ICANNWiki

Module:CountryData: Difference between revisions

← Older edit
MarkWD (talk | contribs)
Bureaucrats, Check users, Interface administrators, Administrators (Semantic MediaWiki), Curators (Semantic MediaWiki), Editors (Semantic MediaWiki), steward, Suppressors, Administrators, Temporary account IP viewers
25,069 edits
// via Wikitext Extension for VSCode
// via Wikitext Extension for VSCode
 
(11 intermediate revisions by the same user not shown)
Line 1: Line 1:
-- Module:CountryData
--[[
-- Unified module for country data management.
* Name: CountryData
--
* Author: Mark W. Datysgeld
-- Features:
* Description: Unified module for country data management with JSON loading, normalization, region mapping, and Semantic MediaWiki integration
--  * Loads country data from JSON stored in Data:CountryDataset.json
* Notes: Loads from Data:CountryDataset.json; normalizes country names to canonical forms; maps countries to ICANN regions; provides extensible property access; formats country lists with region-specific emoji styling; processes countries for category assignment
--  * Normalizes country names to canonical forms
]]
--  * Maps countries to ICANN regions
--  * Provides extensible property access
--  * Integrates with Semantic MediaWiki
--  * Formats country lists with region-specific emoji styling
--  * Processes countries for category assignment


-- Dependencies
-- Dependencies
Line 208: Line 203:
     if not code then
     if not code then
         local stripped = DiacriticNormalization.removeDiacritics(name)
         local stripped = DiacriticNormalization.removeDiacritics(name)
        -- Check if stripping made a difference to avoid redundant normalization
         if stripped ~= name then
         if stripped ~= name then
             local normalizedStripped = NormalizationText.normalizeText(stripped)
             code = nameLookup[NormalizationText.normalizeText(stripped)]
            code = nameLookup[normalizedStripped]
         end
         end
     end
     end
Line 324: Line 317:
                 table.insert(result, {
                 table.insert(result, {
                     code = code,
                     code = code,
                     name = country.name or country.canonical_name
                     name = country.name
                 })
                 })
             end
             end
Line 362: Line 355:
     if data and data.countries then
     if data and data.countries then
         for _, country in pairs(data.countries) do
         for _, country in pairs(data.countries) do
             table.insert(result, country.name or country.canonical_name)
             table.insert(result, country.name)
         end
         end
     end
     end
Line 451: Line 444:
     functionCache[cacheKey] = properties
     functionCache[cacheKey] = properties
     return properties
     return properties
end
function CountryData.getSemanticPropertyName(propertyKey)
    local ConfigRepository = require('Module:ConfigRepository')
    for templateName, templateConfig in pairs(ConfigRepository.templates) do
        if templateConfig.semantics and templateConfig.semantics.additionalProperties then
            if templateConfig.semantics.additionalProperties[propertyKey] then
                return propertyKey
            end
        end
    end
    return nil
end
end


Line 471: Line 452:
     end
     end
      
      
     local countryPropertyName = CountryData.getSemanticPropertyName("Has country")
    local ConfigRepository = require('Module:ConfigRepository')
     local regionPropertyName = CountryData.getSemanticPropertyName("Has ICANN region")
     local countryPropertyName = ConfigRepository.getSemanticPropertyName("Has country")
     local regionPropertyName = ConfigRepository.getSemanticPropertyName("Has ICANN region")
      
      
     if not countryPropertyName or not regionPropertyName then
     if not countryPropertyName or not regionPropertyName then
Line 486: Line 468:
     end
     end
      
      
     for _, country in ipairs(countries) do
     for _, countryName in ipairs(countries) do
         local normalizedCountry = CountryData.normalizeCountryName(country)
         local normalizedCountry = CountryData.normalizeCountryName(countryName)
         if normalizedCountry ~= "(Unrecognized)" then
         if normalizedCountry ~= "(Unrecognized)" then
            -- Initialize property tables if they don't exist
             properties[countryPropertyName] = properties[countryPropertyName] or {}
             properties[countryPropertyName] = properties[countryPropertyName] or {}
             table.insert(properties[countryPropertyName], normalizedCountry)
             table.insert(properties[countryPropertyName], normalizedCountry)
              
              
             local region = CountryData.getRegionByCountry(country)
             local region = CountryData.getRegionByCountry(normalizedCountry)
             if region and region ~= "(Unrecognized)" then
             if region and region ~= "(Unrecognized)" then
                 properties[regionPropertyName] = properties[regionPropertyName] or {}
                 properties[regionPropertyName] = properties[regionPropertyName] or {}
Line 535: Line 518:
function CountryData.formatCountryList(value)
function CountryData.formatCountryList(value)
     if not value or value == "" then return "" end
     if not value or value == "" then return "" end
      
 
     local countries = {}
     local ListGeneration = require('Module:ListGeneration')
     for country in string.gmatch(value, "[^;]+") do
     local itemsToProcess = {}
        local trimmed = country:match("^%s*(.-)%s*$")
 
        if trimmed and trimmed ~= "" then
    -- First, check if the entire string is a single, valid country.
            table.insert(countries, trimmed)
     -- This correctly handles names like "Trinidad and Tobago".
    local singleCountryName = CountryData.normalizeCountryName(value)
    if singleCountryName ~= "(Unrecognized)" then
        -- If it's a valid country, treat it as a single item.
        table.insert(itemsToProcess, value)
    else
        -- If not a single country, assume it's a list and split ONLY by semicolon.
        -- This is safer than letting ListGeneration guess the delimiter.
        for item in string.gmatch(value, "[^;]+") do
            local trimmed = item:match("^%s*(.-)%s*$")
            if trimmed and trimmed ~= "" then
                table.insert(itemsToProcess, trimmed)
            end
         end
         end
     end
     end
      
 
     local normalizedCountries = {}
     -- Define the item hook for country-specific formatting
    for _, country in ipairs(countries) do
     local function countryItemHook(countryName)
         local normalized = CountryData.normalizeCountryName(country)
         local normalized = CountryData.normalizeCountryName(countryName)
         if normalized ~= "(Unrecognized)" then
         if normalized ~= "(Unrecognized)" then
             table.insert(normalizedCountries, normalized)
             local countryRegion = CountryData.getRegionByCountry(normalized)
            -- Return a table with content and class for the li element
            return {
                content = normalized,
                class = getRegionClass(countryRegion)
            }
         end
         end
        return nil -- Exclude unrecognized countries from the list
     end
     end
      
 
     if #normalizedCountries > 0 then
     -- Set the options for the list generation
        local listItems = {}
     local options = {
         for _, country in ipairs(normalizedCountries) do
         mode = 'bullet',
            local countryRegion = CountryData.getRegionByCountry(country)
         listClass = 'template-list-country',
            local regionClass = getRegionClass(countryRegion)
        itemHook = countryItemHook
            table.insert(listItems, string.format("<li class=\"%s\">%s</li>", regionClass, country))
     }
        end
 
         return string.format("<ul class=\"template-list template-list-country\">%s</ul>", table.concat(listItems, ""))
     -- Pass the pre-processed table of items to the list generator.
     end
     return ListGeneration.createList(itemsToProcess, options)
      
     return ""
end
end


Line 573: Line 572:
      
      
     local countries = {}
     local countries = {}
     for country in string.gmatch(value, "[^;]+") do
     for countryName in string.gmatch(value, "[^;]+") do
         local trimmed = country:match("^%s*(.-)%s*$")
         local trimmed = countryName:match("^%s*(.-)%s*$")
         if trimmed and trimmed ~= "" then
         if trimmed and trimmed ~= "" then
             table.insert(countries, trimmed)
             local normalized = CountryData.normalizeCountryName(trimmed)
            if normalized ~= "(Unrecognized)" then
                table.insert(countries, normalized)
            end
         end
         end
     end
     end
      
      
     local normalizedCountries = {}
     return countries
    for _, country in ipairs(countries) do
        local normalized = CountryData.normalizeCountryName(country)
        if normalized ~= "(Unrecognized)" then
            table.insert(normalizedCountries, normalized)
        end
    end
   
    return normalizedCountries
end
end


Line 595: Line 589:
      
      
     local inputName = countryNameOrCode:gsub('_', ' ')
     local inputName = countryNameOrCode:gsub('_', ' ')
     local isoCode = CountryData.getCountryCodeByName(inputName)  
     local isoCode
   
    -- First, try to get the ISO code by treating inputName as a country name.
    isoCode = CountryData.getCountryCodeByName(inputName)  
      
      
    -- If no code was found by name, and the inputName itself is 2 characters long,
    -- it might be an ISO code already. Let's validate it.
     if not isoCode and #inputName == 2 then
     if not isoCode and #inputName == 2 then
         if CountryData.getCountryByCode(inputName) then  
         if CountryData.getCountryByCode(inputName) then  
Retrieved from "https://staging.icannwiki.org/Module:CountryData"