local p = {}
-- Helper to mimic PHP's explode(delimiter, string, 2)
local function split_once(str, delimiter)
local pos = mw.ustring.find(str, delimiter, 1, true)
if pos then
return mw.ustring.sub(str, 1, pos - 1), mw.ustring.sub(str, pos + mw.ustring.len(delimiter))
end
return str, nil
end
function p.parseTablesToData(text)
local lines = mw.text.split(text, "\n")
local parsed_data = {}
local current_table = nil
local current_row = nil
for _, outLine in ipairs(lines) do
local line = mw.text.trim(outLine)
-- Equivalent to skipping via 'continue'
if line ~= '' then
local first_char = mw.ustring.sub(line, 1, 1)
local first_two = mw.ustring.sub(line, 1, 2)
local indent_match, attributes_match = mw.ustring.match(line, "^(:*)%s*%{%|(.*)$")
if indent_match then
-- Start a new table
current_table = {
attributes = mw.text.unstrip(attributes_match),
rows = {}
}
table.insert(parsed_data, current_table)
current_row = nil
elseif not current_table then
-- Not inside a table, do nothing (skip)
elseif first_two == '|}' then
-- End the current table
current_table = nil
current_row = nil
elseif first_two == '|-' then
-- Start a new row
line = mw.ustring.gsub(line, "^|%-+", "")
current_row = {
attributes = mw.text.unstrip(line),
cells = {}
}
table.insert(current_table.rows, current_row)
elseif first_char == '|' or first_char == '!' or first_two == '|+' then
-- Process cells, headers, or captions
-- Handle implicit rows (if cell data appears before a |- tag)
if not current_row and first_two ~= '|+' then
current_row = { attributes = "", cells = {} }
table.insert(current_table.rows, current_row)
end
if first_two == '|+' then
first_char = '+'
line = mw.ustring.sub(line, 3)
current_table.caption = mw.text.trim(mw.text.unstrip(line))
else
line = mw.ustring.sub(line, 2)
if first_char == '!' then
line = mw.ustring.gsub(line, "!!", "||")
end
local cells = mw.text.split(line, "||", true)
for _, cell_str in ipairs(cells) do
local cell_type = (first_char == '!') and "th" or "td"
local cell_attr_str, cell_content_str = split_once(cell_str, "|")
local cell_obj = {
type = cell_type,
attributes = "",
content = ""
}
-- Replicate PHP's logic to check for [[ links or -{ language converter tags
-- to ensure a pipe inside a link isn't mistaken for an attribute delimiter.
if not cell_content_str or mw.ustring.match(cell_attr_str, "%[%[") or mw.ustring.match(cell_attr_str, "%-%{") then
cell_obj.content = mw.text.trim(cell_str)
else
cell_obj.attributes = mw.text.unstrip(cell_attr_str)
cell_obj.content = mw.text.trim(cell_content_str)
end
table.insert(current_row.cells, cell_obj)
end
end
end
end
end
return parsed_data
end
function p.testLoadPage(frame)
-- 1. Define the page title to load.
-- Note: Testing this on the full US Reps page immediately will yield a massive dump string.
-- I recommend creating a subpage with a small test table first.
local page_title = "List of current United States representatives#List of representatives"
local titleObj = mw.title.new(page_title)
if not titleObj then return "Error: Invalid title." end
-- 2. Fetch the raw wikitext
local content = titleObj:getContent()
if not content then return "Error: Page not found or empty." end
-- 3. Run the parser
local parsed_data = p.parseTablesToData(content)[6]
-- 4. Dump the resulting Lua table as a preformatted string so you can read the structure
return "<pre>\n" .. mw.dumpObject(parsed_data) .. "\n</pre>"
end
function p.tallyParties(frame)
-- 1. Load the page and parse the text
local page_title = "List of current United States representatives#List of representatives"
local titleObj = mw.title.new(page_title)
if not titleObj then return "Error: Invalid title." end
local content = titleObj:getContent()
if not content then return "Error: Page not found or empty." end
local parsed_data = p.parseTablesToData(content)[6]
if not parsed_data then return "Error: Table not found." end
-- 2. Build a dynamic column map from the header row
local column_map = {}
local current_col = 1
local header_row = parsed_data.rows[1]
for _, cell in ipairs(header_row.cells) do
-- Clean the header text (removes <ref> tags and hidden spans)
local header_text = mw.ustring.gsub(cell.content, "<[^>]+>", "")
header_text = mw.text.trim(header_text)
-- Check if this header spans multiple columns (e.g., colspan=2)
local colspan = 1
local colspan_match = mw.ustring.match(cell.attributes, "colspan%s*=%s*[\"']?(%d+)[\"']?")
if colspan_match then
colspan = tonumber(colspan_match)
end
column_map[header_text] = {
start_index = current_col,
span = colspan,
-- For Party, the text name is in the right-most cell of the span
end_index = current_col + colspan - 1
}
current_col = current_col + colspan
end
-- Abort if we can't find the Party column
if not column_map["Party"] then return "Error: Could not find 'Party' column." end
-- Extract the exact index we need to check in the data rows
local party_col_index = column_map["Party"].end_index
-- 3. Initialize the tally counters
local tallies = {
Democratic = 0,
Republican = 0,
Independent = 0,
Vacant = 0
}
-- 4. Iterate through the data rows
for i = 2, #parsed_data.rows do
local row = parsed_data.rows[i]
local cells = row.cells
if #cells > 0 then
-- Check for vacant seats (vacant rows use a massive colspan, yielding fewer cells)
if #cells < party_col_index then
assert(#cells == 2, "Row had " .. #cells .. " cells. Expected 2.")
assert(cells[2].content == "''Vacant''", "Expected ''Vacant'', got: " .. tostring(cells[2].content))
tallies.Vacant = tallies.Vacant + 1
else
-- 5. Extract the party using the dynamic index
local party = cells[party_col_index].content
party = mw.ustring.gsub(party, "<[^>]+>", "")
party = mw.text.trim(party)
if tallies[party] ~= nil then
tallies[party] = tallies[party] + 1
else
tallies[party] = 1
end
end
end
end
-- 6. Format the output
local output = "'''Current House Tallies:'''\n"
output = output .. "* Democratic: " .. tallies.Democratic .. "\n"
output = output .. "* Republican: " .. tallies.Republican .. "\n"
output = output .. "* Independent: " .. tallies.Independent .. "\n"
output = output .. "* Vacant: " .. tallies.Vacant .. "\n"
return output
end
return p
