Export canonical viz-setting key types/constants for downstreams

[lmeyerov]

Request: Export canonical Graphistry viz-setting key definitions for downstream type safety

Downstream projects (example: Louie/GraphistryGPT) need to pass and validate Graphistry visualization settings in a typed way, including:

• URL params (embed / graph.html query options)
• React-facing settings used by @graphistry/client-api-react

Today, downstreams often end up hardcoding local key unions / allowlists (e.g., Python Literal[...] + runtime JSON checks), which creates drift risk and maintenance overhead.

Proposed additions in pygraphistry

Please expose a canonical, machine-readable key surface, ideally one or more of:

1. Exported Python constants

   • URL_PARAM_NAMES: tuple[str, ...]
   • REACT_SETTING_NAMES: tuple[str, ...] (if pygraphistry has/owns that view)

2. Typed schemas/models

   • Optional TypedDict / pydantic model for settings payloads
   • Or a JSON Schema artifact that downstreams can consume for codegen

3. Versioned contract file

   • Stable JSON file in package data with explicit key sets and basic value-shape metadata

Why this helps

• Eliminates duplicated key lists across downstream repos
• Improves static typing/autocomplete and safer LLM tool-contract synthesis
• Enables single-source-of-truth tests across Python/TS stacks

Related upstream discussion in graphistry-js:

• Export react bindings + url param names from @graphistry/client-api-react graphistry-js#194

If useful, we can contribute a PR once maintainers confirm preferred API shape.