[feature] Initial OceanOSSE workflow skeleton.

[oj-tooth]

Skeleton OceanOSSE End-to-End Workflow

Milestone: M0 - Foundation Infrastructure

Summary

• Establish the structural skeleton of the OceanOSSE package by implementing a minimal, pass-through end-to-end workflow.
• All core modules and abstract base classes are scaffolded and assembled together in pipeline.py and can be executed using OceanOSSE CLI (cli.py.
• Basic Test: Example ocean model outputs stored in a netCDF input file should be able to traverse the entire OceanOSSE pipeline and be re-written to a netCDF output file.
• No actual transformation applied at any stage (i.e., identity transformations implemented).
• Purpose is to establish initial module interfaces and orchestration logic before any scientific functionality is implemented.

Scope

Following components should be stubbed and connected:

• cli.py — add a run command that accepts a .toml config path, optional log file path and invokes the pipeline.
• pipeline.py — run and describe functions that instantiate and sequences all functions in order.
• io/reader.py — functions to opens a netCDF file and returns xarray.Dataset.
• sampling/base.py — ObsBaseSampler ABC with a sample(dataset) method that returns the dataset unchanged.
• error/base.py — InstrumentBaseError ABC with an apply(dataset) method that returns the dataset unchanged.
• gridding/base.py — ObsBaseRegridder ABC with a grid(dataset) method that returns the dataset unchanged.
• io/writer.py — functions to write the dataset to a netCDF output file.
• config/validation.py — a minimal pydantic .toml schema covering input path, output path, and placeholder sections for each module.