Module:HelperMethods

Documentation for this module may be created at Module:HelperMethods/doc

------------------------------------------------------------------------------------
--                                 HelperMethods                                  --
--                                                                                --
-- This module includes a number of helper functions for dealing with lists,      --
-- e.g. in the person template. It is a meta-module, meant to be called from      --
-- other Lua modules, and should not be called directly from #invoke.             --
------------------------------------------------------------------------------------


local M = {}

-- Required modules for SPARQL queries and HTML table generation
local sparql = require('SPARQL')
local mwHtml = require('mw.html')

-- Return a string if an item's title is not available
local titleNotAvailablePageName = "TitleNotAvailable"
function M.titleNotAvailableStr(arg)
	local entity
	if type(arg) == "table" and arg.args then
		-- Called from template (e.g. Template:Publication)
		entity = arg.args[1]
	elseif type(arg) == "string" then
		-- Called from another module
		entity = arg
	else
		-- No parameter given
		entity = "Title"
	end

	return entity .. " not available ([[" .. titleNotAvailablePageName .. "|Why is that?]])"
end

-- Utility function to trim and lowercase a string
function M.trimAndLower(str)
	if str == nil then return nil end
	str = str:gsub("^%s*(.-)%s*$", "%1")
	return str:lower()
end

-- Utility function to count number of results in JSON answer of the SPQARL query
function M.countElementsInBindings(bindings)
	if not bindings then return 0 end
	local count = 0
	while bindings[count] do
	    count = count + 1
	end
	return count
end


-- This function will replace spaces with + and encode other non-alphanumeric 
-- characters into their percent-encoded representations, making the string 
-- safe to use in a URL.
function M.urlencode(str)
    if str then
        str = string.gsub(str, "\n", "\r\n")
        str = string.gsub(str, "([^%w %-%_%.%~])",
            function (c) return string.format("%%%02X", string.byte(c)) end)
        str = string.gsub(str, " ", "+")
    end
    return str
end


-- Function to convert JSON results into a Lua table
function M.convertJsonToTable(jsonResults)
    local resultsTable = {}
    if jsonResults and jsonResults.results and jsonResults.results.bindings then
        local bindings = jsonResults.results.bindings
        for j = 0, #bindings do
            local row = {}
            for key, value in pairs(bindings[j]) do
                table.insert(row, value.value)
            end
            table.insert(resultsTable, row)
        end
    end
    return resultsTable
end

-- Function to convert JSON results into a Lua table with specified field order
function M.convertJsonToTableOrdered(jsonResults, fieldOrder)
    local resultsTable = {}

    if jsonResults and jsonResults.results and jsonResults.results.bindings then
        local bindings = jsonResults.results.bindings
        -- Iterate through each result row in the bindings
        for j = 0, #bindings do
            local row = {}
            for _, fieldName in ipairs(fieldOrder) do
                -- Extract the value from the current binding
                local value = bindings[j][fieldName]
                if value and value.value then
                    table.insert(row, value.value)
                else
                    -- Use " " for missing or empty fields
                    table.insert(row, ' ')
                end
            end
            table.insert(resultsTable, row)
        end
    end

    return resultsTable
end

function M.makeWikiLinkOLD(link, name)
	if string.sub(link, 1, 34) == "https://portal.mardi4nfdi.de/wiki/" then
	   return "[[" .. string.sub(link, 35) .. "|" .. name .. "]]"
	end
	return "[" .. link .. " " .. name .. "]"
end

function M.makeWikiLink(link, name)
    local baseWiki = "https://portal.mardi4nfdi.de/wiki/"
    local baseEntity = "https://portal.mardi4nfdi.de/entity/"

	-- Case 1: Workaround for the "not available" info links -> return external link
    if string.find(name, titleNotAvailablePageName) then
	    return "[" .. link .. " " .. name .. "]"
    end    

    -- Case 2: full wiki URL -> strip prefix
    if string.sub(link, 1, #baseWiki) == baseWiki then
        local page = string.sub(link, #baseWiki + 1):match("^[^%?]+")
        return "[[" .. page .. "|" .. name .. "]]"
    end

    -- Case 3: entity URL -> convert to wiki/Item:Qxxx
    if string.sub(link, 1, #baseEntity) == baseEntity then
        local qid = string.sub(link, #baseEntity + 1)
        return "[[Item:" .. qid .. "|" .. name .. "]]"
    end
    
    -- fallback
    return "[" .. link .. " " .. name .. "]"
end
				

-- Function to convert JSON results into a comma-separated string
-- E.g. from a SPQRQL query: local jsonResults = sparql.runQuery(sparqlQuery)
-- Expected: "valueLabel" (name or description) and "value" (URL) 
function M.convertJsonToCommaSeparatedList(jsonResults)
	local resultsString = ""
	
	if jsonResults and jsonResults.results and jsonResults.results.bindings then
        local bindings = jsonResults.results.bindings
        for i = 0, #bindings do
            local binding = bindings[i]
            if binding.valueLabel and binding.valueLabel.value then
                if resultsString ~= "" then
                    resultsString = resultsString .. ", "
                end
                
                -- get value
                local value = binding.value.value

                -- get label
                local label = binding.valueLabel.value
--                if string.find(label, "https://") then
--                	label = "N/A"
--                end
                
                local nameAndLink = "[" .. value .. " " .. label .. "]"

                resultsString = resultsString .. M.makeWikiLink(value, label)
            end
        end
    end

    return resultsString
end


-- Additional function to generate the histogram data
-- colWithYear: Contains the index of the column that contains the year information. Expected format "2023-12-30"
function M.generateHistogramChartFromTable(dataTable, colWithYear)
    local yearCounts = {}

    for _, row in ipairs(dataTable) do
        -- Extract the year from the fourth column (publication date)
        local date = row[colWithYear]
        if date then  -- Check if the date exists
            local year = date:sub(1, 4)  -- Extract the year

            if year ~= "" then
                yearCounts[year] = (yearCounts[year] or 0) + 1
            end
        end
    end

    local years = {}
    local counts = {}
    for year, count in pairs(yearCounts) do
        table.insert(years, year)
        table.insert(counts, count)
    end

    -- Sort the years to maintain chronological order
    table.sort(years)
    local sortedCounts = {}
    for _, year in ipairs(years) do
        table.insert(sortedCounts, yearCounts[year])
    end

    local chartData = {
        type = 'bar',
        data = {
            labels = years,  -- x-axis labels (years)
            datasets = {{
                label = 'Number of Publications',
                data = sortedCounts,  -- y-axis data (counts)
                backgroundColor = 'rgba(54, 162, 235, 0.2)',
                borderColor = 'rgba(54, 162, 235, 1)',
                hoverBackgroundColor = 'red',
                borderWidth = 1
            }}
        },
        options = {
            scales = {
                y = {
                    beginAtZero = true
                }
            }
        }
    }

    return chartData
end


-- Function to create a HTML table from a Lua table where columns are merged
-- mergeColumns: contains a list of which columns should be merged, e.g. {{1, 2}, {4}} or {{2, 4}, {1, 3}}
function M.createHtmlTableWithMergedCols(dataTable, headers, mergeColumns, itemprop)
    local htmlTable = mwHtml.create('table')
    htmlTable:addClass('wikitable'):attr('border', '1')
    
    htmlTable:addClass('sortable') -- This line ensures your table has the 'sortable' class

    local headerRow = htmlTable:tag('tr')
    
    -- Use the provided headers
    for _, header in ipairs(headers) do
        headerRow:tag('th'):wikitext(header)
    end

    for _, row in ipairs(dataTable) do
	    if not string.find(row[1], "/entity/statement/") then
	        local dataRow = htmlTable:tag('tr')
	        
	        for _, cols in ipairs(mergeColumns) do
	            local combinedData
	            if #cols == 1 then
	                -- If only one column index is provided, use it as is, default to "N/A" string if nil
	                combinedData = row[cols[1]] or "N/A"
	            else
	                -- If two column indices are provided, merge them, defaulting to "N/A" string if nil
	                local col1Data = row[cols[1]] or "N/A"
	                local col2Data = row[cols[2]] or "N/A"
	                
					-- Check if col1Data is " "
					if col1Data == " " then
					    combinedData = col2Data
					-- Check if both col1Data and col2Data are " "
					elseif col1Data == " " and col2Data == " " then
					    combinedData = " "
					else
					--    combinedData = '[' .. col1Data .. ' ' .. col2Data .. ']'
						combinedData = M.makeWikiLink(col1Data, col2Data)
					end
					if itemprop then
						combinedData = combinedData .. ' <link itemprop="' .. itemprop .. '" href="' .. col1Data .. '"/> '
					end
	            end
	            dataRow:tag('td'):wikitext(combinedData)
	        end
	    end
    end
    
    -- mw.log(htmlTable)
    
	return tostring(htmlTable)
end

-- Helper function to convert table contents to a single string
function tableToString(t)
    local stringParts = {}
    for key, value in pairs(t) do
        if type(value) == 'string' then
            table.insert(stringParts, value)
        elseif type(value) == 'table' then
            -- Recursively convert nested tables to string
            table.insert(stringParts, tableToString(value))
        elseif type(value) == 'number' or type(value) == 'boolean' then
            -- Convert numbers and booleans to string directly
            table.insert(stringParts, tostring(value))
        end
    end
    return table.concat(stringParts, " ") -- Combine with a space as separator
end

-- Function to convert text to JSON valid
function M.validJSON(text)
	    -- Ensure text is always a string (handles nil cases)
    text = tostring(text or "")

    -- Use mw.text.jsonEncode to escape special characters
    return mw.text.jsonEncode(text)
end


-- Function to convert Markdown text to Mediawiki format
function M.markdownToMediawiki(mdText)

    -- Check if the input is a table and convert it
    if type(mdText) == 'table' then
        mdText = tableToString(mdText)
    elseif type(mdText) ~= 'string' then
        error("Expected a string or a table, got " .. type(mdText))
    end

    -- Process the string to convert Markdown to MediaWiki formatting

    -- Replace newline and tab placeholders
    mdText = string.gsub(mdText, "\\N", "\n")
    mdText = string.gsub(mdText, "\\T", "\t")

    -- Convert bold markdown (**text**) to MediaWiki bold ('''text''')
    mdText = string.gsub(mdText, "%*%*(.-)%*%*", "'''%1'''")

    -- Convert markdown links [text](url) to MediaWiki [url text]
    mdText = string.gsub(mdText, "%[(.-)%]%((.-)%)", "[%2 %1]")

    -- Convert headings: match at start of line and beginning of string
    mdText = string.gsub(mdText, "\n### (.-)\n", "\n=== %1 ===\n")
    mdText = string.gsub(mdText, "\n## (.-)\n", "\n== %1 ==\n")
    mdText = string.gsub(mdText, "\n# (.-)\n", "\n= %1 =\n")

    mdText = string.gsub(mdText, "^### (.-)\n", "=== %1 ===\n")
    mdText = string.gsub(mdText, "^## (.-)\n", "== %1 ==\n")
    mdText = string.gsub(mdText, "^# (.-)\n", "= %1 =\n")

    return mdText
end

-- Created a string containing all values for a given property of a given item
function M.getValuesForProperty(frame)
	-- Ensure text is always a string (handles nil cases)
	local target1 = tostring(frame.args[1] or "")
	local PID = tostring(frame.args[2] or "")
	
-- Constructing the SPARQL query with dynamic entity target1
    local sparqlQuery = [[
PREFIX target1: <https://portal.mardi4nfdi.de/entity/]] .. target1 .. [[>
PREFIX wdt: <https://portal.mardi4nfdi.de/prop/direct/>
PREFIX wd: <https://portal.mardi4nfdi.de/entity/>

SELECT ?property ?propertyLabel ?value ?valueLabel
WHERE {
    target1: wdt:]] .. PID .. [[ ?value .
    OPTIONAL {
        ?value rdfs:label ?valueLabel .
        FILTER(LANG(?valueLabel) = "en")
    }
    BIND(wd:]] .. PID .. [[ AS ?property)
    SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
    ]]
    
    -- mw.log( sparqlQuery )
    
	-- Executing the SPARQL query and retrieving results in JSON format
	local jsonResults = sparql.runQuery(sparqlQuery)
	
	-- mw.logObject(jsonResults)

	-- Handle error in SPARQL query execution
	if jsonResults and jsonResults.error then
    	mw.log("Error in SPARQL query: " .. tostring(jsonResults.error))
    	return nil
	end
	
	return jsonResults
end

-- Returns a list of the values for this property.
-- To test:
--    local HM = require("Module:HelperMethods")
--    local fakeFrame = {args = {[1] = "Q6767920", [2] = "P56"}}
--    result = HM.getValuesForPropertyList(fakeFrame)
--    mw.log(result)
function M.getValuesForPropertyList(frame)
	
	local target1 = frame.args[1]
	local PID = frame.args[2]
	
	target1 = tostring(target1 or "")
    if not target1 or target1 == '' then
        return "No records found"
    end    
    
    local fakeFrame = { args = {[1] = target1, [2] = PID } }
	jsonResults = M.getValuesForProperty( fakeFrame )

	if not jsonResults then
        return "Could not fetch data."
	end
	
	if not jsonResults.results then
        return "Could not fetch any data."
	end
	
	if M.countElementsInBindings(jsonResults.results.bindings) == 0 then
        return "No records found."
	end

	local list = M.convertJsonToCommaSeparatedList(jsonResults)
	
	-- mw.log(licenseList) 
	
    return list

end

local supers = {
	["0"]="⁰",["1"]="¹",["2"]="²",["3"]="³",["4"]="⁴",
	["5"]="⁵",["6"]="⁶",["7"]="⁷",["8"]="⁸",["9"]="⁹",
	["n"]="ⁿ",["k"]="ᵏ",["i"]="ⁱ",["j"]="ʲ",["m"]="ᵐ"
}

local subs = {
	["0"]="₀",["1"]="₁",["2"]="₂",["3"]="₃",["4"]="₄",
	["5"]="₅",["6"]="₆",["7"]="₇",["8"]="₈",["9"]="₉"
}

local function to_super(s)
	return (s:gsub(".", function(c) return supers[c] or c end))
end

local function strip_latex(s)
	s = s:gsub("\\%a+%b{}", function(cmd)
		return cmd:match("{(.*)}")
	end)

	s = s:gsub("\\%a+", "")

	s = s:gsub("[{}]", "")

	s = s:gsub("(%a)_(%d)", function(base, sub)
		return base .. (subs[sub] or sub)
	end)

	s = s:gsub("%s+", " ")

	return s
end

local function convert_exponents(txt)
	txt = txt:gsub("(%d)%^{(%w+)}", function(base, exp)
		return base .. to_super(exp)
	end)

	txt = txt:gsub("(%d)%^(%w+)", function(base, exp)
		return base .. to_super(exp)
	end)

	return txt
end

function M.cleanTitle(frame)
	local text = frame.args[1] or ""
	if type(text) ~= "string" then text = tostring(text) end

	text = text:gsub("%$(.-)%$", function(tex)
		local base, exp = tex:match("^(%a)%^(%d)$")
		if base and exp and supers[exp] then
			return base .. supers[exp]
		end
		return tex
	end)

	text = text:gsub("\\%((.-)\\%)", function(inner)
		return strip_latex(inner)
	end)

	text = text:gsub("\\%[(.-)\\%]", function(inner)
		return strip_latex(inner)
	end)

	text = convert_exponents(text)

	text = text:gsub("%s+", " "):gsub("^%s*(.-)%s*$", "%1")
	return text
end



return M