path: root/rst_context.lua

#!/usr/bin/env texlua
--------------------------------------------------------------------------------
--         FILE:  rst_context.lua
--        USAGE:  ./rst_context.lua 
--  DESCRIPTION:  
--      OPTIONS:  ---
-- REQUIREMENTS:  ---
--       AUTHOR:  Philipp Gesang (Phg), <megas.kapaneus@gmail.com>
--      VERSION:  1.0
--      CREATED:  31/08/10 19:35:15 CEST
--------------------------------------------------------------------------------
--


require "lpeg"
require "rst_helpers"

local C, Cb, Cc, Cg, Cmt, Cp, Cs, Ct, P, R, S, V, match = lpeg.C, lpeg.Cb, lpeg.Cc, lpeg.Cg, lpeg.Cmt, lpeg.Cp, lpeg.Cs, lpeg.Ct, lpeg.P, lpeg.R, lpeg.S, lpeg.V, lpeg.match

if not context then -- standard context lpeg stripper from l-string.lua
    local stripper = P{
        [1] = "stripper",
        stripper = V"space"^0 * C((V"space"^0 * V"nospace"^1)^0),
        space    = S(" \t\v\n"),
        nospace  = 1 - V"space",
    }
    function string.strip(str)
        return stripper:match(str) or ""
    end 
end

local rst_context = {}


rst_context.collected_references = {}
rst_context.collected_adornments = {}
rst_context.last_section_level   = 0
rst_context.anonymous_links      = 0
rst_context.context_references   = {}

-- So we can use crefs[n][2] to refer to the place where the reference was
-- created.
local function get_context_reference (str)
    local crefs = rst_context.context_references
    refstring = "__contextref__" .. tostring(#crefs + 1)
    crefs[#crefs + 1] = { refstring, str }
    return refstring
end

function rst_context.emphasis (str)
    return [[{\\em ]] .. str .. [[}]]
end

function rst_context.strong_emphasis (str)
    return [[{\\sc ]] .. str .. [[}]]
end

function rst_context.literal (str)
    str = str:gsub([[\]], [[\\]]) -- evade escaping of backslashes
    return [[\\type{]] .. str .. [[}]]
end


function rst_context.interpreted_text (...)
    local tab = { ... }
    --print (tab, #tab, tab[1], tab[2], tab[3])
    local role, str
    role = tab[1]:match("^:(.*):$") or tab[3]:match("^:(.*):$")
    str  = tab[2]

    if not role then -- implicit role
        role = "emphasis"
    end

    --print(role, str)

    return rst_context[role](str)
end

function rst_context.link_standalone (str)
    return "\n"
        .. [[\\goto{\\hyphenatedurl{]]
        .. str
        .. [[}}[url(]]
        .. str
        .. [=[)]]=]
end

function rst_context.reference (str)
    str = str:match("[^_]*")
    local link = rst_context.collected_references[str]
    if not link then -- TODO make warning instead
        return([[{\\sc UNDEFINED REFERENCE ]] .. str .. [[}.]])
    end
    return [[\\goto{]]
        .. str
        .. [[}[url(]]
        .. link
        .. [=[)]]=]
end

function rst_context.target (tab)
    --print("GOT ONE!")
    --local tab = { ... }
    local refs = rst_context.collected_references
    local target = tab[#tab] -- Ct + C could be clearer but who cares
    tab[#tab] = nil

    local function resolve_indirect (r)
        if r and r:match(".*_$") then -- pointing elsewhere
            return resolve_indirect (refs[r:match("(.*)_$")]) or "need another run!" -- TODO multiple runs && data collection
        end
        return r
    end

    local function create_anonymous ()
        rst_context.anonymous_links = rst_context.anonymous_links + 1
        return "__anon__" .. rst_context.anonymous_links
    end


    target = resolve_indirect (target)

    for i=1,#tab do
        local id = tab[i]:gsub("\\:",":") -- deescaping
        id = id ~= "" and id or create_anonymous ()
        refs[id] = refs[id] or target
    end
    return ""
end

function rst_context.escape (str)
    return str:gsub("\\(.)", "%1")
end

function rst_context.joinindented (tab)
    return table.concat (tab, "")
end

local inline_parser = P{
    [1] = "block",

    block = Cs((V"inline_element" + 1)^1),


    inline_element = Cs(V"precede_inline"
                    * (V"strong_emphasis"
                     + V"emphasis"
                     + V"inline_literal"
                     + V"interpreted_text"
--                   + V"inline_internal_target" -- TODO
                     + V"reference"
--                   + V"footnote_reference"     -- TODO
--                   + V"substitution_reference" -- TODO
                     + V"link_standalone")
                    * V"succede_inline"),

    space = P" ",
    whitespace = (P" " + Cs(P"\t") / "        " + Cs(S"\v") / " "),
    spacing = V"whitespace"^1,

    eol = P"\n",
    inline_delimiters = P"‐" + P"‑" + P"‒" + P"–" + V"emdash" + V"space",        -- inline markup
    --inline_delimiter = P"**" + P"``" + S"*`",
    asterisk = P"*",
    double_asterisk = V"asterisk" * V"asterisk",
    bareia = P"`",
    backslash = P"\\",
    bar = P"|",
    double_bareia = V"bareia" * V"bareia",
    escaped_bareia = (Cs(V"backslash") / "" * V"bareia") + 1,
    colon = P":",
    semicolon = P";",
    underscore = P"_",
    double_underscore = V"underscore" * V"underscore",
    dot = P".",
    interpunct = P"·",
    comma = P",",
    dash = P"-",
    emdash = P"—",
    ellipsis   = P"…" + P"...",
    exclamationmark = P"!",
    questionmark = P"?",
    interrobang = P"‽",
    double_dash = V"dash" * V"dash",
    triple_dash = V"double_dash" * V"dash",
    hyphen = P"‐",
    dashes = V"dash" + P"‒" + P"–" + V"emdash" + P"―",
    letter = R"az" + R"AZ",
    groupchars = S"()[]{}",
    apostrophe = P"’" + P"'",

    guillemets = P"«" + P"»",
    quotationmarks= P"‘" + P"’" + P"“" + P"”",
    solidus= P"⁄",
    slash = P"/",

    punctuation = V"apostrophe"
                + V"colon" 
                + V"comma" 
                + V"dashes"
                + V"dot" 
                + V"ellipsis"
                + V"exclamationmark"
                + V"guillemets"
                + V"hyphen"
                + V"interpunct"
                + V"interrobang"
                + V"questionmark" 
                + V"quotationmarks"
                + V"semicolon" 
                + V"slash"
                + V"solidus"
                + V"underscore"
                ,

    precede_inline = V"spacing"
                   + V"eol"
                   + S[['"([{<-/:]]
                   + P"‘" + P"“" + P"’" + P"«" + P"¡" + P"¿"
                   + V"inline_delimiters"
                   + P"„", -- not in standard Murkin reST

    succede_inline = V"spacing"
                   + S[['")]}>-/:.,;!?\]]
                   + P"’" + P"”" + P"»"
                   + V"inline_delimiters"
                   + P"“", -- non-standard again but who cares

    emphasis        = (V"asterisk" - V"double_asterisk") 
                    * Cs((1 - V"spacing" - V"eol" - V"asterisk")
                       * ((1 - (1 * V"asterisk"))^0 
                        * (1 - V"spacing" - V"eol" - V"asterisk"))^-1) 
                    * V"asterisk" 
                    / rst_context.emphasis,

    strong_emphasis = V"double_asterisk" 
                    * Cs((1 - V"spacing" - V"eol" - V"asterisk")
                       * ((1 - (1 * V"double_asterisk"))^0 
                        * (1 - V"spacing" - V"eol" - V"asterisk"))^-1) 
                    * V"double_asterisk"  
                    / rst_context.strong_emphasis,

    inline_literal  = V"double_bareia"
                    * C ((V"escaped_bareia" - V"spacing" - V"eol" - V"bareia")
                       * ((V"escaped_bareia" - (1 * V"double_bareia"))^0
                        * (V"escaped_bareia" - V"spacing" - V"eol" - V"bareia"))^-1)
                    * V"double_bareia"
                    / rst_context.literal,

    interpreted_text = C(V"role_marker"^-1)
                     * (V"bareia" - V"double_bareia")
                     * C ((1 - V"spacing" - V"eol" - V"bareia")
                        * ((1 - (1 * V"bareia"))^0
                         * (1 - V"spacing" - V"eol" - V"bareia"))^-1)
                     * V"bareia"
                     * C(V"role_marker"^-1)
                     / rst_context.interpreted_text,

    role_marker = V"colon" * (V"letter" + V"dash" + V"underscore" + V"dot")^1 * V"colon",

    link_standalone = C(V"uri")
                    / rst_context.link_standalone,

    reference = Cs(V"_reference")
              / rst_context.reference,

    _reference = (1 - V"underscore" - V"spacing" - V"eol" - V"punctuation" - V"groupchars")^1 * V"underscore",

--------------------------------------------------------------------------------
-- Urls
--------------------------------------------------------------------------------
    uri = V"url_protocol" * V"url_domain" * (V"slash" * V"url_path")^0,

    url_protocol = (P"http" + P"ftp" + P"shttp" + P"sftp") * P"://",
    url_domain_char = 1 - V"dot" - V"spacing" - V"eol" - V"punctuation",
    url_domain = V"url_domain_char"^1 * (V"dot" * V"url_domain_char"^1)^0,
    url_path_char = R("az", "AZ", "09") + S"-_.!~*'()",
    url_path = V"slash" * (V"url_path_char"^1 * V"slash"^-1)^1,
}

function rst_context.paragraph (tab)
    local str = inline_parser:match(table.concat(tab, " "))
    print(inline_parser:match(table.concat(tab, " ")))
    return string.format([[

\\startparagraph
%s
\\stopparagraph
]], str)
end

local sectionlevels = {
    [1] = "chapter",
    [2] = "section",
    [3] = "subsection",
    [4] = "subsubsection",
    [5] = "subsubsubsection",
}

local function get_line_pattern (chr)
    return P(chr)^1 * (-P(1))
end

function rst_context.section (...)  -- TODO general cleanup; move validity
    local tab = { ... }             -- checking to parser.
    local section, str = true, ""
    local adornchar 
    if #tab == 3 then -- TODO use unicode length with ConTeXt
        --print(">>"..tab[1].."<>"..tab[2].."<<")
        adornchar = tab[1]:sub(1,1)
        --  overline == underline && len(overline) = len(sectionstring)
        section = tab[1] == tab[3] and #tab[1] >= #tab[2] 
        -- if overline consists only of one char then keep truth value else
        -- false
        section = get_line_pattern(adornchar):match(tab[1]) ~= nil and section
        str = string.strip(tab[2])
    else -- no overline
        --print(">>"..tab[1].."<>"..tab[2].."<<")
        adornchar = tab[2]:sub(1,1)
        section = #tab[1] <= #tab[2]
        section = get_line_pattern(adornchar):match(tab[2]) ~= nil and section
        str = tab[1]
    end

    if section then -- determine level
        local level = rst_context.last_section_level
        local rca = rst_context.collected_adornments
        if rca[adornchar] then
            level = rca[adornchar]
        else
            level = level + 1
            rca[adornchar] = level
            rst_context.last_section_level = level
        end

        ref = get_context_reference (str)

        str = string.format("\n\\\\%s[%s]{%s}\n", sectionlevels[level], ref, str)
    end

    return section and str or ""
end

-- Prime time for the fancybreak module.
function rst_context.transition (str)
    return "\n\\hrule\n"
end

function rst_context.bullet_marker(str)
    return "marker"
end

do
    local space    = lpeg.S(" \t\v\n")
    local nospace  = 1 - space
    local stripper = space^0 * lpeg.C((space^0 * nospace^1)^0)
    function string.strip(str)
        return match(stripper,str) or ""
    end 
end

local enumeration_types = {
    ["*"] = "*", -- unordered bulleted
    ["+"] = "*",
    ["-"] = "*",
    ["•"] = "*",
    ["‣"] = "*",
    ["⁃"] = "*",

    ["#"] = "1", -- numbered lists and conversion
    ["A"] = "A",
    ["a"] = "a",
    ["I"] = "R",
    ["i"] = "r",
}

-- \setupitemize[left=(, right=), margin=4em, stopper=]

local stripme   = S"()."
local dontstrip = 1 - stripme
local itemstripper = stripme^0 * C(dontstrip^1) * stripme^0
local function parse_itemstring(str)
    local setup = [[\setupitemize[]]
    -- string.match is slightly faster than string.find
    if str:match("^%(") then
        setup = setup .. [[left=(,]]
    end
    if str:match("%)$") then
        setup = setup .. [[right=)]]
    end
    if str:match("%.$") then
        setup = setup .. [[stopper=.]]
    end
    setup = setup .. "]\n"

    str = itemstripper:match(str)
    return {setup = setup, str=str}
end

function rst_context.startitemize(str)
    local setup = ""
    str = string.strip(str)

    local listtype = enumeration_types[str] or parse_itemstring(str)

    if type(listtype) == "table" then
        print(type(listtype), listtype[2])
        setup = listtype.setup
        listtype = listtype.str
    end

    return setup .. [[
\\startitemize[]] .. listtype .. [[]
]] 
end

function rst_context.stopitemize(str)
    return str .. [[
\\stopitemize
]]
end

function rst_context.bullet_item (str)
    return [[

\\item ]] .. str .. [[

]]
end

--------------------------------------------------------------------------------
-- Definition lists 
--------------------------------------------------------------------------------
-- TODO define proper setups (probably bnf-like and some narrower for def-paragraphs)

function rst_context.deflist (str)
    return [[

\\startdefinitionlist
]] .. str .. [[

\\stopdefinitionlist
]]
end

function rst_context.deflist_item (str)
    return [[\\definitionitem{%]] .. str .. [[}% end definition item]]
end

function rst_context.deflist_classifier (str)
    return [[\\definitionclassifier{]] .. str .. [[}]]
end

function rst_context.deflist_term (str)
    return [[
    
  \\definitionterm{]] .. str .. [[}]]
end

function rst_context.deflist_def (str)
    return [[
    
  \\definitiondef{%]] .. str .. [[}]]
end

--------------------------------------------------------------------------------
-- Field lists
--------------------------------------------------------------------------------

-- TODO Do something useful with field lists. For now I'm not sure what as the
-- bibliography directives from the reST specification seem to make sense only
-- when using docinfo and, after all, we have .bib files that are portable.

function rst_context.field_list (str)
    return [[

\\startfieldlist]] .. str .. [[\\stopfieldlist
]]
end

function rst_context.field_name (str)
    return [[\\fieldname{]] .. str .. [[}]]
end

function rst_context.field_body (str)
    return [[\\fieldbody{]] .. str .. [[}]]
end

function rst_context.field (tab)
    local name, body = tab[1], tab[2]
    return string.format([[

  \\startfield
    \\fieldname{%s}
    \\fieldbody{%s}
  \\stopfield
]], name, inline_parser:match(body))
end

function rst_context.line_comment (str)
    return "% " .. str
end

function rst_context.block_comment (str)
    return string.format([[

\iffalse
%s
\fi
]], str)
end

function rst_context.option_list (str)
    return [[
\\setupTABLE[c][first] [background=color, backgroundcolor=grey, style=\tt]
\\setupTABLE[c][each]  [frame=off]
\\setupTABLE[r][each]  [frame=off]
\\bTABLE[split=repeat,option=stretch]
\\bTABLEhead
\\bTR
  \\bTH  Option \\eTH
  \\bTH  Description \\eTH
\\eTR
\\eTABLEhead
\\bTABLEbody
]] .. inline_parser:match(str) .. [[

\\eTABLEbody
\\eTABLE
]]
end

function rst_context.option_item (tab)
    return string.format([[\\bTR\\bTC %s \\eTC\\bTC %s \\eTC\\eTR
]], tab[1], tab[2])
end

function rst_context.literal_block (str)
    local indent = P" "^1
    local stripme = indent:match(str) or 0

    local strip = P{
        [1] = "strip",
        strip = Cs(V"line"^1),
        eol = P"\n",
        restofline = (1 - V"eol")^0,
        stop = Cs(V"eol" * P" "^0) * -P(1) / "", -- remove trailing blank lines
        line = Cs(V"restofline" * (V"stop" + V"eol")) / function (line)
            return #line > stripme and line:sub(stripme) or line
        end,
    }

    str = strip:match(str)
    --strip:print() -- grammar consists of 45 rules only; wheras a plain
                    -- pattern has 60+
    return [[

\\starttyping[lines=hyphenated]
]] .. str .. [[

\\stoptyping
]]
end

function rst_context.line_block (str)
    return [[

\\startlines
]] .. inline_parser:match(str) .. [[\\stoplines
]]
end

function rst_context.line_block_line(str)
    str = str:gsub("\n", " ")
    return str .. "\n"
end

function rst_context.line_block_empty()
    return "\n"
end

function rst_context.block_quote (tab)
    local str = [[
\\setupdelimitedtext  [blockquote][style={\\setupbodyfont[11pt]}] % awful placeholder
\\definedelimitedtext[attribution][blockquote]
\\setupdelimitedtext [attribution][style={\\setupbodyfont[11pt]\\it}]

\\startlinecorrection
\\startblockquote
]] .. inline_parser:match(tab[1]) .. [[

\\stopblockquote
]]

    return tab[2] and str .. [[
\\startattribution
]] .. tab[2] .. [[
\\stopattribution
\\stoplinecorrection
]]  or str .. [[
\\stoplinecorrection
]] 
end

function rst_context.table (str)
    return [[
\\startlinecorrection
]] .. str .. [[

\\stoplinecorrection
]]
end

function rst_context.grid_table (tab)
    local head
    if tab.has_head then
        head = [[
\\setupTABLE[c][first] [background=color, backgroundcolor=grey, style=\tt]
\\setupTABLE[c][each]  [frame=on]
\\setupTABLE[r][each]  [frame=on]
\\bTABLE[split=repeat,option=stretch]
\\bTABLEhead
\\bTR
  \\bTH  first \\eTH
  \\bTH  second \\eTH
  \\bTH  third \\eTH
  \\bTH  fourth \\eTH
\\eTR
\\eTABLEhead
\\bTABLEbody
]] 
    else
        head = [[
\\setupTABLE[c][each]  [frame=on]
\\setupTABLE[r][each]  [frame=on]
\\bTABLE[split=repeat,option=stretch]
\\bTABLEbody
]] 
    end
    local tail = [[

\\eTABLEbody
\\eTABLE
]]
    local body = ""
    for i,r in ipairs(tab.rows) do
        local isempty = true
        for n, cell in ipairs(r) do
            if cell.variant == "normal" then
                isempty = false
                break
            end
        end

        if not isempty then
            local row = [[\\bTR]]
            for n,c in ipairs(r) do
                if not (c.parent or
                        c.variant == "separator") then
                    local celltext = inline_parser:match(c.stripped)
                    if c.span.x or c.span.y then
                        local span_exp = "["
                        if c.span.x then
                            span_exp = span_exp .. "nc=" .. c.span.x .. ","
                        end
                        if c.span.y then
                            span_exp = span_exp .. "nr=" .. c.span.y
                        end
                        celltext  = span_exp .. "] " .. celltext

                    end

                    row = row .. "\n  " .. [[\\bTC ]] .. celltext .. [[\\eTC]]
                end
            end
            body = body .. row .. "\n" .. [[\\eTR]] .. "\n"
        end
    end
    return head .. body .. tail
end

function rst_context.table_row (tab)
    local tmp = [[\\bTR]]
    for n, cell in ipairs(tab) do
        tmp = tmp .. [[\\bTC]] .. cell .. [[\\eTC]] .. "\n"
        print (cell)
    end

    tmp = tmp .. [[\\eTR
]]
    return tmp
end

return rst_context