feat(frames): add dataframe conversion foundation

[suhailkakar]

Note

Medium Risk
New public API surface and pagination drain behavior can surprise callers who pass paginators without limit; type inference and decimal handling affect downstream analytics correctness, though behavior is heavily tested.

Overview
Adds a new polymarket.frames layer that turns Pydantic models, Page results, and homogeneous model sequences into PyArrow, pandas, or polars tables, with optional extras (arrow, pandas, polars, quant) and MissingOptionalDependencyError when backends are missing.

The Arrow path infers column types (including nested structs/lists, decimal precision/scale, and guarded numeric promotion) and supports a per-type override registry; OrderBook is flattened to per-level rows (with market/token_id when converting multiple books). Page, Paginator, AsyncPaginator, and OrderBook gain to_arrow / to_pandas / to_polars via a lazy _frames_bridge; paginator conversion requires limit (rejecting raw paginators on the free functions) and can set df.attrs["polymarket_truncated"] when capped.

Packaging updates optional dependencies, dev test deps, pyright relaxations under src/polymarket/frames, and a broad unit test suite for conversion, overrides, order books, pagination, and missing deps.

^{Reviewed by Cursor Bugbot for commit 8ae911c. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit a4a9601. Configure here.}