Count infinite histogram samples once in edge buckets

[sylvesterkaczmarek]

Fixes #5360.

Description

Remove the extra degenerate infinity bucket from get_std_histogram_buckets: those samples have already been included in the first or last bucket. Make the positive-infinity check independent of the negative-infinity check so a one-bin histogram can include both endpoints.

The result retains the requested number of bins and counts each infinite sample once. Finite input handling, explicit-range behavior and NaN exclusion remain unchanged. Automatic range inference with infinite inputs is not changed.

Validation

Ten new cases cover negative/positive infinities separately and together, one-bin and all-infinite inputs, finite controls, NaNs, out-of-range finite values, empty inputs, read-only strided views and nonmutation. A dataframe-to-global-histogram check verifies two client datasets produce counts [3, 1, 3], totaling their seven nonmissing samples.

• New tests against unchanged main: 7 failed, 3 passed.
• Fixed revision, complete common-statistics test directory: 83 passed, 12 skipped.
• Scoped equivalents of the style checks in ./runtest.sh -s passed on both changed files: Black, isort and flake8. git diff --check passed.

python -m pytest -o addopts= -q tests/unit_test/app_common/statistics/

Tested on macOS with Python 3.12.11 and NumPy 2.5.3. The full runtest.sh suite and a deployed federated job were not run. No dependency, workflow or configuration changes.

Types of changes

• Non-breaking correction to histogram counts.
• New regression tests added.
• Targeted statistics tests and scoped style checks passed locally.

[sylvesterkaczmarek]

@yhwen @IsaacYangSLA The histogram regression tests and common-statistics suite pass locally (83 passed, 12 skipped), along with scoped style checks. Could you review this and trigger /build when appropriate?

[greptile-apps]

Confidence Score: 4/5

[Medium risk] Fixes histogram bucket counting for infinite values.

The PR should not merge until global histograms retain counts from clients with differing infinity presence.

Findings

1.  Mixed clients lose histogram counts ▶

Summary

The PR removes a duplicate infinity bucket, counts both infinities in a one-bin histogram, and adds local and two-client regression tests.

• The global histogram correction remains incomplete when clients have different infinity patterns.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart LR
    A["Client A: infinite outer edges"] --> M["Aggregate by exact bin range"]
    B["Client B: finite outer edges"] --> M
    M --> D["Unmatched edge counts discarded"]

Loading

_{Reviews (1) · Last reviewed commit: "Count infinite histogram samples once in..."}

[greptile-apps]

Mixed clients lose histogram counts

When one client has infinite values and another does not, their outer bins have different boundaries. For example, with three bins over (-1, 1), [-inf, 0, inf] produces infinite outer edges, while [-1, 0, 1] produces finite ones. accumulate_hists matches bins by exact boundaries and keeps only the first client's bins, so the combined histogram counts four of the six samples. The new two-client test gives both clients the same infinity pattern and does not catch this case.

Knowledge Base Used: Built-in application components