feature provide column headers via parameter for asciitable

[DEvil0000]

I would like to provide column headers via a parameter for asciitable.
also defining a separator char might be helpfull.

example command:
cat $file | jc --asciitable --headers="a b" --separator=" "

example input file:
aaaaaa xyz
bbbbb uvw

example output:
[{"a": "aaaaaa", "b": "xyz"},{"a": "bbbbb", "b": "uvw"}]

[snoblenet]

Indeed.

Either that, or a new parser for headerless ASCII tables that returns an array of arrays:

[
  ['row_1_col_1', 'row_1_col_2', 'row_1_col_3'],
  ['row_2_col_1', 'row_2_col_2', 'row_2_col_3'],
  ['row_3_col_1', 'row_3_col_2', 'row_3_col_3'],
]

Use cases would include:

mas list | jc --headerless-asciitable

[snoblenet]

We'd need a robust, generalisable solution in Python, but here's a naive, MAS-specific implementation as a shellscript:

LIST=$(
  mas list |
    awk '
      {
        id = $1
        match($0, /\([0-9]/)
        version = substr($0, RSTART)
        gsub(/\(/, "",version)
        gsub(/\)/, "",version)
        name = substr($0, index($0, $2), RSTART - index($0, $2) - 1)
        gsub(/^[ \t]+|[ \t]+$/, "", name)
        print "[\"" id "\", \"" name "\", \"" version "\"],"
      }' |
    sed '$s/,\s*$//'
)

echo "[\n$LIST\n]" | jq -s '[.[][] | {id: .[0], name: .[1], version: .[2]}]'

The naive, MAS-specific implementation returns an object like this:

[
  {
    "id": "1352778147",
    "name": "Bitwarden",
    "version": "2025.4.2"
  },
  {
    "id": "6448461551",
    "name": "Command X",
    "version": "1.5.1"
  },
  {
    "id": "409222199",
    "name": "Cyberduck",
    "version": "9.1.4"
  },
  {
    "id": "1572239625",
    "name": "EPUB Viewer Pro",
    "version": "5.1.2"
  },
  {
    "id": "880764359",
    "name": "Export for iTunes",
    "version": "3.7.6"
  },
  {
    "id": "1289583905",
    "name": "Pixelmator Pro",
    "version": "3.6.18"
  },
  {
    "id": "1611378436",
    "name": "Pure Paste",
    "version": "1.12.1"
  },
  {
    "id": "1522267256",
    "name": "Shareful",
    "version": "1.11.0"
  },
  {
    "id": "507257563",
    "name": "Sip",
    "version": "4.5.2"
  },
  {
    "id": "803453959",
    "name": "Slack",
    "version": "4.44.60"
  },
  {
    "id": "1666327168",
    "name": "Spaced",
    "version": "1.2.0"
  },
  {
    "id": "6636491400",
    "name": "Text Capture",
    "version": "2.0"
  },
  {
    "id": "425424353",
    "name": "The Unarchiver",
    "version": "4.3.9"
  },
  {
    "id": "1463298887",
    "name": "Userscripts",
    "version": "4.7.1"
  },
  {
    "id": "1607635845",
    "name": "Velja",
    "version": "2.1.1"
  },
  {
    "id": "310633997",
    "name": "WhatsApp",
    "version": "25.15.75"
  },
  {
    "id": "497799835",
    "name": "Xcode",
    "version": "16.3"
  }
]

[kellyjonbrazil]

One way to do this would be by using environment variables. This way we don't need to add arguments to jc:

$ JC_ASCIITABLE_COLS="column1,column2,column3" cat $file | jc --asciitable

We would just need to add env variable functionality to the existing parser code or create new, similar parsers with the capability.

[kellyjonbrazil]

In fact, the parse() function in lib.py already accepts *kwargs, so when using as a library you could call the parser with parser-specific arguments:

jc/jc/lib.py

Lines 392 to 399 in ac8120e

	def parse(
	parser_mod_name: Union[str, ModuleType],
	data: Union[str, bytes, Iterable[str]],
	quiet: bool = False,
	raw: bool = False,
	ignore_exceptions: Optional[bool] = None,
	**kwargs
	) -> Union[JSONDictType, List[JSONDictType], Iterator[JSONDictType]]:

So the parser's parse() function could be extended to use a new kwarg for something like this. Within the function you would allow it to accept the passed value or pull it from the environment variable for jc command-line support. It would be a heavy lift to add additional argument support in the jc command-line so I think environment variables would be the best solution there.

[snoblenet]

Promising. It's years since I've done anything meaningful with Python so I won't volunteer my own efforts on this one, if that's OK.