/sdk/metric/aggregator/exponential/structure, where it will be a sibling of the go.opentelemetry.io/otel/sdk/metric/aggregator/exponential/mapping library which was merged previously.

[codecov-commenter]

Codecov Report

Merging #222 (0a34d83) into main (b60529d) will decrease coverage by 0.00%.
The diff coverage is 97.82%.

@@            Coverage Diff             @@
##             main     #222      +/-   ##
==========================================
- Coverage   94.26%   94.25%   -0.01%     
==========================================
  Files          62       64       +2     
  Lines        3312     3360      +48     
==========================================
+ Hits         3122     3167      +45     
- Misses        145      148       +3     
  Partials       45       45              

Impacted Files	Coverage Δ
lightstep/sdk/metric/aggregator/aggregator.go	87.50% <77.77%> (-12.50%)	⬇️
lightstep/sdk/metric/aggregator/gauge/gauge.go	90.56% <90.00%> (ø)
...metric/aggregator/minmaxsumcount/minmaxsumcount.go	94.28% <90.00%> (ø)
...tstep/sdk/metric/aggregator/histogram/histogram.go	96.72% <98.21%> (-1.45%)	⬇️
...tric/aggregator/histogram/structure/exponential.go	96.55% <99.20%> (ø)
...dk/metric/aggregator/histogram/structure/config.go	100.00% <100.00%> (ø)
.../sdk/metric/aggregator/histogram/structure/test.go	100.00% <100.00%> (ø)
lightstep/sdk/metric/aggregator/sum/sum.go	100.00% <100.00%> (ø)
...ghtstep/sdk/metric/internal/viewstate/viewstate.go	98.63% <100.00%> (+<0.01%)	⬆️
lightstep/sdk/metric/view/views.go	100.00% <100.00%> (ø)
... and 1 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b60529d...0a34d83. Read the comment docs.