Create autocomplete+ suggestions from CSV files

This package adds contents of user-defined files to the autocomplete+ suggestions list.

Suggestions

Suggestions are loaded from a user-defined csv-file (custom delimiters possible) with the column headers as indicated in bold:

text (required): the replacement text
display: main UI text for the suggestion
left: text to be shown left of the display text
right: text to be shown right of the display text
className: CCS class names
icon: the octicons icon to be used (without the icon- prefix)
description: a description to be displayed on the bottom of the autocomplete box
url: the URL linking to further information; is only shown when description is given

See also the autocomplete+ documentation on suggestions.

Configuration

All autocomplete+ options but disableForSelector and onDidInsertSuggestion can be configured on a by file basis. The GUI supports two files at the moment, but new files can easily added with new provider-entries in the autocomplete-from-file section in the config.cson. Default values are applied to these as well.

All further settings are explained in the Settings panel; please be aware that many settings provided for autocomplete+ will only apply after a restart, as I have not figured out yet how to reload a provider.

Examples

text,icon,left,display
Bob goes home.,home,to home,Bob's Action
Alice has the keys.,key,has keys,Alice's Action

If the prefix --[A-Za-z]*$ is matched --Al shows what is indicated in the screenshot, and would be replaced with Alice has the keys.

Bob-and-Alice example.

On a more serious note, I use this to load a file with pandoc-crossref references. The references are generated via a grep -rnP ' {#' | sed -re ... script from all files in a project directory, so I can easily access them. The loaded csv-file looks similar to this:

text	icon	left	display	right	class
sec:introduction	bookmark	sec:introduction	Introduction	introduction.md	header-1
tbl:example	list-unordered	tbl:example	Example table	another_file.md	table
fig:example	file-media	fig:example	Example Figure	third-file.md	figure

And these are the corresponding settings; the match prefix triggers the completion only when a colon occurs after an @ to distinguish them from literature references. The first matching group until the @ is retained when selecting an auto completion:

provider1:
  enabled: true
  fileProvided: "/path/to/file.csv"
  matchPrefix: "((?:^|[ ;\\[])@)([\\w0-9_-]*:[\\w0-9_-]*)$"
  retainFirstCapture: true
  numberOfSuggestions: 10
  selector: ".text.md, .text.plain"