`luau`: add advanced Excel writer helper functions about qsv

[ondohotola]

Very helpful Message ID: ***@***.***>

[jqnatividad]

Include a reference recipe as well that creates a formatted Excel with:

• Sheet 1 named "Data" - containing the "prettified" data with conditional formatting for alternating row colors and outlier highlighting. Also each column can have sparklines just below/above the column name showing its distribution.
• Sheet 2 named "Data Dictionary" with an annotated Data Dictionary
• Sheet 3 named "Summary Statistics" - with Stats
• Sheet 4 named "Frequency" with Frequency tables
• Sheet 5 named "Charts" with different "smart" visualizations using stats, frequency, etc.

[jqnatividad]

Per DOI-DO/dcat-us#131, also add a "Machine-readable" sheet as the last sheet in the workbook.

The Machine-readable sheet will have all the relevant metadata elements for the package/resource, and should hew closely to the configured DCAT profile (use DCAT-US 3.0 for the initial reference implementation).

The format of the "Machine-readable" sheet should be as follows:

• The first few rows of the sheet will be the "metadata header" section. The first row will be "# METADATA-BEGIN"
• All metadata elements will be key-value pairs, with each kv pair on its own line, with the key and value separated by a colon. The line starts with the "#" symbol (e.g. # Title:"The title of the dataset")
• The "metadata header" section ends with a line with "# METADATA-END"

After the metadata header section, a "normal", RFC-4180 compliant CSV file follows. The CSV file will be the machine-readable version of the "Data" in Sheet 1.

The recipe can then be called from Datapusher+ when populating the CKAN datastore.

Data exported from CKAN can also invoke the recipe to create an "Annotated Excel export" of a dataset.