feat(python/sedonadb): add aggregate expression methods on Expr

[jiayuasu]

First step toward grouping/aggregation, per the reprioritization recorded on #791 (aggregation/join/UDFs ahead of the remaining small schema ops).

What's new

Aggregate builder methods on Expr:

col("x").sum()      # Expr(sum(x))
col("x").count()    # Expr(count(x))
col("x").mean()     # Expr(avg(x))
col("x").min()      # Expr(min(x))
col("x").max()      # Expr(max(x))

• Methods on Expr (matching Polars / PySpark) rather than free functions — consistent with the operator-method pattern already on Expr.
• mean maps to DataFusion's avg aggregate but is named for the pandas/Polars vocabulary.
• Each wraps the corresponding helper from datafusion::functions_aggregate::expr_fn.

Scope

This PR ships only the expression builders. The resulting Expr is an aggregate AST node — valid only inside an aggregation context (DataFrame.agg(...) / group_by().agg()), which come in follow-up PRs. Shipping the builders in isolation mirrors how the original Expr foundation (#807) landed before any DataFrame integration existed; tests are repr-level.

Planned follow-ups:

1. df.agg(*exprs) — global (ungrouped) aggregation.
2. df.group_by(*keys).agg(*exprs) — the GroupedDataFrame layer.

Test plan

6 tests in tests/expr/test_expression.py:

• Parametrized exact-repr() check for all five aggregates.
• Aggregate over a compound expression ((col("x") + col("y")).sum() → Expr(sum(x + y))).

Local: 54 expression tests + 15 doctests + ruff format + ruff check all clean.

[Copilot]

Pull request overview

Adds aggregate expression builder methods (sum, count, mean, min, max) to the Python Expr type, wrapping DataFusion's aggregate function helpers. This is a small, isolated step toward the upcoming DataFrame.agg / group_by().agg() integration tracked in #791.

Changes:

• Rust PyExpr gains five aggregate methods that wrap datafusion::functions_aggregate::expr_fn helpers.
• Python Expr exposes corresponding methods with docstring examples; mean maps to DataFusion's avg.
• Tests verify repr() output for each aggregate and for an aggregate over a compound expression.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
python/sedonadb/src/expr.rs	Adds sum/count/mean/min/max PyO3 methods on PyExpr delegating to DataFusion aggregate helpers.
python/sedonadb/python/sedonadb/expr/expression.py	Adds the Python-side Expr aggregate methods with docstring examples.
python/sedonadb/tests/expr/test_expression.py	Parametrized repr tests for the five aggregates plus a compound-expression case.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[jiayuasu]

Closing in favor of the approach in #885 — exposing scalar and aggregate UDFs through a registry-walking sd.funcs.<name>(args) dispatch rather than hand-rolling per-function methods on Expr. That mechanism:

• avoids growing the Rust bindings by one method per aggregate,
• automatically surfaces plugin and Python-registered UDFs that a hard-coded list would miss,
• uses one path for scalar (spatial st_*) and aggregate functions instead of two parallel mechanisms.

Once #885 lands I'll build DataFrame.agg(...) and DataFrame.group_by(...).agg(...) on top, with call sites like sd.funcs.sum(col("x")). The repr-level tests from this PR will move to that follow-up.