fix(data): remove MinMaxPyramid::reduce, repoint its two test callers

The pyramid's public query surface is the bucket-grained trio
(reduce_buckets/sum_buckets/sumsq_buckets); reduce() was the one
non-*_buckets query and silently over-covered non-aligned ranges, which
is exactly the misuse the exactness split exists to prevent. Exact range
queries stay on Series.

- delete the declaration, its doc paragraph and the definition; no
  helper became dead (fold_range/cover_buckets/ceil_to_bucket are all
  still used by the *_buckets queries)
- pyramid_clear_and_rebuild: complete bucket 0 and assert its extrema via
  reduce_buckets(0, kBucket) -- a pyramid has no exact sub-bucket query
- stream_series_acquire_publishes_snapshot: s.pyramid().reduce(0, 100)
  -> s.reduce(0, 100), keeping the == 99.0 assertion

src/data/minmax_pyramid.cpp

@@ -221,15 +221,6 @@ f64 MinMaxPyramid::sumsq_buckets(u64 begin, u64 end) const {
   return total;
 }
 
-Range MinMaxPyramid::reduce(u64 begin, u64 end) const {
-  const u64 e = std::min(end, size_);
-  const u64 b = std::min(begin, e);
-  if (e <= b) return Range{0.0, 0.0};
-  const u64 b0 = b / kBucket;                   // first bucket holding a sample
-  const u64 b1 = (e - 1) / kBucket + 1;         // one past the last one
-  return fold_range(*this, b0, b1);
-}
-
 u32 MinMaxPyramid::level_for(u64 n, f64 samples_per_pixel) const {
   // The ratio test also rejects NaN, and the clamp keeps a huge (or infinite)
   // samples-per-pixel from making the floor-to-u32 conversion ill-formed.