Broaden Polars input support via narwhals normalization

[Copilot]

TimeCopilot accepted Polars only in limited paths. This extends Polars acceptance across public APIs while clarifying that inputs are normalized to pandas for downstream model compatibility.

• Library-wide dataframe normalization

  • Apply to_pandas at all model and forecaster entry points (stats, ML, neural, foundation, adapters, ensembles, parallel/gluonts utilities) to accept Polars and other narwhals-backed frames consistently.

• Docs update

  • Clarify that Polars inputs are supported but normalized to pandas before modeling.

• Polars-focused validation

  • Strengthen tests around narwhals conversion, fallback behavior, and dataset validation.

import polars as pl
from timecopilot import TimeCopilot

df = pl.DataFrame({"unique_id": ["a"], "ds": ["2024-01-01"], "y": [1.0]})
tc = TimeCopilot()
fcst = tc.forecast(df=df, h=7, freq="D")

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.