docs: introduce the DataFrame API in the README

[nielspardon]

What

Adds a Building plans with the DataFrame API section to the README, leading
with the ergonomic substrait.dataframe API (introduced in #204), and reframes
the existing raw substrait.proto / substrait.builders examples as the
low-level API.

Previously the README led straight into verbose raw-proto plan construction,
with no mention of the DataFrame API that already ships on main.

Why a separate PR

These README changes are extracted from the user guide PR (#214) so they can
land independently while that larger guide is being finalized. Everything here
is self-contained against main:

• The DataFrame snippet is the module's own docstring example and was run
  end to end (it produces a substrait.proto.Plan).
• The extensions extra referenced for function resolution exists in
  pyproject.toml.

The guide-link paragraph from #214 (pointing at docs/index.md and
pixi run docs-serve) is intentionally omitted here, since those docs and
tasks only exist in #214 — including it now would leave a broken link on main.
#214 re-adds that pointer when the guide lands.

Verification

Ran the new README example against the current API:

import substrait.dataframe as sub

plan = (
    sub.read_named_table("people", {"id": sub.i64, "age": sub.i64, "name": sub.string})
    .filter(sub.col("age") > 25)
    .with_columns(adult=sub.col("age") >= 18)
    .select("id", "name", "adult")
    .to_plan()
)
# -> substrait.proto.Plan

🤖 Generated with AI