Add $isNumber compatibility tests (#212)

[xMayble]

What does this PR do?

Adds compatibility test coverage for the $isNumber aggregation expression operator. Three new test files exercise the operator against numeric BSON types (int32, int64, double, and Decimal128 — including the NaN, Infinity, and negative-zero edge cases), non-numeric BSON types (string, bool, array, object, ObjectId, Date, Timestamp, Binary, Regex, MinKey/MaxKey, Code), and null/missing field inputs. Every case is run both as a literal expression and via a document field reference.

Why was this PR needed?

Issue #212 requests a second-pass compatibility test for $isNumber. Only a single smoke test (test_smoke_isNumber.py) existed; there was no systematic coverage of the numeric-vs-non-numeric type boundary or of null/missing inputs, which is where $isNumber behavior is most easily mis-implemented across engines.

What are the relevant issue numbers?

Closes #212

Does this PR meet the acceptance criteria?

• Tests added for new/changed behavior
• Follows project style guide (black-formatted, tagged @pytest.mark.aggregate, uses pytest_params / BaseTestCase / execute_expression* helpers)
• No breaking changes introduced (purely additive — new files only)
• All tests passing (validated locally to compile/collect; relying on CI to run against MongoDB)

[xMayble]

Hey, this is my first contribution to the project. This PR addresses #212 by adding $isNumber compatibility coverage across numeric types, non-numeric types, and null/missing inputs. Would appreciate a review when you have time. Thanks!

[documentdb-triage-tool]

🤖 Auto-triaged by documentdb-triage-tool.

Applied: compatibility test, enhancement
Project fields suggested: Component test-coverage · Priority P2 · Effort M · Status Needs Review
Confidence: 0.30 (deterministic)

Reasoning

component from path globs (test-coverage); effort from diff stats (321+0 LOC, 3 files); LLM failed: Invalid response body while trying to fetch https://api.anthropic.com/v1/messages: Premature close

If a label is wrong, remove it manually and ping @patty-chow so the rules can be tuned. The bot will not re-label items that already have component labels.