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"
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.paragraph (str)
    return "\n" .. [[\\startparagraph]] .. "\n" .. str .. "\n".. [[\\stopparagraph]] .. "\n"
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 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

return rst_context