PandocAnsiFilter/ansi2html.lua

-- this is essentially 
-- https://github.com/jupyter/nbconvert/blob/main/nbconvert/filters/ansi.py
-- converted to lua

-- good list of ANSI escape sequences:
-- https://gist.github.com/fnky/458719343aabd01cfb17a3a4f7296797

local css = [[
<!-- CSS added by ANSI escape sequences filter -->
<style>
.ansi {line-height:1; padding:0; margin: 0; letter-spacing:0;font-family:JuliaMono;}

/* console foregrounds and backgrounds */
pre .ansi-black-fg { color: #3e424d; }
pre .ansi-red-fg { color: #e75c58; }
pre .ansi-green-fg { color: #00a250; }
pre .ansi-yellow-fg { color: #ddb62b; }
pre .ansi-blue-fg { color: #208ffb; }
pre .ansi-magenta-fg { color: #d160c4; }
pre .ansi-cyan-fg { color: #60c6c8; }
pre .ansi-white-fg { color: #c5c1b4; }
pre .ansi-black-bg { background-color: #3e424d; }
pre .ansi-red-bg { background-color: #e75c58; }
pre .ansi-green-bg { background-color: #00a250; }
pre .ansi-yellow-bg { background-color: #ddb62b; }
pre .ansi-blue-bg { background-color: #208ffb; }
pre .ansi-magenta-bg { background-color: #d160c4; }
pre .ansi-cyan-bg { background-color: #60c6c8; }
pre .ansi-white-bg { background-color: #c5c1b4; }
pre .ansi-black-intense-fg { color: #282c36; }
pre .ansi-red-intense-fg { color: #b22b31; }
pre .ansi-green-intense-fg { color: #007427; }
pre .ansi-yellow-intense-fg { color: #b27d12; }
pre .ansi-blue-intense-fg { color: #0065ca; }
pre .ansi-magenta-intense-fg { color: #a03196; }
pre .ansi-cyan-intense-fg { color: #258f8f; }
pre .ansi-white-intense-fg { color: #a1a6b2; }
pre .ansi-black-intense-bg { background-color: #282c36; }
pre .ansi-red-intense-bg { background-color: #b22b31; }
pre .ansi-green-intense-bg { background-color: #007427; }
pre .ansi-yellow-intense-bg { background-color: #b27d12; }
pre .ansi-blue-intense-bg { background-color: #0065ca; }
pre .ansi-magenta-intense-bg { background-color: #a03196; }
pre .ansi-cyan-intense-bg { background-color: #258f8f; }
pre .ansi-white-intense-bg { background-color: #a1a6b2; }
pre .ansi-default-inverse-fg { color: rgba(255, 255, 255, 1); }
pre .ansi-default-inverse-bg { background-color: #111111; }
pre .ansi-bold { font-weight: bold; }
pre .ansi-underline { text-decoration: underline; }
</style>
]]

local latexpreamble = [[
\usepackage{fancyvrb} 
\usepackage{fontspec}
\usepackage{xcolor}   
\renewenvironment{verbatim}{%
  \VerbatimEnvironment
  \begin{Verbatim}[commandchars=\\\{\}]%
}{%
  \end{Verbatim}%
}
\setmonofont{JuliaMono}[
  Scale = MatchLowercase,
  UprightFont = *-Regular,
  BoldFont = *-Bold,
  ItalicFont = *-RegularItalic,
  BoldItalicFont = *-BoldItalic,
  ]
  \definecolor{ansi-black}{HTML}{3E424D}
  \definecolor{ansi-black-intense}{HTML}{282C36}
  \definecolor{ansi-red}{HTML}{E75C58}
  \definecolor{ansi-red-intense}{HTML}{B22B31}
  \definecolor{ansi-green}{HTML}{00A250}
  \definecolor{ansi-green-intense}{HTML}{007427}
  \definecolor{ansi-yellow}{HTML}{DDB62B}
  \definecolor{ansi-yellow-intense}{HTML}{B27D12}
  \definecolor{ansi-blue}{HTML}{208FFB}
  \definecolor{ansi-blue-intense}{HTML}{0065CA}
  \definecolor{ansi-magenta}{HTML}{D160C4}
  \definecolor{ansi-magenta-intense}{HTML}{A03196}
  \definecolor{ansi-cyan}{HTML}{60C6C8}
  \definecolor{ansi-cyan-intense}{HTML}{258F8F}
  \definecolor{ansi-white}{HTML}{C5C1B4}
  \definecolor{ansi-white-intense}{HTML}{A1A6B2}
  \definecolor{ansi-default-inverse-fg}{HTML}{FFFFFF}
  \definecolor{ansi-default-inverse-bg}{HTML}{000000}
]]

local ANSI_COLORS = {
    "ansi-black",
    "ansi-red",
    "ansi-green",
    "ansi-yellow",
    "ansi-blue",
    "ansi-magenta",
    "ansi-cyan",
    "ansi-white",
    "ansi-black-intense",
    "ansi-red-intense",
    "ansi-green-intense",
    "ansi-yellow-intense",
    "ansi-blue-intense",
    "ansi-magenta-intense",
    "ansi-cyan-intense",
    "ansi-white-intense"
}

local flag = false -- set to true if we find ANSI sequences, used by Meta

local function get_extended_color(numbers)
    local n = table.remove(numbers, 1)
    local r,g,b,idx
    if n == 2 and #numbers >=3 then
        -- 24bit RGB
        r = table.remove(numbers, 1)
        g = table.remove(numbers, 1)
        b = table.remove(numbers, 1)
    elseif n == 5 and #numbers >= 1 then
        -- 256 colors
        idx = table.remove(numbers, 1)
        if idx < 16 then
            -- 16 default terminal colors
            return idx
        elseif idx < 232 then
            -- 6x6x6 color cube, see http://stackoverflow.com/a/27165165/500098
            r = (idx - 16) // 36
            r = 55 + r * 40
            if r < 0 then r = 0 end
            g = ((idx - 16) % 36) // 6
            g = 55 + g * 40
            if g < 0 then g = 0 end
            b = (idx - 16) % 6
            b = 55 + b * 40
            if b < 0 then b = 0 end
        elseif idx < 256 then
            -- grayscale, see http://stackoverflow.com/a/27165165/500098
            r = (idx - 232) * 10 + 8
            g = r
            b = r
        end
    end
    return {r, g, b}
end


--[=[
local re = require "re"

local ANSI = re.compile [[
    '\x1b%[' {.*?} {[@-~]} 
    ]]

--]=]

local function LaTeXconverter(fg, bg, bold, underline, inverse)
    if not (fg or bg or bold or underline or inverse) then
        return "",""
    end

    local starttag = ""
    local endtag = ""

    if inverse then
        fg, bg = bg, fg
    end

    if type(fg) == "number" then
        starttag = starttag .. [[\textcolor{]] .. ANSI_COLORS[fg+1] .. "}{"
        endtag = "}" .. endtag
    elseif type(fg) == "table" then
        -- See http://tex.stackexchange.com/a/291102/13684
        starttag = starttag .. [[\def\tcRGB{\textcolor[RGB]}\expandafter]]
        starttag = starttag .. string.format([[\tcRGB\expandafter{\detokenize{%d,%d,%d}}{]], fg[1], fg[2], fg[3])
        endtag = "}" .. endtag
    elseif inverse then
        starttag = starttag .. [[\textcolor{ansi-default-inverse-fg}{]]
        endtag = "}" .. endtag
    end

    if type(bg) == "number" then
        starttag = starttag .. [[\setlength{\fboxsep}{0pt}]]
        starttag = starttag .. [[\colorbox{]] .. ANSI_COLORS[bg+1] .. "}{"
        endtag = [[\strut}]] .. endtag
    elseif type(bg) == "table" then
        -- See http://tex.stackexchange.com/a/291102/13684
        starttag = starttag .. [[\setlength{\fboxsep}{0pt}]]
        starttag = starttag .. [[\def\cbRGB{\colorbox[RGB]}\expandafter]]
        starttag = starttag .. string.format([[\cbRGB\expandafter{\detokenize{%d,%d,%d}}{]], bg[1], bg[2], bg[3])
        endtag = [[\strut}]] .. endtag
    elseif inverse then
        starttag = starttag .. [[\setlength{\fboxsep}{0pt}]]
        starttag = starttag .. [[\colorbox{ansi-default-inverse-bg}{]]
        endtag = [[\strut}]] .. endtag
    end

    if bold then
        starttag = starttag .. [[\textbf{]]
        endtag  = "}" .. endtag
    end

    if underline then
        starttag = starttag .. [[\underline{]]
        endtag  = "}" .. endtag
    end
    return starttag, endtag
end

local function HTMLconverter(fg, bg, bold, underline, inverse)
    if not (fg or bg or bold or underline or inverse) then
        return "",""
    end
    local classes = {}
    local styles  = {}
    local type = type  -- more efficient? 
    local next = next
    if inverse then
        fg, bg = bg, fg
    end

    if type(fg) == "number" then
        table.insert(classes, ANSI_COLORS[fg+1] .. "-fg")
    elseif type(fg) == "table" then
        table.insert(styles, string.format("color: rgb(%d,%d,%d)", fg[1], fg[2], fg[3]))
    elseif inverse then
        table.insert(classes, "ansi-default-inverse-fg")
    end

    if type(bg) == "number" then
        table.insert(classes, ANSI_COLORS[bg+1] .. "-bg")
    elseif type(bg) == "table" then
        table.insert(styles, string.format("background-color: rgb(%d,%d,%d)",
            bg[1], bg[2], bg[3]))
    elseif inverse then
        table.insert(classes, "ansi-default-inverse-bg")
    end

    if bold then
        table.insert(classes, "ansi-bold")
    end

    if underline then
        table.insert(classes, "ansi-underline")
    end

    local starttag = "<span"
    if next(classes) ~= nil   then
        starttag = starttag .. ' class="' ..  table.concat(classes, " ") .. '"'
    end

    if next(styles) ~= nil then
        starttag = starttag .. ' style="' ..  table.concat(styles, " ") .. '"'
    end

    return starttag..">","</span>"
end

   
local function codeBlockTrans(e)
    local converter, fmt
    if FORMAT:match 'latex' then
        converter = LaTeXconverter
        fmt = 'latex'
    elseif FORMAT:match 'html' then
        converter = HTMLconverter
        fmt = 'html'
    else
        return
    end
    local out=""
    if string.find(e.text, "\x1b%[") then
        flag = true
        local bold = false
        local underline = false
        local inverse = false
        local text = e.text
        local chunk = ""
        local fg = nil
        local bg = nil
        local starttag = ""
        local endtag = ""
        local numbers={}

        while text ~= "" do
            numbers = {}
            local s1, e1, c1, d1 = string.find(text, "\x1b%[(.-)([@-~])")
            if s1 then
                if d1 == "m" then
                    for i in string.gmatch(c1, "([^;]*)") do
                        table.insert(numbers, tonumber(i))
                    end
                else
                    io.stderr:write("Unsupported ANSI sequence ESC["..c1..d1.." ignored\n" ) 
                end
                chunk, text = text:sub(1, s1-1), text:sub(e1+1)
            else
                chunk, text = text, ""
            end

            if chunk ~= "" then
                if bold and type(fg)=="number" and fg<8 then
                    starttag, endtag = converter(fg+8, bg, bold, underline, inverse)
                else
                    starttag, endtag = converter(fg, bg, bold, underline, inverse)
                end    
                out = out .. starttag .. chunk .. endtag
            end

            while next(numbers) ~= nil do
                local n = table.remove(numbers, 1)
                if n == 0 then
                    fg = nil
                    bg = nil
                    bold = false
                    inverse = false
                    underline = false
                elseif n == 1 then
                    bold = true
                elseif n == 4 then
                    underline = true
                elseif n == 5 then
                    bold = true -- 'blinking'
                elseif n == 7 then
                    inverse = true
                elseif n == 21 or n == 22 then
                    bold = false
                elseif n == 24 then
                    underline = false
                elseif n == 27 then
                    inverse = false
                elseif n >= 30 and n <= 37 then
                    fg = n - 30
                elseif n == 38 then
                    fg = get_extended_color(numbers)
                elseif n == 39 then
                    fg = nil
                elseif n >= 40 and n <= 47 then
                    bg = n - 40
                elseif n == 48 then
                    bg = get_extended_color(numbers)
                elseif n == 49  then
                    bg = nil
                elseif n >= 90 and n <= 97 then
                    fg = n + 8 - 90
                elseif n >= 100 and n <= 107 then
                    bg = n + 8 - 100
                else
                    io.stderr:write(string.format("ESC sequence with unknown code %d before:\n",n))
                    io.stderr:write(chunk.."\n")
                end
            end
        end
        if fmt == 'html' then
            return pandoc.RawBlock(fmt, '<pre class="ansi"><code class="ansi">'..out..'</code></pre>')
        end
        if fmt == 'latex' then
            return pandoc.RawBlock(fmt, "\\begin{verbatim}\n"..out.."\n\\end{verbatim}")
        end
    end
end

-- algo from https://github.com/jgm/pandoc/commit/77faccb505992c944cd1b92f50e4e00d2927682b
-- simulate --ipynb-outbut=best, but w/o removing ANSI colors
-- consider only formats html and latex
local function divTrans(e)
    local fmt
    if FORMAT:match 'latex' then
        fmt = 'latex'
    elseif FORMAT:match 'html' then
        fmt = 'html'
    else
        return
    end

    local function cmp(a,b) 
        return a["rank"] < b["rank"]
    end

    if e.classes[1] == "output" then
        local c = e.content
        local ranks = {}
        for i, el in pairs(c) do
            if el.t == "RawBlock" then
                if el.format == fmt then
                    table.insert(ranks, {rank = 1, el})
                else
                    table.insert(ranks, {rank = 3, el})
                end
            else
                table.insert(ranks, {rank = 2, el})
            end
        end
        table.sort(ranks, cmp)
        local winner = ranks[1][1]
        e.content = pandoc.List:new()
        table.insert(e.content, winner)
        return e
    end
end


local function metaAdd(meta)
    if flag then
        -- read current "header-includes" from metadata, or make a new one
        -- and add css to the end of "header-includes"
        local current = meta['header-includes'] or pandoc.MetaList{meta['header-includes']}
        if FORMAT:match 'html' then
            current[#current+1] = pandoc.MetaBlocks(pandoc.RawBlock("html", css))
        end
        if FORMAT:match 'latex' then
            current[#current+1] = pandoc.MetaBlocks(pandoc.RawBlock("latex", latexpreamble))
        end
        meta['header-includes'] = current
        return(meta)
    end
end

return {
    {Div = divTrans},
    {CodeBlock = codeBlockTrans},
    {Meta = metaAdd}
}