Module:Import table: Difference between revisions

Module documentation[view] [edit] [history] [purge]

Usage

{{#invoke:Import table|import
|page=
|config=
|common=
}}

This module is designed to import data from tables in Wikipedia articles into Wikidata. The first column of the table must be the name of the item.

• If this is a link (or a redirect) to an existing article which has a corresponding Wikidata item, then the data will be imported into this item.
• If the first column contains a valid identifier for a Wikidata item (e.g. Q123456) then the data will be imported into this item.
• If this is plain text or a redlink, then the data will be imported into a new Wikidata item.

Parameters

• page - the name of the article/page to parse, e.g. |page=List of dams in South Africa
• config - details about what type of data is held in each column of the table formatted as type-property-option. Currently recognised types are:
  • label - the name of the article
  • wikilink - a link to an article which should be the value of the property, e.g. wikilink-P131
  • quantity - specify the unit in the third parameter, e.g. quantity-P2048-Q11573
  • year - a 4-digit year, e.g. year-P571
  • text - any text to import into a string datatype, e.g. text-P3562
  • coord - coordinate position, inside the {{coord}} template, e.g. coord-P625
  • donotuse (or any other unrecognised type) - indicates a column that will not be imported
• common - a set of values that every item in the list should have, formatted in pairs as property-value, e.g. |common=P31-Q12323,P17-Q258

Notes

• The module will not import any claim if there is already a statement for that property in Wikidata, no matter if the value is the same or different to that being imported, and even if the value is marked as deprecated.

The above documentation is transcluded from Module:Import table/doc. (edit | history)
Editors can experiment in this module's sandbox (edit | diff) and testcases (create) pages.
Subpages of this module.

require("strict")
local p = {}

local function resolveqid(title)
	local resolveEntity = require("Module:ResolveEntityId")._id
	local qid
	if title then
		local rawarticle = string.match(title,'%[%[(.+)%|') or string.match(title,'%[%[(.+)%]%]')
		if rawarticle and rawarticle ~= "" then
			qid = resolveEntity(rawarticle)
		end
	end
	return qid
end

function p.import(frame)
	local tab = "|"
	local wikiqid = "Q328" -- QID for English Wikipedia
	local eol = "<br>" -- end of line string
	local coord2text = require("Module:Coordinates/sandbox")._coord2text
	local args = frame.args or frame:getParent().args
	if not args.page then
		return "No page specified."
	end
	if not args.config then
		return "No configuation."
	end
	local config = {}
	for c1,col in ipairs(mw.text.split(args.config,",")) do
		config[c1] = mw.text.split(col,"-")
	end
	local common = {}
	if args.common then
		for c1,prop in ipairs(mw.text.split(args.common,",")) do
			common[c1] = mw.text.split(prop,"-")
			common[common[c1][1]] = common[c1][2] -- create index, e.g. common["P31"]=Q39715
		end
	end
	local content = mw.title.new(args.page):getContent() -- read page specified
	content = string.match(content,"%{%|(.+)%|%}") -- keep table only
	content = string.gsub(content,"||","\n|") -- use \n| for column breaks
	local rows = mw.text.split(content,"|-",true) -- split table into rows
	table.remove(rows,1) -- remove table definition
	table.remove(rows,1) -- remove heading row
	local output = ""
	local v2 = ""
	for rn,row in ipairs(rows) do
		local columns = mw.text.split(row,"\n|") -- split table rom into columns
		table.remove(columns,1) -- remove content before the first \n| character
		local label = columns[1]
		label = mw.text.trim(label)
		local qid = resolveqid(label) -- resolve qid if first column is link
		if not qid then
			v2 = v2 .. "CREATE" .. eol
			qid = "LAST" -- creating new item, so use LAST
			v2 = v2 .. qid .. tab .. 'Len' .. tab .. '"' .. label .. '"' .. eol
			if common["P31"] then -- create auto-description
				v2 = v2 .. qid .. tab .. 'Den' .. tab .. '"' .. mw.wikibase.getLabel(common["P31"])
				if common["P17"] then -- add country
					v2 = v2 .. ' in ' .. mw.wikibase.getLabel(common["P17"])
				end
				v2 = v2  .. '"' .. eol
			end
			for c = 1,#common do
				v2 = v2 .. qid .. tab .. common[c][1] .. tab .. common[c][2] .. eol
			end
		end
		local function addtov2(prop,val)
			v2 = v2 .. qid .. tab .. prop .. tab .. val .. tab .. "S143" .. tab .. wikiqid .. eol
			return v2
		end
		for cn,col in ipairs(columns) do
			if config[cn] then
				for ref in mw.ustring.gmatch(col,"%<ref.+%<%/ref%>") do
					col = mw.ustring.gsub(col,ref,"")
				end
				col = mw.text.trim(col)
				if  config[cn][1] == "wikilink" then
					local val = resolveqid(col)
					if val then
						v2 = addtov2(config[cn][2],val)
					end
				elseif config[cn][1] == "year" then
					local val = string.match(col,"%d%d%d%d")
					if val then
						v2 = addtov2(config[cn][2],"+" ..val .. "-00-00T00:00:00Z/9")
					end
				elseif config[cn][1] == "quantity" then
					local val = string.gsub(col,",","") -- remove any commas
					val = string.match(val,"%d+%.?%d*") -- extract value, possibly from inside convert template
					if val then
						if string.upper(string.sub(config[cn][3],1,1)) == "Q" then
							config[cn][3] = string.sub(config[cn][3],2)
						end
						v2 = addtov2(config[cn][2],val .. "U" .. config[cn][3])
					end
				elseif config[cn][1] == "text" then
					if col ~= "" then
						v2 = addtov2(config[cn][2],'"' .. col .. '"')
					end
				--elseif config[cn][1] == "coord" then
					--import[#import+1] = {}
					--import[#import].property = config[cn][2]
					--import[#import].value = coord2text(col,"lat") .. "/" .. coord2text(col,"long")
				end
			end
		end
	end
	return v2
end

return p