feat: IOFactory and explicit Dataclass usage

[NJManganelli]

This PR is a precursor to both parquet reading being re-enabled (pre-processing data, most importantly) and column-joining. This exposes and expands the explicit Dataclasses, and provides an IOFactory which can convert between the currently-used dictionaries and explicit dataclass versions of them. preprocess is updated to handle dataclasses explicitly. DatasetJoinableSpec needs to be compatible with preprocess (and upcoming PR feature, preprocess_parquet) so it is upstreamed from the column-join repo

These dataclasses are especially crucial for column-joining, as creating a joinable dataset spec becomes complicated when one must combine two disparate datasets (with different forms, if one wishes to understand their structure and determine necessary columns automatically) AND a non-trivial join specification (which tells column-joining how to handle the inputs to deliver a joined dataset to the user), and that spec needs to be preprocessed (separately for root and parquet sub-datasets to be joined) then recomposed into a joinable (explicit) version of the joinable spec

We might be able to utilize these in the future to simplify some code (at the expense of this IOFactory code, and dual-paths in the preprocess function) by always coercing user inputs to dataclasses (and converting back, where appropriate)

Still could use some tests and examples in the docs, hopefully Copilot can provide these.

Serialization of these is not yet considered, may require some custom methods to be enabled (which might suggest pydantic instead?), but as-is they can always be converted to json-friendly dictionaries and vice-versa after deserialization.

[Copilot]

Pull Request Overview

This PR introduces IOFactory and explicit dataclass structures to handle dataset specifications, transitioning from dictionary-based representations to strongly-typed dataclasses. This is a preparatory change for enabling parquet reading and column-joining functionality.

• Adds new dataclass structures for different file formats (ParquetFileSpec, CoffeaParquetFileSpec variants)
• Introduces IOFactory class to convert between dictionaries and dataclass representations
• Updates preprocess function to handle both dictionary and dataclass inputs

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 5 comments.

File	Description
src/coffea/dataset_tools/preprocess.py	Adds new dataclass definitions, IOFactory class, and updates preprocess function to handle dataclass inputs
src/coffea/dataset_tools/init.py	Exports new dataclass types and IOFactory for public API

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}