Do you need the sunspot dataset?

[Nimrais]

Hi, I don't know what are the rules for the contributions sorry. I haven't not found it in the documentation.

Thanks for the nice package!

Do you need the sunspot dataset? It's a dataset of the sunspots over the years.

const SUNSPOTS_DEPNAME = "Sunspots"
const SUNSPOTS_LINK = "https://raw.githubusercontent.com/STAT-JET-ASU/Datasets/master/Instructor/"
const SUNSPOTS_DOCS = "https://stat-jet-asu.github.io/Datasets/InstructorDescriptions/sunspots.html"
const SUNSPOTS_FILE = "sunspots.csv"
const SUNSPOTS_SHA256 = "657169832a67571b46385046a2966a6056890ad22da82c26f7a26e615adde840"

function __init__sunspots()
    register(DataDep(SUNSPOTS_DEPNAME,
                     """
                     Dataset: Monthly mean relative sunspot numbers (1749-).
                     Website: $SUNSPOTS_DOCS
                     """,
                     SUNSPOTS_LINK * SUNSPOTS_FILE,
                     SUNSPOTS_SHA256))
end

"""
    Sunspots(; as_df = true, dir = nothing)

Monthly mean relative sunspot numbers from 1749 onwards, a classic
univariate time-series dataset.

Each row reports, for one month, the average daily sunspot number together
with its standard deviation. The data is downloaded on first use from the
[STAT-JET-ASU datasets collection]($(SUNSPOTS_DOCS)) and cached locally
via DataDeps.jl.

Columns:

- `year`: calendar year of the observation.
- `month`: month of the year (1-12).
- `date`: observation date as a fractional year.
- `average`: monthly mean relative sunspot number — used as the target.
- `sd`: standard deviation of the daily values within the month.

NOTE: this is a time series with no pre-defined train-test split.

# Arguments

- If `as_df = true` (default), `features`/`targets` are `DataFrame`s, and
  `dataframe` holds the full table; otherwise they are plain `Matrix`es and
  `dataframe` is `nothing`.
- `dir`: directory to load/download into. Defaults to the DataDeps cache.

# Fields

- `metadata`: dictionary with `path`, `n_observations`, `feature_names`,
  `target_names`.
- `features`: the `year`, `month`, `date`, `sd` columns.
- `targets`: the `average` column.
- `dataframe`: the full table, or `nothing` when `as_df = false`.

The full implementation attached in the file below, github does not render the full code in a readiable way for whatever reason.

sunspots.jl.txt