Wikibooks zhwikibooks https://zh.wikibooks.org/wiki/Wikibooks:%E9%A6%96%E9%A1%B5 MediaWiki 1.47.0-wmf.3 first-letter Media Special Talk User User talk Wikibooks Wikibooks talk File File talk MediaWiki MediaWiki talk Template Template talk Help Help talk Category Category talk Transwiki Transwiki talk Wikijunior Wikijunior talk Subject Subject talk TimedText TimedText talk Module Module talk Event Event talk 0 22141 184517 123999 2026-05-21T03:31:37Z TunnelESON 317 修正重定向 184517 wikitext text/x-wiki #重定向 [[中國歷史/附錄三 中國歷代君主全表]] 1rb483x6t4dn8jsbxgilbk5oj1y6r4w 0 34064 184526 182867 2026-05-21T06:28:21Z P1ayer 1197 /* 派 */ 184526 wikitext text/x-wiki {{W|武林}} == 組織 == === 派 === <div style="text-align: left; column-count: 3; display: block;"> *{{W|少林派}} *{{W|武當派}} *{{W|點蒼派}} *{{W|華山派}} *{{W|峨嵋派}} *{{w|崑崙派}} *{{w|崆峒派}} *{{w|青城派}} *{{w|茅山宗|茅山派}} </div> === 幫 === *{{w|丐幫}} === 門 === *下五門 === 世家 === *諸葛 *司馬 *南宮 *四川{{w|唐門}} === 宗教 === *{{w|明教}} *{{w|全真教}} *天魔神教 === 其他 === *{{w|鏢局}} == 地點 == *{{w|中原}} 14wvs2ijzyaq4uayuxne81dtmdlh1zj 184527 184526 2026-05-21T06:51:47Z P1ayer 1197 /* 門 */ 184527 wikitext text/x-wiki {{W|武林}} == 組織 == === 派 === <div style="text-align: left; column-count: 3; display: block;"> *{{W|少林派}} *{{W|武當派}} *{{W|點蒼派}} *{{W|華山派}} *{{W|峨嵋派}} *{{w|崑崙派}} *{{w|崆峒派}} *{{w|青城派}} *{{w|茅山宗|茅山派}} </div> === 幫 === *{{w|丐幫}} === 門 === *{{w|洪門}} *下五門 *:指傳統社會三教九流中的五個下層社會職業(車、船、店、腳、牙)為掩護的情報蒐集組織。 === 世家 === *諸葛 *司馬 *南宮 *四川{{w|唐門}} === 宗教 === *{{w|明教}} *{{w|全真教}} *天魔神教 === 其他 === *{{w|鏢局}} == 地點 == *{{w|中原}} n09rve76lh9w34p4dpgy11hpl9ukmuz 10 34330 184518 2026-05-21T05:47:19Z P1ayer 1197 创建页面，内容为“<onlyinclude>{{safesubst:<noinclude/>#invoke:columns|display|sort=1|collapse=1|columns=3}}</onlyinclude><!-- -->{{documentation}}” 184518 wikitext text/x-wiki <onlyinclude>{{safesubst:<noinclude/>#invoke:columns|display|sort=1|collapse=1|columns=3}}</onlyinclude><!-- -->{{documentation}} 89fsybsde8t9sdhj0q4itbm3fgjicl2 828 34331 184519 2026-05-21T05:55:57Z P1ayer 1197 创建页面，内容为“local export = {} local collation_module = "Module:collation" local debug_track_module = "Module:debug/track" local headword_data_module = "Module:headword/data" local JSON_module = "Module:JSON" local languages_module = "Module:languages" local links_module = "Module:links" local pages_module = "Module:pages" local parameter_utilities_module = "Module:parameter utilities" local parameters_module = "Module:parameters" local parse_utilities_module = "Module:par…” 184519 Scribunto text/plain local export = {} local collation_module = "Module:collation" local debug_track_module = "Module:debug/track" local headword_data_module = "Module:headword/data" local JSON_module = "Module:JSON" local languages_module = "Module:languages" local links_module = "Module:links" local pages_module = "Module:pages" local parameter_utilities_module = "Module:parameter utilities" local parameters_module = "Module:parameters" local parse_utilities_module = "Module:parse utilities" local pron_qualifier_module = "Module:pron qualifier" local qualifier_module = "Module:qualifier" local string_utilities_module = "Module:string utilities" local table_module = "Module:table" local utilities_module = "Module:utilities" local yesno_module = "Module:yesno" local m_str_utils = require(string_utilities_module) local concat = table.concat local html = mw.html.create local is_substing = mw.isSubsting local insert = table.insert local rmatch = m_str_utils.match local remove = table.remove local sub = string.sub local trim = m_str_utils.trim local u = m_str_utils.char local dump = mw.dumpObject local function track(page) require(debug_track_module)("columns/" .. page) return true end local function deepEquals(...) deepEquals = require(table_module).deepEquals return deepEquals(...) end local function term_already_linked(term) return term == "?" or -- signals an unknown term -- optimization to avoid unnecessarily loading [[Module:parse utilities]] (term:find("[<{]") and require(parse_utilities_module).term_already_linked(term)) end local function convert_delimiter_to_separator(item, itemind, args) if itemind == 1 then item.separator = nil elseif item.delimiter == " " then item.separator = args.space_delim elseif item.delimiter == "~" then item.separator = args.tilde_delim else item.separator = args.comma_delim end end local function get_horizontal_separator(args_horiz, embedded_comma) return args_horiz == "bullet" and " · " or embedded_comma and "; " or ", " end -- Suppress false positives in categories like [[Category:English links with redundant wikilinks]] so people won't -- be tempted to "correct" them; terms like embedded ~ like [[Micros~1]] or embedded comma not followed by a space -- such as [[1,6-Cleves acid]] need to have a link around them to avoid the tilde or comma being interpreted as a -- delimiter. local function suppress_redundant_wikilink_cat(term, alt) return term:find("~") or term:find(",%S") end local function full_link_and_track_self_links(item, face, nochecktr) if item.term then local pagename = mw.loadData(headword_data_module).pagename local term_is_pagename = item.term == pagename local term_contains_pagename = item.term:find("%[%[" .. m_str_utils.pattern_escape(pagename) .. "[|%]]") if term_is_pagename or term_contains_pagename then local current_L2 = require(pages_module).get_current_L2() if current_L2 then local current_L2_lang = require(languages_module).getByCanonicalName(current_L2) if current_L2_lang and current_L2_lang:getCode() == item.lang:getCode() then if term_is_pagename then track("term-is-pagename") else track("term-contains-pagename") end end end end end item.suppress_redundant_wikilink_cat = suppress_redundant_wikilink_cat item.never_call_transliteration_module = nochecktr return require(links_module).full_link(item, face) end local function format_subitem(subitem, lang, face, compute_embedded_comma, nochecktr) local embedded_comma = false local text if subitem.term and term_already_linked(subitem.term) then text = subitem.term if compute_embedded_comma then embedded_comma = not not require(utilities_module).get_plaintext(text):find(",") end else text = full_link_and_track_self_links(subitem, face, nochecktr) if compute_embedded_comma then -- We don't check qualifier, label or reference text for commas as it's inside parens or displayed -- elsewhere. local subitem_plaintext = subitem.alt or subitem.term if subitem_plaintext then embedded_comma = not not subitem_plaintext:find(",") end end end -- We could use the "show qualifiers" flag to full_link() but not when term_already_linked(). if subitem.q and subitem.q[1] or subitem.qq and subitem.qq[1] or subitem.l and subitem.l[1] or subitem.ll and subitem.ll[1] or subitem.refs and subitem.refs[1] then text = require(pron_qualifier_module).format_qualifiers { lang = subitem.lang or args.lang, text = text, q = subitem.q, qq = subitem.qq, l = subitem.l, ll = subitem.ll, refs = subitem.refs, } end return text, embedded_comma end function export.format_item(item, args, face) local compute_embedded_comma = args.horiz == "comma" local embedded_comma = false local nochecktr = args.noautotr if type(item) == "table" then if item.terms then local parts = {} local is_first = true for _, subitem in ipairs(item.terms) do if subitem == false then -- omitted subitem; do nothing else local separator = subitem.separator or not is_first and (args.subitem_separator or ", ") if separator then if compute_embedded_comma then embedded_comma = embedded_comma or not not separator:find(",") end insert(parts, separator) end local formatted, this_embedded_comma = format_subitem(subitem, args.lang, face, compute_embedded_comma, nochecktr) embedded_comma = embedded_comma or this_embedded_comma insert(parts, formatted) is_first = false end end return concat(parts), embedded_comma else return format_subitem(item, args.lang, face, compute_embedded_comma, nochecktr) end else if compute_embedded_comma then embedded_comma = not not require(utilities_module).get_plaintext(item):find(",") end if args.lang and not term_already_linked(item) then return full_link_and_track_self_links({lang = args.lang, term = item, sc = args.sc}, face, nochecktr), embedded_comma else return item, embedded_comma end end end function export.construct_old_style_header(header, horiz) local old_style_header local function ib_colon() return tostring(html("span"):addClass("ib-colon"):addClass("ib-content"):wikitext(":")) end if horiz then old_style_header = require(qualifier_module).format_qualifiers { qualifiers = header, open = false, close = false, } .. ib_colon() .. " " else old_style_header = require(qualifier_module).format_qualifiers { qualifiers = header } .. ib_colon() old_style_header = tostring(html("div"):wikitext(old_style_header)) end return old_style_header end -- Construct the sort base of a single term. As a hack, sort appendices after mainspace items. local function term_sortbase(val) if not val then -- This should not normally happen. return u(0x10FFFF) elseif val:find("^%[*Appendix:") then return u(0x10FFFE) .. val else return val end end -- Construct the sort base of a single item, using the display form preferentially, otherwise the term itself. -- As a hack, sort appendices after mainspace items. local function item_sortbase(item) return term_sortbase(item.alt or item.term) end local function make_sortbase(item) if item == false then return "*" -- doesn't matter, will be omitted in create_list() elseif type(item) == "table" then if item.terms then -- Optimize for the common case of only a single term if item.terms[2] then local parts = {} -- multiple terms local first = true for _, subitem in ipairs(item.terms) do if subitem ~= false then if not first then insert(parts, ", ") end insert(parts, item_sortbase(subitem)) first = false end end if parts[1] then return concat(parts) end else local subitem = item.terms[1] if subitem ~= false then return item_sortbase(subitem) end end return "*" -- doesn't matter, entire group will be omitted in create_list() else return item_sortbase(item) end else return item end end local function make_node_sortbase(node) return make_sortbase(node.item) end -- Sort a sublist of `list` in place, keeping the first `keepfirst` and last `keeplast` items fixed. -- `lang` is the language of the items and `make_sortbase` creates the appropriate sort base. local function sort_sublist(list, lang, make_sortbase, keepfirst, keeplast) if keepfirst == 0 and keeplast == 0 then require(collation_module).sort(list, lang, make_sortbase) else local sublist = {} for i = keepfirst + 1, #list - keeplast do sublist[i - keepfirst] = list[i] end require(collation_module).sort(sublist, lang, make_sortbase) for i = keepfirst + 1, #list - keeplast do list[i] = sublist[i - keepfirst] end end end -- URL-encode only the characters that serve as template delimiters (left and right brace, vertical bar, equal sign -- and percent sign since it's the escape character). local function bot_url_encode(txt) return (txt:gsub("[%%|{}=&]", {["%"] = "%25", ["|"] = "%7C", ["{"] = "%7B", ["}"] = "%7D", ["="] = "%3D", ["&"] = "%26"})) end -- Reverse the action of bot_url_encode(). local function bot_url_decode(txt) return (txt:gsub("%%7([BCD])", {B = "{", C = "|", D = "}"}):gsub("%%3D", "="):gsub("%%26", "&"):gsub("%%25", "%%")) end --[==[ Bot-callable function to generate a number of sortkeys simultaneously. {{para|1}} contains the langcode, and remaining numeric parameters contain "bot-URL-encoded" strings whose sort keys will be computed and returned as a JSON array. Here, "bot-URL-encoded" means that the six characters `{ | } = & %` should be converted to their URL-encoded representation (respectively <code>%7B %7C %7D %3D %26 %25</code>), and will be decoded appropriately before computing the sortkey. ]==] function export.make_sortkey(frame) local iparams = { [1] = {type = "language"}, [2] = {list = true}, } local iargs = require(parameters_module).process(frame.args, iparams) local make_sortkey = require(collation_module).make_lang_sortkey_function(iargs[1], term_sortbase) local retval = {} for _, arg in ipairs(iargs[2]) do arg = bot_url_decode(arg) insert(retval, make_sortkey(arg)) end return require(JSON_module).toJSON(retval) end local large_text_scripts = { ["Arab"] = true, ["Beng"] = true, ["Deva"] = true, ["Gujr"] = true, ["Guru"] = true, ["Hebr"] = true, ["Khmr"] = true, ["Knda"] = true, ["Laoo"] = true, ["Mlym"] = true, ["Mong"] = true, ["Mymr"] = true, ["Orya"] = true, ["Sinh"] = true, ["Syrc"] = true, ["Taml"] = true, ["Telu"] = true, ["Tfng"] = true, ["Thai"] = true, ["Tibt"] = true, } --[==[ Format a list of items using HTML. `args` is an object specifying the items to add and related properties, with the following fields: * `content`: A list of the items to format. See below for the format of the items. * `lang`: The language object of the items to format, if the items in `content` are strings. * `sc`: The script object of the items to format, if the items in `content` are strings. * `raw`: If true, return the list raw, without any collapsing or columns. * `class`: The CSS class of the surrounding <div>. * `column_count`: Number of columns to format the list into. * `alphabetize`: If true, sort the items in the table. * `collapse`: If true, make the table partially collapsed by default, with a "Show more" button at the bottom. * `toggle_category`: Value of `data-toggle-category` property grouping collapsible elements. * `header`: If specified, Wikicode to prepend to the output. * `title_new_style`: If true, the header is treated as a title and displayed in a new style. This is ignored if `horiz` is non-nil. * `subitem_separator`: Separator used between subitems when multiple subitems occur on a line, if not specified in the subitem itself (using the `separator` field). Defaults to {", "}. * `keepfirst`: If > 0, keep this many rows unsorted at the beginning of the top level. * `keeplast`: If > 0, keep this many rows unsorted at the end of the top level. * `horiz`: If non-nil, format the items horizontally. If the value is "bullet", put a center dot/bullet (·) between items. If the value is "comma", put a comma between items (but if there is an embedded comma in any item, put a semicolon between all items). Each item in `content` is in one of the following formats: * A string. This is for compatibility and should not be used by new callers. * An object describing an item to format, in the format expected by full_link() in [[Module:links]] but can also have left or right qualifiers, left or right labels, or references. * An object describing a list of subitems to format, displayed side-by-side, separated by a comma or other separator. This format is identified by the presence of a key `terms` specifying the list of subitems. Each subitem is in the same format as for a single top-level item, except that it should also have a `separator` field specifying the separator to display before each item (which will typically be a blank string before the first item). ]==] function export.create_list(args) if type(args) ~= "table" then error("expected table, got " .. type(args)) end local column_count = args.column_count or 1 local toggle_category = args.toggle_category or "derived terms" local keepfirst = args.keepfirst or 0 local keeplast = args.keeplast or 0 if keepfirst > 0 then track("keepfirst") end if keeplast > 0 then track("keeplast") end -- maybe construct old-style header local old_style_header = nil if args.header and (args.horiz or not args.title_new_style) then old_style_header = export.construct_old_style_header(args.header, args.horiz) end if args.horiz then old_style_header = "* " .. (old_style_header or "") end local list local any_extra_indented_item = false for _, item in ipairs(args.content) do if item == false then -- do nothing elseif type(item) == "table" and item.extra_indent and item.extra_indent > 0 then any_extra_indented_item = true break end end -- If any extra indented item, convert the items to a nested structure, which is necessary both for sorting and -- for converting to HTML. if any_extra_indented_item then local function make_node(item) return { item = item } end local root_node = make_node(nil) local node_stack = {root_node} local last_indent = 0 local function append_subnode(node, subnode) if not node.subnodes then node.subnodes = {} end insert(node.subnodes, subnode) end for i, item in ipairs(args.content) do if item == false then -- do nothing else local this_indent if type(item) ~= "table" then this_indent = 1 else this_indent = (item.extra_indent or 0) + 1 end local node = make_node(item) if this_indent == last_indent then append_subnode(node_stack[#node_stack], node) elseif this_indent > last_indent + 1 then error(("Element #%s (%s) has indent %s, which is more than one greater than the previous item with indent %s"):format( i, make_sortbase(item), this_indent, last_indent)) elseif this_indent > last_indent then -- Start a new sublist attached to the last item of the sublist one level up; but we need special -- handling for the root node (last_indent == 0). if last_indent > 0 then local subnodes = node_stack[#node_stack].subnodes if not subnodes then error(("Internal error: Not first item and no subnodes at preceding level %s: %s"):format( #node_stack, dump(node_stack))) end insert(node_stack, subnodes[#subnodes]) end append_subnode(node_stack[#node_stack], node) last_indent = this_indent else while last_indent > this_indent do local finished_node = table.remove(node_stack) if args.alphabetize then require(collation_module).sort(finished_node.subnodes, args.lang, make_node_sortbase) end last_indent = last_indent - 1 end append_subnode(node_stack[#node_stack], node) end end end if args.alphabetize then while node_stack[1] do local finished_node = table.remove(node_stack) if node_stack[1] then -- We're sorting something other than the root node. require(collation_module).sort(finished_node.subnodes, args.lang, make_node_sortbase) else -- We're sorting the root node; honor `keepfirst` and `keeplast`. sort_sublist(finished_node.subnodes, args.lang, make_node_sortbase, keepfirst, keeplast) end end end local function format_node(node, depth) local sublist local embedded_comma = false if node.subnodes then if args.horiz then sublist = {} else sublist = html("ul") end local prevnode = nil for _, subnode in ipairs(node.subnodes) do local thisnode, this_embedded_comma = format_node(subnode, depth + 1) embedded_comma = embedded_comma or this_embedded_comma if not prevnode or not args.alphabetize or not deepEquals(prevnode, thisnode) then if args.horiz then table.insert(sublist, thisnode) else sublist = sublist:node(thisnode) end prevnode = thisnode end end if args.horiz then sublist = table.concat(sublist, get_horizontal_separator(args.horiz, embedded_comma)) end end if not node.item then -- At the root. return sublist, embedded_comma end local formatted, listitem -- Ignore embedded commas in subitems inside of parens or square brackets. formatted, embedded_comma = export.format_item(node.item, args) if args.horiz then listitem = formatted if sublist then -- Use parens for the first, third, fifth, etc. sublists and square brackets for the remainder. if depth % 2 == 1 then listitem = ("%s (%s)"):format(listitem, sublist) else listitem = ("%s [%s]"):format(listitem, sublist) end end else listitem = html("li"):wikitext(formatted) if sublist then listitem = listitem:node(sublist) end end return listitem, embedded_comma end list = format_node(root_node, 0) else if args.alphabetize then sort_sublist(args.content, args.lang, make_sortbase, keepfirst, keeplast) end if args.horiz then list = {} else list = html("ul") end local previtem = nil local embedded_comma = false for _, item in ipairs(args.content) do if item == false then -- omitted item; do nothing else local thisitem, this_embedded_comma = export.format_item(item, args) embedded_comma = embedded_comma or this_embedded_comma if not previtem or not args.alphabetize or previtem ~= thisitem then if args.horiz then table.insert(list, thisitem) else list = list:node(html("li"):wikitext(thisitem)) end previtem = thisitem end end end if args.horiz then list = table.concat(list, get_horizontal_separator(args.horiz, embedded_comma)) end end local output if args.horiz then output = list else output = html("div"):addClass("term-list"):node(list) if args.class then output:addClass(args.class) end if not args.raw then output:addClass("ul-column-count") :attr("data-column-count", column_count) if args.collapse then output = html("div") :node(output) :addClass("list-switcher") :attr("data-toggle-category", toggle_category) -- identify commonly used scripts that use large text and -- provide a special CSS class to make the template bigger local sc = args.sc if sc == nil then local scripts = args.lang:getScripts() if #scripts > 0 then sc = scripts[1] end end if sc ~= nil then local scriptcode = sc:getParentCode() if scriptcode == "top" then scriptcode = sc:getCode() end if large_text_scripts[scriptcode] then output:addClass("list-switcher-large-text") end end end end if args.collapse or args.title_new_style then -- wrap in wrapper to prevent interference from floating elements local list_switcher_wrapper = html("div") :addClass("list-switcher-wrapper") if args.title_new_style then list_switcher_wrapper :node( html("div") :addClass("list-switcher-header") :wikitext(args.header) ) end list_switcher_wrapper:node(output) output = list_switcher_wrapper end output = tostring(output) end return (old_style_header or "") .. output end -- This function is for compatibility with earlier version of [[Module:columns]] -- (now found in [[Module:columns/old]]). function export.create_table(...) -- Earlier arguments to create_table: -- n_columns, content, alphabetize, bg, collapse, class, title, column_width, line_start, lang local args = {} args.column_count, args.content, args.alphabetize, args.collapse, args.class, args.header, args.column_width, args.line_start, args.lang = ... return export.create_list(args) end function export.display_from(frame_args, parent_args, frame) local boolean = {type = "boolean"} local iparams = { ["class"] = true, -- Default for auto-collapse. Overridable by template |collapse= param. ["collapse"] = boolean, -- If specified, this specifies the number of columns, and no columns parameter is available on the template. -- Otherwise, the columns parameter is named |n=. ["columns"] = {type = "number"}, -- If specified, this specifies the default language code, which can be overridden using |lang= in the template. -- Otherwise, the language-code parameter is required and normally found in |1=, but for compatibility can be -- specified as |lang= (which leads to deprecation handling). ["lang"] = {type = "language"}, -- Default for auto-sort. Overridable by template |sort= param. ["sort"] = boolean, ["toggle_category"] = true, -- Minimum number of rows required to format into a multicolumn list. If below this, the list is displayed "raw" -- (no columns, no collapsbility). ["minrows"] = {type = "number", default = 5}, -- Disables automatic transliteration; entries without a manual transliteration will have none at all. -- Used on large pages, especially Chinese ones, because zh-translit works by fetching and parsing -- the target of the page, which is a performance killer on large pages with potentially thousands -- of link targets. -- Note: noautotr also disables redundant transliteration checks. ["noautotr"] = boolean, } local iargs = require(parameters_module).process(frame_args, iparams) local langcode_in_lang = iargs.lang or parent_args.lang local lang_param = langcode_in_lang and "lang" or 1 local deprecated = not iargs.lang and langcode_in_lang local ret = export.handle_display_from_or_topic_list(iargs, parent_args, nil) return deprecated and frame:expandTemplate{title = "check deprecated lang param usage", args = {ret, lang = args[lang_param]}} or ret end --[==[ Implement `display_from()` [the internal entry point for {{tl|col}} and variants, which enter originally through `display()`] as well as regular (column-oriented) topic lists, invoked through [[Module:topic list]]. `iargs` are the invocation args of {{tl|col}}, and `raw_item_args` are the arguments specifying the values of each row as well as other properties, corresponding to the user-specified template arguments of {{tl|col}}. Note that `show()` in [[Module:topic list]] is normally invoked directly by a topic list template, whose invocation arguments are passed in using `raw_item_args` and are similar to the template arguments of {{tl|col}}. `iargs` for topic-list invocations is hard-coded, and template arguments to a topic-list template are processed in [[Module:topic list]] itself. Note that the handling of topic lists is currently implemented almost entirely through callbacks in `topic_list_data` (which is nil if we're processing {{tl|col}} rather than a topic list) in an attempt to reduce the coupling and keep the topic-list-specific code in [[Module:topic list]], but IMO the coupling is still too tight. Probably the control structure should be reversed and the following function split up into subfunctions, which are invoked as needed by {{tl|col}} and/or [[Module:topic list]]. ]==] function export.handle_display_from_or_topic_list(iargs, raw_item_args, topic_list_data) local boolean = {type = "boolean"} local langcode_in_lang = iargs.lang or raw_item_args.lang local lang_param = langcode_in_lang and "lang" or 1 local first_content_param = langcode_in_lang and 1 or 2 local params = { [lang_param] = {required = not iargs.lang, type = "language", template_default = not iargs.lang and "und" or nil}, ["n"] = not iargs.columns and {type = "number"} or nil, [first_content_param] = {list = true, allow_holes = true}, ["title"] = {}, ["collapse"] = boolean, ["sort"] = boolean, ["sc"] = {type = "script"}, -- used when calling from [[Module:saurus]] so the page displaying the synonyms/antonyms doesn't occur in the -- list ["omit"] = {list = true}, ["keepfirst"] = {type = "number", default = 0}, ["keeplast"] = {type = "number", default = 0}, ["horiz"] = {}, ["notr"] = boolean, ["noautotr"] = boolean, ["allow_space_delim"] = boolean, ["tilde_delim"] = {}, ["space_delim"] = {}, ["comma_delim"] = {}, } if topic_list_data then topic_list_data.add_topic_list_params(params) end local m_param_utils = require(parameter_utilities_module) local param_mods = m_param_utils.construct_param_mods { {default = true, require_index = true}, {group = "link"}, -- sc has separate_no_index = true; that's the only one -- It makes no sense to have overall l=, ll=, q= or qq= params for columnar display. {group = {"ref", "l", "q"}, require_index = true}, } m_param_utils.augment_params_with_modifiers(params, param_mods) local processed_args = require(parameters_module).process(raw_item_args, params) local horiz = processed_args.horiz if horiz and horiz ~= "comma" and horiz ~= "bullet" then horiz = require(yesno_module)(horiz) if horiz == nil then error(("Unrecognized value |horiz=%s; should be 'comma', 'bullet' or a recognized Boolean value such " .. "as 'yes' or '1' (same as 'bullet') or 'no' or '0'"):format(processed_args.horiz)) end if horiz == true then horiz = "bullet" end processed_args.horiz = horiz end -- If default argument values specified, set them after parsing the caller-specified arguments in `raw_item_args`. if topic_list_data then topic_list_data.set_default_arguments(processed_args) end -- Now set defaults for the various delimiters, depending in some cases on whether horiz was set. -- We can't set these defaults (even regardless of their dependency on horiz=) in `local params` above -- because we want any defaults specified in `default_props` to override these. if not processed_args.tilde_delim then local tilde_with_abbr = '<abbr title="near equivalent">~</abbr>' processed_args.tilde_delim = processed_args.horiz and tilde_with_abbr or " " .. tilde_with_abbr .. " " end if not processed_args.space_delim then processed_args.space_delim = "&nbsp;" end if not processed_args.comma_delim then processed_args.comma_delim = processed_args.horiz and "/" or ", " end -- Check for extra term indent. Do this before calling parse_list_with_inline_modifiers_and_separate_params() -- because sometimes space is a delimiter and the space in the indent will confuse things and get interpreted as a -- delimiter. local extra_indent_by_termno = {} local termargs = processed_args[first_content_param] for i = 1, termargs.maxindex do local term = termargs[i] if term then local extra_indent, actual_term = rmatch(term, "^(%*+)%s+(.-)$") if extra_indent then termargs[i] = actual_term extra_indent_by_termno[i] = #extra_indent end end end local groups, args = m_param_utils.parse_list_with_inline_modifiers_and_separate_params { param_mods = param_mods, processed_args = processed_args, termarg = first_content_param, parse_lang_prefix = true, allow_multiple_lang_prefixes = true, disallow_custom_separators = true, track_module = "columns", lang = iargs.lang or lang_param, sc = "sc.default", splitchar = processed_args.allow_space_delim and "[,~ ]" or "[,~]", } local lang = iargs.lang or args[lang_param] local langcode = lang:getCode() local fulllangcode = lang:getFullCode() local sc = args.sc.default local sort = iargs.sort if args.sort ~= nil then if not args.sort then track("nosort") end sort = args.sort else -- HACK! For Japanese-script languages (Japanese, Okinawan, Miyako, etc.), sorting doesn't yet work properly, so -- disable it. for _, langsc in ipairs(lang:getScriptCodes()) do if langsc == "Jpan" then sort = false break end end end local collapse = iargs.collapse if args.collapse ~= nil then if not args.collapse then track("nocollapse") end collapse = args.collapse end local title = args.title local formatted_cats if topic_list_data then title, formatted_cats = topic_list_data.get_title_and_formatted_cats(args, lang, sc, topic_list_data) end local number_of_groups = 0 for i, group in ipairs(groups) do local number_of_items = 0 group.extra_indent = extra_indent_by_termno[group.orig_index] for j, item in ipairs(group.terms) do convert_delimiter_to_separator(item, j, args) if args.notr then item.tr = "-" elseif args.noautotr then item.tr = item.tr or "-" end -- If a separate language code was given for the term, display the language name as a right qualifier. -- (Briefly we made them labels but this leads to non-obvious behavior e.g. "French" becoming "France" under -- some circumstances.) Otherwise it may not be obvious that the term is in a separate language (e.g. if the -- main language is 'zh' and the term language is a Chinese lect such as Min Nan). But don't do this for -- Translingual terms, which are often added to the list of English and other-language terms. if item.termlangs then local qqs = {} for _, termlang in ipairs(item.termlangs) do local termlangcode = termlang:getCode() if termlangcode ~= langcode and termlangcode ~= "mul" then insert(qqs, termlang:getCanonicalName()) end end if item.qq then for _, qq in ipairs(item.qq) do insert(qqs, qq) end end item.qq = qqs end local omitted = false for _, omitted_item in ipairs(args.omit) do if omitted_item == item.term then omitted = true break end end if omitted then -- signal create_list() to omit this item group.terms[j] = false else number_of_items = number_of_items + 1 end end if number_of_items == 0 then -- omit the whole group groups[i] = false else number_of_groups = number_of_groups + 1 end end local column_count = iargs.columns or args.n -- FIXME: This needs a total rewrite. if column_count == nil then column_count = number_of_groups <= 3 and 1 or number_of_groups <= 9 and 2 or number_of_groups <= 27 and 3 or number_of_groups <= 81 and 4 or 5 end local raw = number_of_groups < iargs.minrows local horiz_edit_button if topic_list_data and args.horiz then -- append edit button to title horiz_edit_button = topic_list_data.make_horiz_edit_button(topic_list_data.topic_list_template) end return export.create_list { column_count = column_count, raw = raw, content = groups, alphabetize = sort, header = title, title_new_style = (title ~= nil and title ~= ''), collapse = collapse, toggle_category = iargs.toggle_category, -- columns-bg (in [[MediaWiki:Gadget-Site.css]]) provides the background color class = (iargs.class and iargs.class .. " columns-bg" or "columns-bg"), lang = lang, sc = sc, subitem_separator = ", ", keepfirst = args.keepfirst, keeplast = args.keeplast, horiz = args.horiz, noautotr = args.noautotr, } .. (horiz_edit_button or "") .. (formatted_cats or "") end function export.display(frame) if not is_substing() then return export.display_from(frame.args, frame:getParent().args, frame, false) end -- If substed, unsubst template with newlines between each term, redundant wikilinks removed, and remove duplicates + sort terms if sort is enabled. local m_table = require("Module:table") local m_template_parser = require("Module:template parser") local parent = frame:getParent() local elems = m_table.shallowCopy(parent.args) local code = remove(elems, 1) code = code and trim(code) local lang = require("Module:languages").getByCode(code, 1) local i = 1 while true do local elem = elems[i] while elem do elem = trim(elem, "%s") if elem ~= "" then break end remove(elems, i) elem = elems[i] end if not elem then break elseif not ( -- Strip redundant wikilinks. not elem:match("^()%[%[") or elem:find("[[", 3, true) or elem:find("]]", 3, true) ~= #elem - 1 or elem:find("|", 3, true) ) then elem = sub(elem, 3, -3) elem = trim(elem, "%s") end elems[i] = elem .. "\n" i = i + 1 end -- If sort is enabled, remove duplicates then sort elements. if require("Module:yesno")(frame.args.sort) then elems = m_table.removeDuplicates(elems) require("Module:collation").sort(elems, lang) end -- Readd the langcode. insert(elems, 1, code .. "\n") -- TODO: Place non-numbered parameters after 1 and before 2. local template = m_template_parser.getTemplateInvocationName(mw.title.new(parent:getTitle())) return "{{" .. concat(m_template_parser.buildTemplate(template, elems), "|") .. "}}" end return export g1591p4wiwxvv6n8shajus3a4el65bm 828 34332 184520 2026-05-21T06:03:35Z P1ayer 1197 创建页面，内容为“local export = {} local function_module = "Module:fun" local load_module = "Module:load" local memoize_module = "Module:memoize" local string_char_module = "Module:string/char" local string_charset_escape_module = "Module:string/charsetEscape" local mw = mw local string = string local table = table local ustring = mw.ustring local byte = string.byte local char = string.char local concat = table.concat local find = string.find local format = string.format loc…” 184520 Scribunto text/plain local export = {} local function_module = "Module:fun" local load_module = "Module:load" local memoize_module = "Module:memoize" local string_char_module = "Module:string/char" local string_charset_escape_module = "Module:string/charsetEscape" local mw = mw local string = string local table = table local ustring = mw.ustring local byte = string.byte local char = string.char local concat = table.concat local find = string.find local format = string.format local gmatch = string.gmatch local gsub = string.gsub local insert = table.insert local len = string.len local lower = string.lower local match = string.match local next = next local require = require local reverse = string.reverse local select = select local sort = table.sort local sub = string.sub local tonumber = tonumber local tostring = tostring local type = type local ucodepoint = ustring.codepoint local ufind = ustring.find local ugcodepoint = ustring.gcodepoint local ugmatch = ustring.gmatch local ugsub = ustring.gsub local ulower = ustring.lower local umatch = ustring.match local unpack = unpack or table.unpack -- Lua 5.2 compatibility local upper = string.upper local usub = ustring.sub local uupper = ustring.upper local memoize = require(memoize_module) -- Defined below. local codepoint local explode_utf8 local format_fun local get_charset local gsplit local pattern_escape local pattern_simplifier local replacement_escape local title_case local trim local ucfirst local ulen --[==[ Loaders for functions in other modules, which overwrite themselves with the target function when called. This ensures modules are only loaded when needed, retains the speed/convenience of locally-declared pre-loaded functions, and has no overhead after the first call, since the target functions are called directly in any subsequent calls. ]==] local function charset_escape(...) charset_escape = require(string_charset_escape_module) return charset_escape(...) end local function is_callable(...) is_callable = require(function_module).is_callable return is_callable(...) end local function load_data(...) load_data = require(load_module).load_data return load_data(...) end local function u(...) u = require(string_char_module) return u(...) end local function prepare_iter(str, pattern, str_lib, plain) local callable = is_callable(pattern) if str_lib or plain then return pattern, #str, string, callable elseif not callable then local simple = pattern_simplifier(pattern) if simple then return simple, #str, string, false end end return pattern, ulen(str), ustring, callable end --[==[ Returns {nil} if the input value is the empty string, or otherwise the same value. If the input is a string and `do_trim` is set, the input value will be trimmed before returning; if the trimmed value is the empty string, returns {nil}. If `quote_delimiters` is set, then any outer pair of quotation marks ({' '} or {" "}) surrounding the rest of the input string will be stripped, if present. The string will not be trimmed again, converted to {nil}, or have further quotation marks stripped, as it exists as a way to embed spaces or the empty string in an input. Genuine quotation marks may also be embedded this way (e.g. {"''foo''"} returns {"'foo'"}). ]==] function export.is_not_empty(str, do_trim, quote_delimiters) if str == "" then return nil elseif not (str and type(str) == "string") then return str elseif do_trim then str = trim(str) if str == "" then return nil end end return quote_delimiters and gsub(str, "^(['\"])(.*)%1$", "%2") or str end --[==[ Explodes a string into an array of UTF-8 characters. '''Warning''': this function assumes that the input is valid UTF-8 in order to optimize speed and memory use. Passing in an input containing non-UTF-8 byte sequences could result in unexpected behaviour. ]==] function export.explode_utf8(str) local text, i = {}, 0 for ch in gmatch(str, ".[\128-\191]*") do i = i + 1 text[i] = ch end return text end explode_utf8 = export.explode_utf8 --[==[ Returns {true} if `str` is a valid UTF-8 string. This is true if, for each character, all of the following are true: * It has the expected number of bytes, which is determined by value of the leading byte: 1-byte characters are `0x00` to `0x7F`, 2-byte characters start with `0xC2` to `0xDF`, 3-byte characters start with `0xE0` to `0xEF`, and 4-byte characters start with `0xF0` to `0xF4`. * The leading byte must not fall outside of the above ranges. * The trailing byte(s) (if any), must be between `0x80` to `0xBF`. * The character's codepoint must be between U+0000 (`0x00`) and U+10FFFF (`0xF4 0x8F 0xBF 0xBF`). * The character cannot have an overlong encoding: for each byte length, the lowest theoretical encoding is equivalent to U+0000 (e.g. `0xE0 0x80 0x80`, the lowest theoretical 3-byte encoding, is exactly equivalent to U+0000). Encodings that use more than the minimum number of bytes are not considered valid, meaning that the first valid 3-byte character is `0xE0 0xA0 0x80` (U+0800), and the first valid 4-byte character is `0xF0 0x90 0x80 0x80` (U+10000). Formally, 2-byte characters have leading bytes ranging from `0xC0` to `0xDF` (rather than `0xC2` to `0xDF`), but `0xC0 0x80` to `0xC1 0xBF` are overlong encodings, so it is simpler to say that the 2-byte range begins at `0xC2`. If `allow_surrogates` is set, surrogates (U+D800 to U+DFFF) will be treated as valid UTF-8. Surrogates are used in UTF-16, which encodes codepoints U+0000 to U+FFFF with 2 bytes, and codepoints from U+10000 upwards using a pair of surrogates, which are taken together as a 4-byte unit. Since surrogates have no use in UTF-8, as it encodes higher codepoints in a different way, they are not considered valid in UTF-8 text. However, there are limited circumstances where they may be necessary: for instance, JSON escapes characters using the format `\u0000`, which must contain exactly 4 hexadecimal digits; under the scheme, codepoints above U+FFFF must be escaped as the equivalent pair of surrogates, even though the text itself must be encoded in UTF-8 (e.g. U+10000 becomes `\uD800\uDC00`). ]==] function export.isutf8(str, allow_surrogates) for ch in gmatch(str, "[\128-\255][\128-\191]*") do if #ch > 4 then return false end local b1, b2, b3, b4 = byte(ch, 1, 4) if not (b2 and b2 >= 0x80 and b2 <= 0xBF) then return false -- 1-byte is always invalid, as gmatch excludes 0x00 to 0x7F elseif not b3 then -- 2-byte if not (b1 >= 0xC2 and b1 <= 0xDF) then -- b1 == 0xC0 or b1 == 0xC1 is overlong return false end elseif not (b3 >= 0x80 and b3 <= 0xBF) then -- trailing byte return false elseif not b4 then -- 3-byte if b1 > 0xEF then return false elseif b2 < 0xA0 then if b1 < 0xE1 then -- b1 == 0xE0 and b2 < 0xA0 is overlong return false end elseif b1 < 0xE0 or (b1 == 0xED and not allow_surrogates) then -- b1 == 0xED and b2 >= 0xA0 is a surrogate return false end elseif not (b4 >= 0x80 and b4 <= 0xBF) then -- 4-byte return false elseif b2 < 0x90 then if not (b1 >= 0xF1 and b1 <= 0xF4) then -- b1 == 0xF0 and b2 < 0x90 is overlong return false end elseif not (b1 >= 0xF0 and b1 <= 0xF3) then -- b1 == 0xF4 and b2 >= 0x90 is too high return false end end return true end do local charset_chars = { ["\0"] = "%z", ["%"] = "%%", ["-"] = "%-", ["]"] = "%]", ["^"] = "%^" } charset_chars.__index = charset_chars local chars = setmetatable({ ["$"] = "%$", ["("] = "%(", [")"] = "%)", ["*"] = "%*", ["+"] = "%+", ["."] = "%.", ["?"] = "%?", ["["] = "%[" }, charset_chars) --[==[ Escapes the magic characters used in a [[mw:Extension:Scribunto/Lua reference manual#Patterns|pattern]] (Lua's version of regular expressions): {$%()*+-.?[]^}, and converts the null character to {%z}. For example, {"^$()%.[]*+-?\0"} becomes {"%^%$%(%)%%%.%[%]%*%+%-%?%z"}. This is necessary when constructing a pattern involving arbitrary text (e.g. from user input). ]==] function export.pattern_escape(str) return (gsub(str, "[%z$%%()*+%-.?[%]^]", chars)) end pattern_escape = export.pattern_escape --[==[ Escapes only {%}, which is the only magic character used in replacement [[mw:Extension:Scribunto/Lua reference manual#Patterns|patterns]] with string.gsub and mw.ustring.gsub. ]==] function export.replacement_escape(str) return (gsub(str, "%%", "%%%%")) end replacement_escape = export.replacement_escape local function case_insensitive_char(ch) local upper_ch = uupper(ch) if upper_ch == ch then ch = ulower(ch) if ch == upper_ch then return chars[ch] or ch end end return "[" .. (charset_chars[upper_ch] or upper_ch) .. (charset_chars[ch] or ch) .. "]" end local function iterate(str, str_len, text, n, start, _gsub, _sub, loc1, loc2) if not (loc1 and start <= str_len) then -- Add final chunk and return. n = n + 1 text[n] = _gsub(_sub(str, start), ".", chars) return elseif loc2 < loc1 then if _sub == sub then local b = byte(str, loc1) if b and b >= 128 then loc1 = loc1 + (b < 224 and 1 or b < 240 and 2 or 3) end end n = n + 1 text[n] = _gsub(_sub(str, start, loc1), ".", chars) start = loc1 + 1 if start > str_len then return end else -- Add chunk up to the current match. n = n + 1 text[n] = _gsub(_sub(str, start, loc1 - 1), ".", chars) -- Add current match. n = n + 1 text[n] = _gsub(_sub(str, loc1, loc2), ".", case_insensitive_char) start = loc2 + 1 end return n, start end --[==[ Escapes the magic characters used in a [[mw:Extension:Scribunto/Lua reference manual#Patterns|pattern]], and makes all characters case-insensitive. An optional pattern or find function (see {split}) may be supplied as the second argument, the third argument (`str_lib`) forces use of the string library, while the fourth argument (`plain`) turns any pattern matching facilities off in the optional pattern supplied. ]==] function export.case_insensitive_pattern(str, pattern_or_func, str_lib, plain) if pattern_or_func == nil then return (gsub(str, str_lib and "[^\128-\255]" or ".[\128-\191]*", case_insensitive_char)) end local text, n, start, str_len, _string, callable = {}, 0, 1 pattern_or_func, str_len, _string, callable = prepare_iter(str, pattern_or_func, str_lib, plain) local _find, _gsub, _sub = _string.find, _string.gsub, _string.sub if callable then repeat n, start = iterate(str, str_len, text, n, start, _gsub, _sub, pattern_or_func(str, start)) until not start -- Special case if the pattern is anchored to the start: "^" always -- anchors to the start position, not the start of the string, so get -- around this by only attempting one match with the pattern, then match -- the end of the string. elseif byte(pattern_or_func) == 0x5E then -- ^ n, start = iterate(str, str_len, text, n, start, _gsub, _sub, _find(str, pattern_or_func, start, plain)) if start ~= nil then iterate(str, str_len, text, n, start, _gsub, _sub, _find(str, "$", start, plain)) end else repeat n, start = iterate(str, str_len, text, n, start, _gsub, _sub, _find(str, pattern_or_func, start, plain)) until not start end return concat(text) end end do local character_classes local function get_character_classes() character_classes, get_character_classes = { [0x41] = true, [0x61] = true, -- Aa [0x43] = true, [0x63] = true, -- Cc [0x44] = true, [0x64] = true, -- Dd [0x4C] = true, [0x6C] = true, -- Ll [0x50] = true, [0x70] = true, -- Pp [0x53] = true, [0x73] = true, -- Ss [0x55] = true, [0x75] = true, -- Uu [0x57] = true, [0x77] = true, -- Ww [0x58] = true, [0x78] = true, -- Xx [0x5A] = true, -- z dealt with separately. }, nil return character_classes end local function check_sets_equal(set1, set2) local k2 for k1, v1 in next, set1 do local v2 = set2[k1] if v1 ~= v2 and (v2 == nil or not check_sets_equal(v1, v2)) then return false end k2 = next(set2, k2) end return next(set2, k2) == nil end local function check_sets(bytes) local key, set1, set = next(bytes) if set1 == true then return true elseif not check_sets(set1) then return false end while true do key, set = next(bytes, key) if not key then return true elseif not check_sets_equal(set, set1) then return false end end end local function make_charset(range) if #range == 1 then return char(range[1]) end sort(range) local compressed, n, start = {}, 0, range[1] for i = 1, #range do local this, nxt = range[i], range[i + 1] if nxt ~= this + 1 then n = n + 1 compressed[n] = this == start and char(this) or char(start) .. "-" .. char(this) start = nxt end end return "[" .. concat(compressed) .. "]" end local function parse_1_byte_charset(pattern, pos) local ch while true do pos, ch = match(pattern, "()([%%%]\192-\255])", pos) if ch == "%" then local nxt = byte(pattern, pos + 1) if not nxt or nxt >= 128 or (character_classes or get_character_classes())[nxt] then -- acdlpsuwxACDLPSUWXZ, but not z return false end pos = pos + 2 elseif ch == "]" then pos = pos + 1 return pos else return false end end end --[==[ Parses `pattern`, a ustring library pattern, and attempts to convert it into a string library pattern. If conversion isn't possible, returns false. ]==] function pattern_simplifier(pattern) if type(pattern) == "number" then return tostring(pattern) end local pos, capture_groups, start, n, output, ch, nxt_pos = 1, 0, 1, 0 while true do -- FIXME: use "()([%%(.[\128-\255])[\128-\191]?[\128-\191]?[\128-\191]?()" and ensure non-UTF8 always fails. pos, ch, nxt_pos = match(pattern, "()([%%(.[\192-\255])[\128-\191]*()", pos) if not ch then break end local nxt = byte(pattern, nxt_pos) if ch == "%" then if nxt == 0x62 then -- b local nxt2, nxt3 = byte(pattern, pos + 2, pos + 3) if not (nxt2 and nxt2 < 128 and nxt3 and nxt3 < 128) then return false end pos = pos + 4 elseif nxt == 0x66 then -- f nxt_pos = nxt_pos + 2 local nxt2, nxt3 = byte(pattern, nxt_pos - 1, nxt_pos) -- Only possible to convert a positive %f charset which is -- all ASCII, so use parse_1_byte_charset. if not (nxt2 == 0x5B and nxt3 and nxt3 ~= 0x5E and nxt3 < 128) then -- [^ return false elseif nxt3 == 0x5D then -- Initial ] is non-magic. nxt_pos = nxt_pos + 1 end pos = parse_1_byte_charset(pattern, nxt_pos) if not pos then return false end elseif nxt == 0x5A then -- Z nxt = byte(pattern, nxt_pos + 1) if nxt == 0x2A or nxt == 0x2D then -- *- pos = pos + 3 else if output == nil then output = {} end local ins = sub(pattern, start, pos - 1) .. "[\1-\127\192-\255]" n = n + 1 if nxt == 0x2B then -- + output[n] = ins .. "%Z*" pos = pos + 3 elseif nxt == 0x3F then -- ? output[n] = ins .. "?[\128-\191]*" pos = pos + 3 else output[n] = ins .. "[\128-\191]*" pos = pos + 2 end start = pos end elseif not nxt or (character_classes or get_character_classes())[nxt] then -- acdlpsuwxACDLPSUWX, but not Zz return false -- Skip the next character if it's ASCII. Otherwise, we will -- still need to do length checks. else pos = pos + (nxt < 128 and 2 or 1) end elseif ch == "(" then if nxt == 0x29 or capture_groups == 32 then -- ) return false end capture_groups = capture_groups + 1 pos = pos + 1 elseif ch == "." then if nxt == 0x2A or nxt == 0x2D then -- *- pos = pos + 2 else if output == nil then output = {} end local ins = sub(pattern, start, pos - 1) .. "[^\128-\191]" n = n + 1 if nxt == 0x2B then -- + output[n] = ins .. ".*" pos = pos + 2 elseif nxt == 0x3F then -- ? output[n] = ins .. "?[\128-\191]*" pos = pos + 2 else output[n] = ins .. "[\128-\191]*" pos = pos + 1 end start = pos end elseif ch == "[" then -- Fail negative charsets. TODO: 1-byte charsets should be safe. if nxt == 0x5E then -- ^ return false -- If the first character is "%", ch_len is determined by the -- next one instead. elseif nxt == 0x25 then -- % nxt = byte(pattern, nxt_pos + 1) elseif nxt == 0x5D then -- Initial ] is non-magic. nxt_pos = nxt_pos + 1 end if not nxt then return false end local ch_len = nxt < 128 and 1 or nxt < 224 and 2 or nxt < 240 and 3 or 4 if ch_len == 1 then -- Single-byte charset. pos = parse_1_byte_charset(pattern, nxt_pos) if not pos then return false end else -- Multibyte charset. -- TODO: 1-byte chars should be safe to mix with multibyte chars. CONFIRM THIS FIRST. local charset_pos, bytes = pos pos = pos + 1 while true do -- TODO: non-ASCII charset ranges. pos, ch, nxt_pos = match(pattern, "^()([^\128-\191])[\128-\191]*()", pos) -- If escaped, get the next character. No need to -- distinguish magic characters or character classes, -- as they'll all fail for having the wrong length -- anyway. if ch == "%" then pos, ch, nxt_pos = match(pattern, "^()([^\128-\191])[\128-\191]*()", nxt_pos) elseif ch == "]" then pos = nxt_pos break end if not (ch and nxt_pos - pos == ch_len) then return false elseif bytes == nil then bytes = {} end local bytes, last = bytes, nxt_pos - 1 for i = pos, last - 1 do local b = byte(pattern, i) local bytes_b = bytes[b] if bytes_b == nil then bytes_b = {} bytes[b] = bytes_b end bytes[b], bytes = bytes_b, bytes_b end bytes[byte(pattern, last)] = true pos = nxt_pos end if not pos then return false end nxt = byte(pattern, pos) if ( (nxt == 0x2A or nxt == 0x2D or nxt == 0x3F) or -- *-? (nxt == 0x2B and ch_len > 2) or -- + not check_sets(bytes) ) then return false end local ranges, b, key, next_byte = {}, 0 repeat key, next_byte = next(bytes) local range, n = {key}, 1 -- Loop starts on the second iteration. for key in next, bytes, key do n = n + 1 range[n] = key end b = b + 1 ranges[b] = range bytes = next_byte until next_byte == true if nxt == 0x2B then -- + local range1, range2 = ranges[1], ranges[2] ranges[1], ranges[3] = make_charset(range1), make_charset(range2) local n = #range2 for i = 1, #range1 do n = n + 1 range2[n] = range1[i] end ranges[2] = make_charset(range2) .. "*" pos = pos + 1 else for i = 1, #ranges do ranges[i] = make_charset(ranges[i]) end end if output == nil then output = {} end nxt = byte(pattern, pos) n = n + 1 output[n] = sub(pattern, start, charset_pos - 1) .. concat(ranges) .. ((nxt == 0x2A or nxt == 0x2B or nxt == 0x2D or nxt == 0x3F) and "%" or "") -- following *+-? now have to be escaped start = pos end elseif not nxt then break elseif nxt == 0x2B then -- + if nxt_pos - pos ~= 2 then return false elseif output == nil then output = {} end pos, nxt_pos = pos + 1, nxt_pos + 1 nxt = byte(pattern, nxt_pos) local ch2 = sub(pattern, pos, pos) n = n + 1 output[n] = sub(pattern, start, pos - 1) .. "[" .. ch .. ch2 .. "]*" .. ch2 .. ((nxt == 0x2A or nxt == 0x2B or nxt == 0x2D or nxt == 0x3F) and "%" or "") -- following *+-? now have to be escaped pos, start = nxt_pos, nxt_pos elseif nxt == 0x2A or nxt == 0x2D or nxt == 0x3F then -- *-? return false else pos = nxt_pos end end if start == 1 then return pattern end return concat(output) .. sub(pattern, start) end pattern_simplifier = memoize(pattern_simplifier, true) export.pattern_simplifier = pattern_simplifier end --[==[ Parses `charset`, the interior of a string or ustring library character set, and normalizes it into a string or ustring library pattern (e.g. {"abcd-g"} becomes {"[abcd-g]"}, and {"[]"} becomes {"[[%]]"}). The negative (`^`), range (`-`) and literal (`%`) magic characters work as normal, and character classes may be used (e.g. `%d` and `%w`), but opening and closing square brackets are sanitized so that they behave like ordinary characters. ]==] function get_charset(charset) if type(charset) == "number" then return tostring(charset) end local pos, start, n, output = 1, 1, 0 if byte(charset) == 0x5E then -- ^ pos = pos + 1 end -- FIXME: "]" is non-magic if it's the first character in a charset. local nxt_pos, nxt while true do local new_pos, ch = match(charset, "()([%%%-%]])", pos) if not ch then break -- Skip percent escapes. Ranges can't start with them, either. elseif ch == "%" then pos = new_pos + 2 else -- If `ch` is a hyphen, get the character before iff it's at or ahead of `pos`. if ch == "-" and new_pos > pos then pos, nxt_pos, nxt = new_pos - 1, new_pos, ch ch = sub(charset, pos, pos) else pos, nxt_pos = new_pos, new_pos + 1 nxt = sub(charset, nxt_pos, nxt_pos) end -- Range. if nxt == "-" then if output == nil then output = {} end n = n + 1 output[n] = sub(charset, start, pos - 1) nxt_pos = nxt_pos + 1 nxt = sub(charset, nxt_pos, nxt_pos) -- Ranges fail if they end with a percent escape, so escape the hyphen to avoid undefined behaviour. if nxt == "" or nxt == "%" then n = n + 1 output[n] = (ch == "]" and "%]" or ch) .. "%-" start = nxt_pos nxt_pos = nxt_pos + 2 -- Since ranges can't contain "%]", since it's escaped, range inputs like "]-z" or "a-]" must be -- adjusted to the character before or after, plus "%]" (e.g. "%]^-z" or "a-\\%]"). The escaped "%]" is -- omitted if the range would be empty (i.e. if the first byte is greater than the second). else n = n + 1 output[n] = (ch == "]" and (byte(nxt) >= 0x5D and "%]^" or "^") or ch) .. "-" .. (nxt == "]" and (byte(ch) <= 0x5D and "\\%]" or "\\") or nxt) nxt_pos = nxt_pos + 1 start = nxt_pos end elseif ch == "-" or ch == "]" then if output == nil then output = {} end n = n + 1 output[n] = sub(charset, start, pos - 1) .. "%" .. ch start = nxt_pos end pos = nxt_pos end end if start == 1 then return "[" .. charset .. "]" end return "[" .. concat(output) .. sub(charset, start) .. "]" end get_charset = memoize(get_charset, true) export.get_charset = get_charset function export.len(str) return type(str) == "number" and len(str) or #str - #gsub(str, "[^\128-\191]+", "") end ulen = export.len function export.sub(str, i, j) str, i = type(str) == "number" and tostring(str) or str, i or 1 if i < 0 or j and j < 0 then return usub(str, i, j) elseif j and i > j or i > #str then return "" end local n, new_i = 0 for loc1, loc2 in gmatch(str, "()[^\128-\191]+()[\128-\191]*") do n = n + loc2 - loc1 if not new_i and n >= i then new_i = loc2 - (n - i) - 1 if not j then return sub(str, new_i) end end if j and n > j then return sub(str, new_i, loc2 - (n - j) - 1) end end return new_i and sub(str, new_i) or "" end do local function _find(str, loc1, loc2, ...) if loc1 and not match(str, "^()[^\128-\255]*$") then -- Use raw values of loc1 and loc2 to get loc1 and the length of the match. loc1, loc2 = ulen(sub(str, 1, loc1)), ulen(sub(str, loc1, loc2)) -- Offset length with loc1 to get loc2. loc2 = loc1 + loc2 - 1 end return loc1, loc2, ... end --[==[A version of find which uses string.find when possible, but otherwise uses mw.ustring.find.]==] function export.find(str, pattern, init, plain) init = init or 1 if init ~= 1 and not match(str, "^()[^\128-\255]*$") then return ufind(str, pattern, init, plain) elseif plain then return _find(str, find(str, pattern, init, true)) end local simple = pattern_simplifier(pattern) if simple then return _find(str, find(str, simple, init)) end return ufind(str, pattern, init) end end --[==[A version of match which uses string.match when possible, but otherwise uses mw.ustring.match.]==] function export.match(str, pattern, init) init = init or 1 if init ~= 1 and not match(str, "^()[^\128-\255]*$") then return umatch(str, pattern, init) end local simple = pattern_simplifier(pattern) if simple then return match(str, simple, init) end return umatch(str, pattern, init) end --[==[A version of gmatch which uses string.gmatch when possible, but otherwise uses mw.ustring.gmatch.]==] function export.gmatch(str, pattern) local simple = pattern_simplifier(pattern) if simple then return gmatch(str, simple) end return ugmatch(str, pattern) end --[==[A version of gsub which uses string.gsub when possible, but otherwise uses mw.ustring.gsub.]==] function export.gsub(str, pattern, repl, n) local simple = pattern_simplifier(pattern) if simple then return gsub(str, simple, repl, n) end return ugsub(str, pattern, repl, n) end --[==[ Like gsub, but pattern-matching facilities are turned off, so `pattern` and `repl` (if a string) are treated as literal. ]==] function export.plain_gsub(str, pattern, repl, n) return gsub(str, pattern_escape(pattern), type(repl) == "string" and replacement_escape(repl) or repl, n) end --[==[ Reverses a UTF-8 string; equivalent to string.reverse. ]==] function export.reverse(str) return reverse((gsub(str, "[\192-\255][\128-\191]*", reverse))) end function export.char(...) -- To be moved to [[Module:string/char]]. return u(...) end do local function utf8_err(func_name) error(format("bad argument #1 to '%s' (string is not UTF-8)", func_name), 4) end local function get_codepoint(func_name, b1, b2, b3, b4) if b1 <= 0x7F then return b1, 1 elseif not (b2 and b2 >= 0x80 and b2 <= 0xBF) then utf8_err(func_name) elseif b1 <= 0xDF then local cp = 0x40 * b1 + b2 - 0x3080 return cp >= 0x80 and cp or utf8_err(func_name), 2 elseif not (b3 and b3 >= 0x80 and b3 <= 0xBF) then utf8_err(func_name) elseif b1 <= 0xEF then local cp = 0x1000 * b1 + 0x40 * b2 + b3 - 0xE2080 return cp >= 0x800 and cp or utf8_err(func_name), 3 elseif not (b4 and b4 >= 0x80 and b4 <= 0xBF) then utf8_err(func_name) end local cp = 0x40000 * b1 + 0x1000 * b2 + 0x40 * b3 + b4 - 0x3C82080 return cp >= 0x10000 and cp <= 0x10FFFF and cp or utf8_err(func_name), 4 end function export.codepoint(str, i, j) if str == "" then return -- return nothing elseif type(str) == "number" then return byte(str, i, j) end i, j = i or 1, j == -1 and #str or i or 1 if i == 1 and j == 1 then return (get_codepoint("codepoint", byte(str, 1, 4))) elseif i < 0 or j < 0 then return ucodepoint(str, i, j) -- FIXME end local n, nb, ret, nr = 0, 1, {}, 0 while n < j do n = n + 1 if n < i then local b = byte(str, nb) nb = nb + (b < 128 and 1 or b < 224 and 2 or b < 240 and 3 or 4) else local b1, b2, b3, b4 = byte(str, nb, nb + 3) if not b1 then break end nr = nr + 1 local add ret[nr], add = get_codepoint("codepoint", b1, b2, b3, b4) nb = nb + add end end return unpack(ret) end codepoint = export.codepoint function export.gcodepoint(str, i, j) i, j = i or 1, j ~= -1 and j or nil if i < 0 or j and j < 0 then return ugcodepoint(str, i, j) -- FIXME end local n, nb = 1, 1 while n < i do local b = byte(str, nb) if not b then break end nb = nb + (b < 128 and 1 or b < 224 and 2 or b < 240 and 3 or 4) n = n + 1 end return function() if j and n > j then return nil end n = n + 1 local b1, b2, b3, b4 = byte(str, nb, nb + 3) if not b1 then return nil end local ret, add = get_codepoint("gcodepoint", b1, b2, b3, b4) nb = nb + add return ret end end end do local _ulower = ulower --[==[A version of lower which uses string.lower when possible, but otherwise uses mw.ustring.lower.]==] function export.lower(str) return (match(str, "^()[^\128-\255]*$") and lower or _ulower)(str) end end do local _uupper = uupper --[==[A version of upper which uses string.upper when possible, but otherwise uses mw.ustring.upper.]==] function export.upper(str) return (match(str, "^()[^\128-\255]*$") and upper or _uupper)(str) end end do local function add_captures(t, n, ...) if ... == nil then return end -- Insert any captures from the splitting pattern. local offset, capture = n - 1, ... while capture do n = n + 1 t[n] = capture capture = select(n - offset, ...) end return n end --[==[ Reimplementation of mw.text.split() that includes any capturing groups in the splitting pattern. This works like Python's re.split() function, except that it has Lua's behavior when the split pattern is empty (i.e. advancing by one character at a time; Python returns the whole remainder of the string). When possible, it will use the string library, but otherwise uses the ustring library. There are two optional parameters: `str_lib` forces use of the string library, while `plain` turns any pattern matching facilities off, treating `pattern` as literal. In addition, `pattern` may be a custom find function (or callable table), which takes the input string and start index as its two arguments, and must return the start and end index of the match, plus any optional captures, or nil if there are no further matches. By default, the start index will be calculated using the ustring library, unless `str_lib` or `plain` is set. ]==] function export.split(str, pattern_or_func, str_lib, plain) local iter, t, n = gsplit(str, pattern_or_func, str_lib, plain), {}, 0 repeat n = add_captures(t, n, iter()) until n == nil return t end export.capturing_split = export.split -- To be removed. end --[==[ Returns an iterator function, which iterates over the substrings returned by {split}. The first value returned is the string up the splitting pattern, with any capture groups being returned as additional values on that iteration. ]==] function export.gsplit(str, pattern_or_func, str_lib, plain) local start, final, str_len, _string, callable = 1 pattern_or_func, str_len, _string, callable = prepare_iter(str, pattern_or_func, str_lib, plain) local _find, _sub = _string.find, _string.sub local function iter(loc1, loc2, ...) -- If no match, or there is but we're past the end of the string -- (which happens when the match is the empty string), then return -- the final chunk. if not loc1 then final = true return _sub(str, start) end -- Special case: If we match the empty string, then eat the -- next character; this avoids an infinite loop, and makes -- splitting by the empty string work the way mw.text.gsplit() does -- (including non-adjacent empty string matches with %f). If we -- reach the end of the string this way, set `final` to true, so we -- don't get stuck matching the empty string at the end. local chunk if loc2 < loc1 then -- If using the string library, we need to make sure we advance -- by one UTF-8 character. if _sub == sub then local b = byte(str, loc1) if b and b >= 128 then loc1 = loc1 + (b < 224 and 1 or b < 240 and 2 or 3) end end chunk = _sub(str, start, loc1) if loc1 >= str_len then final = true else start = loc1 + 1 end -- Eat chunk up to the current match. else chunk = _sub(str, start, loc1 - 1) start = loc2 + 1 end return chunk, ... end if callable then return function() if not final then return iter(pattern_or_func(str, start)) end end -- Special case if the pattern is anchored to the start: "^" always -- anchors to the start position, not the start of the string, so get -- around this by only attempting one match with the pattern, then match -- the end of the string. elseif byte(pattern_or_func) == 0x5E then -- ^ local returned return function() if not returned then returned = true return iter(_find(str, pattern_or_func, start, plain)) elseif not final then return iter(_find(str, "$", start, plain)) end end end return function() if not final then return iter(_find(str, pattern_or_func, start, plain)) end end end gsplit = export.gsplit function export.count(str, pattern, plain) if plain then return select(2, gsub(str, pattern_escape(pattern), "")) end local simple = pattern_simplifier(pattern) if simple then return select(2, gsub(str, pattern, "")) end return select(2, ugsub(str, pattern, "")) end function export.trim(str, charset, str_lib, plain) if charset == nil then -- "^.*%S" is the fastest trim algorithm except when strings only consist of characters to be trimmed, which are -- very slow due to catastrophic backtracking. gsub with "^%s*" gets around this by trimming such strings to "" -- first. return match(gsub(str, "^%s*", ""), "^.*%S") or "" elseif charset == "" then return str end charset = plain and ("[" .. charset_escape(charset) .. "]") or get_charset(charset) -- The pattern uses a non-greedy quantifier instead of the algorithm used for %s, because negative character sets -- are non-trivial to compute (e.g. "[^^-z]" becomes "[%^_-z]"). Plus, if the ustring library has to be used, there -- would be two callbacks into PHP, which is slower. local pattern = "^" .. charset .. "*(.-)" .. charset .. "*$" if not str_lib then local simple = pattern_simplifier(pattern) if not simple then return umatch(str, pattern) end pattern = simple end return match(str, pattern) end trim = export.trim do local entities local function get_entities() entities, get_entities = load_data("Module:data/entities"), nil return entities end local function decode_entity(hash, x, code) if hash == "" then return (entities or get_entities())[x .. code] end local cp if x == "" then cp = match(code, "^()%d+$") and tonumber(code) else cp = match(code, "^()%x+$") and tonumber(code, 16) end return cp and (cp <= 0xD7FF or cp >= 0xE000 and cp <= 0x10FFFF) and u(cp) or nil end -- Non-ASCII characters aren't valid in proper HTML named entities, but MediaWiki uses them in some custom aliases -- which have also been included in [[Module:data/entities]]. function export.decode_entities(str) local amp = find(str, "&", nil, true) return amp and find(str, ";", amp, true) and gsub(str, "&(#?)([xX]?)([%w\128-\255]+);", decode_entity) or str end end do local entities local function get_entities() -- Memoized HTML entities (taken from mw.text.lua). entities, get_entities = { ["\""] = "&quot;", ["&"] = "&amp;", ["'"] = "&#039;", ["<"] = "&lt;", [">"] = "&gt;", ["\194\160"] = "&nbsp;", }, nil return entities end local function encode_entity(ch) local entity = (entities or get_entities())[ch] if entity == nil then local cp = codepoint(ch) -- U+D800 to U+DFFF are surrogates, so can't be encoded as entities. entity = cp and (cp <= 0xD7FF or cp >= 0xE000) and format("&#%d;", cp) or false entities[ch] = entity end return entity or nil end function export.encode_entities(str, charset, str_lib, plain) if charset == nil then return (gsub(str, "[\"&'<>\194]\160?", entities or get_entities())) elseif charset == "" then return str end local pattern = plain and ("[" .. charset_escape(charset) .. "]") or charset == "." and charset or get_charset(charset) if not str_lib then local simple = pattern_simplifier(pattern) if not simple then return (ugsub(str, pattern, encode_entity)) end pattern = simple end return (gsub(str, pattern, encode_entity)) end end do local function decode_path(code) return char(tonumber(code, 16)) end local function decode(lead, trail) if lead == "+" or lead == "_" then return " " .. trail elseif #trail == 2 then return decode_path(trail) end return lead .. trail end function export.decode_uri(str, enctype) enctype = enctype and upper(enctype) or "QUERY" if enctype == "PATH" then return find(str, "%", nil, true) and gsub(str, "%%(%x%x)", decode_path) or str elseif enctype == "QUERY" then return (find(str, "%", nil, true) or find(str, "+", nil, true)) and gsub(str, "([%%%+])(%x?%x?)", decode) or str elseif enctype == "WIKI" then return (find(str, "%", nil, true) or find(str, "_", nil, true)) and gsub(str, "([%%_])(%x?%x?)", decode) or str end error("bad argument #2 to 'decode_uri' (expected QUERY, PATH, or WIKI)", 2) end end do local function _remove_comments(str, pre) local head = find(str, "<!--", nil, true) if not head then return str end local ret, n = {sub(str, 1, head - 1)}, 1 while true do local loc = find(str, "-->", head + 4, true) if not loc then return pre and concat(ret) or concat(ret) .. sub(str, head) end head = loc + 3 loc = find(str, "<!--", head, true) if not loc then return concat(ret) .. sub(str, head) end n = n + 1 ret[n] = sub(str, head, loc - 1) head = loc end end --[==[ Removes any HTML comments from the input text. `stage` can be one of three options: * {"PRE"} (default) applies the method used by MediaWiki's preprocessor: all {{code|html|<nowiki><!-- ... --></nowiki>}} pairs are removed, as well as any text after an unclosed {{code|html|<nowiki><!--</nowiki>}}. This is generally suitable when parsing raw template or [[mw:Parser extension tags|parser extension tag]] code. (Note, however, that the actual method used by the preprocessor is considerably more complex and differs under certain conditions (e.g. comments inside nowiki tags); if full accuracy is absolutely necessary, use [[Module:template parser]] instead). * {"POST"} applies the method used to generate the final page output once all templates have been expanded: it loops over the text, removing any {{code|html|<nowiki><!-- ... --></nowiki>}} pairs until no more are found (e.g. {{code|html|<nowiki><!-<!-- ... -->- ... --></nowiki>}} would be fully removed), but any unclosed {{code|html|<nowiki><!--</nowiki>}} is ignored. This is suitable for handling links embedded in template inputs, where the {"PRE"} method will have already been applied by the native parser. * {"BOTH"} applies {"PRE"} then {"POST"}. ]==] function export.remove_comments(str, stage) if not stage or stage == "PRE" then return _remove_comments(str, true) end local processed = stage == "POST" and _remove_comments(str) or stage == "BOTH" and _remove_comments(str, true) or error("bad argument #2 to 'remove_comments' (expected PRE, POST, or BOTH)", 2) while processed ~= str do str = processed processed = _remove_comments(str) end return str end end do local byte_escapes local function get_byte_escapes() byte_escapes, get_byte_escapes = load_data("Module:string utilities/data").byte_escapes, nil return byte_escapes end local function escape_byte(b) return (byte_escapes or get_byte_escapes())[b] or format("\\%03d", byte(b)) end function export.escape_bytes(str) return (gsub(str, ".", escape_byte)) end end function export.format_fun(str, fun) str = str:gsub("%-%{", "\5"):gsub("%}%-", "\6") -- this str is for l10n! str = (gsub(str, "{(\\?)((\\?)[^{}]*)}", function(p1, name, p2) if #p1 + #p2 == 1 then return name == "op" and "{" or name == "cl" and "}" or error(mw.getCurrentFrame():getTitle() .. " format: unrecognized escape sequence '{\\" .. name .. "}'") elseif fun(name) and type(fun(name)) ~= "string" then error(mw.getCurrentFrame():getTitle() .. " format: \"" .. name .. "\" is a " .. type(fun(name)) .. ", not a string") end return fun(name) or error(mw.getCurrentFrame():getTitle() .. " format: \"" .. name .. "\" not found in table") end)) return (str:gsub("\5", "-{"):gsub("\6", "}-")) end format_fun = export.format_fun --[==[ This function, unlike {string.format} and {mw.ustring.format}, takes just two parameters, a format string and a table, and replaces all instances of { {param_name} } in the format string with the table's entry for {param_name}. The opening and closing brace characters can be escaped with { {\op} } and { {\cl} }, respectively. A table entry beginning with a slash can be escaped by doubling the initial slash. ====Examples==== * {string_utilities.format("{foo} fish, {bar} fish, {baz} fish, {quux} fish", {["foo"]="one", ["bar"]="two", ["baz"]="red", ["quux"]="blue"}) } *: produces: {"one fish, two fish, red fish, blue fish"} * {string_utilities.format("The set {\\op}1, 2, 3{\\cl} contains {\\\\hello} elements.", {["\\hello"]="three"})} *: produces: {"The set {1, 2, 3} contains three elements."} *:* Note that the single and double backslashes should be entered as double and quadruple backslashes when quoted in a literal string. ]==] function export.format(str, tbl) return format_fun(str, function(key) return tbl[key] end) end do local function do_uclcfirst(str, case_func) -- Re-case the first letter. local first, remainder = match(str, "^(.[\128-\191]*)(.*)") return first and (case_func(first) .. remainder) or "" end local function uclcfirst(str, case_func) -- Strip off any HTML tags at the beginning. This currently does not handle comments or <ref>...</ref> -- correctly; it's intended for text wrapped in <span> or the like, as happens when passing text through -- [[Module:links]]. local html_at_beginning = nil if str:match("^<") then while true do local html_tag, rest = str:match("^(<.->)(.*)$") if not html_tag then break end if not html_at_beginning then html_at_beginning = {} end insert(html_at_beginning, html_tag) str = rest end end -- If there's a link at the beginning, re-case the first letter of the -- link text. This pattern matches both piped and unpiped links. -- If the link is not piped, the second capture (linktext) will be empty. local link, linktext, remainder = match(str, "^%[%[([^|%]]+)%|?(.-)%]%](.*)$") local retval if link then retval = "[[" .. link .. "|" .. do_uclcfirst(linktext ~= "" and linktext or link, case_func) .. "]]" .. remainder else retval = do_uclcfirst(str, case_func) end if html_at_beginning then retval = concat(html_at_beginning) .. retval end return retval end --[==[ Uppercase the first character of the input string, correctly handling one-part and two-part links, optionally surrounded by HTML tags such as `<nowiki><span>...</span></nowiki>`, possibly nested. Intended to correctly uppercase the first character of text that may include links that have been passed through `full_link()` in [[Module:links]] or a similar function. ]==] function export.ucfirst(str) return uclcfirst(str, uupper) end ucfirst = export.ucfirst --[==[ Lowercase the first character of the input string, correctly handling one-part and two-part links, optionally surrounded by HTML tags such as `<nowiki><span>...</span></nowiki>`, possibly nested. Intended to correctly lowercase the first character of text that may include links that have been passed through `full_link()` in [[Module:links]] or a similar function. ]==] function export.lcfirst(str) return uclcfirst(str, ulower) end --[==[Capitalizes each word of the input string. WARNING: May be broken in the presence of multiword links.]==] function export.capitalize(str) -- Capitalize multi-word that is separated by spaces -- by uppercasing the first letter of each part. return (ugsub(str, "%w+", ucfirst)) end local function do_title_case(first, remainder) first = uupper(first) return remainder == "" and first or (first .. ulower(remainder)) end --[==[ Capitalizes each word of the input string, with any further letters in each word being converted to lowercase. ]==] function export.title_case(str) return str == "" and "" or ugsub(str, "(%w)(%w*)", do_title_case) end title_case = export.title_case --[==[ Converts the input string to {{w|Camel case|CamelCase}}. Any non-word characters are treated as breaks between words. If `lower_first` is set, then the first character of the string will be lowercase (e.g. camelCase). ]==] function export.camel_case(str, lower_first) str = ugsub(str, "%W*(%w*)", title_case) return lower_first and do_uclcfirst(str, ulower) or str end end do local function do_snake_case(nonword, word) return nonword == "" and word or "_" .. word end --[==[ Converts the input string to {{w|Snake case|snake_case}}. Any non-word characters are treated as breaks between words. ]==] function export.snake_case(str) return (ugsub(str, "(%W*)(%w*)", do_snake_case)) end end return export 5539a01p7ihwjc0hufg4p7q6ayyna31 828 34333 184521 2026-05-21T06:08:46Z P1ayer 1197 创建页面，内容为“local math_module = "Module:math" local table_pack_module = "Module:table/pack" local require = require local select = select local unpack = unpack or table.unpack -- Lua 5.2 compatibility -- table.pack: in Lua 5.2+, this is a function that wraps the parameters given -- into a table with the additional key `n` that contains the total number of -- parameters given. This is not available on Lua 5.1, so [[Module:table/pack]] -- provides the same functionality. l…” 184521 Scribunto text/plain local math_module = "Module:math" local table_pack_module = "Module:table/pack" local require = require local select = select local unpack = unpack or table.unpack -- Lua 5.2 compatibility -- table.pack: in Lua 5.2+, this is a function that wraps the parameters given -- into a table with the additional key `n` that contains the total number of -- parameters given. This is not available on Lua 5.1, so [[Module:table/pack]] -- provides the same functionality. local function pack(...) pack = require(table_pack_module) return pack(...) end local function sign(...) sign = require(math_module).sign return sign(...) end ----- M E M O I Z A T I O N----- -- Memoizes a function or callable table. -- Supports any number of arguments and return values. -- If the optional parameter `simple` is set, then the memoizer will use a faster implementation, but this is only compatible with one argument and one return value. If `simple` is set, additional arguments will be accepted, but this should only be done if those arguments will always be the same. -- Sentinels. local _nil, neg_0, pos_nan, neg_nan = {}, {}, {}, {} -- Certain values can't be used as table keys, so they require sentinels as well: e.g. f("foo", nil, "bar") would be memoized at memo["foo"][_nil]["bar"][memo]. These values are: -- nil. -- -0, which is equivalent to 0 in most situations, but becomes "-0" on conversion to string; it also behaves differently in some operations (e.g. 1/a evaluates to inf if a is 0, but -inf if a is -0). -- NaN and -NaN, which are the only values for which n == n is false; they only seem to differ on conversion to string ("nan" and "-nan"). local function get_key(x) if x == x then return x == nil and _nil or x == 0 and 1 / x < 0 and neg_0 or x end return sign(x) == 1 and pos_nan or neg_nan end -- Return values are memoized as tables of return values, which are looked up using each input argument as a key, followed by `memo`. e.g. if the input arguments were (1, 2, 3), the memo would be located at t[1][2][3][memo]. `memo` is always used as the final lookup key so that (for example) the memo for f(1, 2, 3), f[1][2][3][memo], doesn't interfere with the memo for f(1, 2), f[1][2][memo]. local function get_memo(memo, n, nargs, key, ...) key = get_key(key) local next_memo = memo[key] if next_memo == nil then next_memo = {} memo[key] = next_memo end memo = next_memo return n == nargs and memo or get_memo(memo, n + 1, nargs, ...) end -- Used to catch the function output values instead of using a table directly, -- since pack() returns a table with the key `n`, giving the number of return -- values, even if they are nil. This ensures that any nil return values after -- the last non-nil value will always be present (e.g. pack() gives {n = 0}, -- pack(nil) gives {n = 1}, pack(nil, "foo", nil) gives {[2] = "foo", n = 3} -- etc.). The distinction between nil and nothing affects some native functions -- (e.g. tostring() throws an error, but tostring(nil) returns "nil"), so it -- needs to be reconstructable from the memo. local function memoize_then_return(memo, _memo, ...) _memo[memo] = pack(...) return ... end return function(func, simple) local memo = {} if simple then return function(...) local key = get_key((...)) local output = memo[key] if output == nil then output = func(...) memo[key] = output == nil and _nil or output return output elseif output == _nil then return nil end return output end end return function(...) local nargs = select("#", ...) -- Since all possible inputs need to be memoized (including true, false -- and nil), the memo table itself is used as a sentinel to ensure that -- the table of arguments will always have a unique key. local _memo = nargs == 0 and memo or get_memo(memo, 1, nargs, ...) local output = _memo[memo] -- If get_memo() returned nil, call `func` with the arguments and catch -- the output with memoize_then_return(); this packs the return values -- into a table to memoize them, then returns them. Since the return -- values are available to it as `...`, this avoids the need to call -- unpack() on the memoized table on the first call, as they can be -- returned directly. if output == nil then return memoize_then_return(memo, _memo, func(...)) end -- Unpack from 1 to the original number of return values (memoized at -- key `n`); unpack() returns nil for any values not in output. return unpack(output, 1, output.n) end end 8gmyvbgmtmwh2pu7neciscs7liw1v30 828 34334 184522 2026-05-21T06:12:50Z P1ayer 1197 创建页面，内容为“--[==[TODO: * Change certain flag names, as some are misnomers: * Change `allow_holes` to `keep_holes`, because it's not the inverse of `disallow_holes`. * Change `allow_empty` to `keep_empty`, as it causes them to be kept as "" instead of deleted. * Sort out all the internal error calls. Manual error(format()) calls are used when certain parameters shouldn't be dumped, so find a way to avoid that. ]==] local export = {} local collation_module = "Module:col…” 184522 Scribunto text/plain --[==[TODO: * Change certain flag names, as some are misnomers: * Change `allow_holes` to `keep_holes`, because it's not the inverse of `disallow_holes`. * Change `allow_empty` to `keep_empty`, as it causes them to be kept as "" instead of deleted. * Sort out all the internal error calls. Manual error(format()) calls are used when certain parameters shouldn't be dumped, so find a way to avoid that. ]==] local export = {} local collation_module = "Module:collation" local families_module = "Module:families" local functions_module = "Module:fun" local gender_and_number_utilities_module = "Module:gender and number utilities" local labels_module = "Module:labels" local languages_module = "Module:languages" local math_module = "Module:math" local pages_module = "Module:pages" local parameters_finalize_set_module = "Module:parameters/finalizeSet" local parameters_track_module = "Module:parameters/track" local parse_utilities_module = "Module:parse utilities" local references_module = "Module:references" local scribunto_module = "Module:Scribunto" local scripts_module = "Module:scripts" local string_utilities_module = "Module:string utilities" local table_module = "Module:table" local wikimedia_languages_module = "Module:wikimedia languages" local yesno_module = "Module:yesno" local mw = mw local mw_title = mw.title local string = string local table = table local dump = mw.dumpObject local find = string.find local format = string.format local gsub = string.gsub local insert = table.insert local ipairs = ipairs local list_to_text = mw.text.listToText local make_title = mw_title.makeTitle local match = string.match local max = math.max local new_title = mw_title.new local next = next local pairs = pairs local pcall = pcall local require = require local sub = string.sub local tonumber = tonumber local type = type local unpack = unpack or table.unpack -- Lua 5.2 compatibility local current_title_text, current_namespace, sets -- Defined when needed. local namespaces = mw.site.namespaces --[==[ Loaders for functions in other modules, which overwrite themselves with the target function when called. This ensures modules are only loaded when needed, retains the speed/convenience of locally-declared pre-loaded functions, and has no overhead after the first call, since the target functions are called directly in any subsequent calls.]==] local function decode_entities(...) decode_entities = require(string_utilities_module).decode_entities return decode_entities(...) end local function extend(...) extend = require(table_module).extend return extend(...) end local function finalize_set(...) finalize_set = require(parameters_finalize_set_module) return finalize_set(...) end local function get_family_by_code(...) get_family_by_code = require(families_module).getByCode return get_family_by_code(...) end local function get_family_by_name(...) get_family_by_name = require(families_module).getByCanonicalName return get_family_by_name(...) end local function get_language_by_code(...) get_language_by_code = require(languages_module).getByCode return get_language_by_code(...) end local function get_language_by_name(...) get_language_by_name = require(languages_module).getByCanonicalName return get_language_by_name(...) end local function get_script_by_code(...) get_script_by_code = require(scripts_module).getByCode return get_script_by_code(...) end local function get_script_by_name(...) get_script_by_name = require(scripts_module).getByCanonicalName return get_script_by_name(...) end local function get_wm_lang_by_code(...) get_wm_lang_by_code = require(wikimedia_languages_module).getByCode return get_wm_lang_by_code(...) end local function get_wm_lang_by_code_with_fallback(...) get_wm_lang_by_code_with_fallback = require(wikimedia_languages_module).getByCodeWithFallback return get_wm_lang_by_code_with_fallback(...) end local function gsplit(...) gsplit = require(string_utilities_module).gsplit return gsplit(...) end local function is_callable(...) is_callable = require(functions_module).is_callable return is_callable(...) end local function is_integer(...) is_integer = require(math_module).is_integer return is_integer(...) end local function is_internal_title(...) is_internal_title = require(pages_module).is_internal_title return is_internal_title(...) end local function is_positive_integer(...) is_positive_integer = require(math_module).is_positive_integer return is_positive_integer(...) end local function iterate_list(...) iterate_list = require(table_module).iterateList return iterate_list(...) end local function num_keys(...) num_keys = require(table_module).numKeys return num_keys(...) end local function parse_gender_and_number_spec(...) parse_gender_and_number_spec = require(gender_and_number_utilities_module).parse_gender_and_number_spec return parse_gender_and_number_spec(...) end local function parse_references(...) parse_references = require(references_module).parse_references return parse_references(...) end local function pattern_escape(...) pattern_escape = require(string_utilities_module).pattern_escape return pattern_escape(...) end local function php_trim(...) php_trim = require(scribunto_module).php_trim return php_trim(...) end local function scribunto_parameter_key(...) scribunto_parameter_key = require(scribunto_module).scribunto_parameter_key return scribunto_parameter_key(...) end local function sort(...) sort = require(collation_module).sort return sort(...) end local function sorted_pairs(...) sorted_pairs = require(table_module).sortedPairs return sorted_pairs(...) end local function split(...) split = require(string_utilities_module).split return split(...) end local function split_labels_on_comma(...) split_labels_on_comma = require(labels_module).split_labels_on_comma return split_labels_on_comma(...) end local function split_on_comma(...) split_on_comma = require(parse_utilities_module).split_on_comma return split_on_comma(...) end local function tonumber_extended(...) tonumber_extended = require(math_module).tonumber_extended return tonumber_extended(...) end local function track(...) track = require(parameters_track_module) return track(...) end local function yesno(...) yesno = require(yesno_module) return yesno(...) end --[==[ intro: This module is used to standardize template argument processing and checking. A typical workflow is as follows (based on [[Module:translations]]): { ... local parent_args = frame:getParent().args local params = { [1] = {required = true, type = "language", default = "und"}, [2] = true, [3] = {list = true}, ["alt"] = true, ["id"] = true, ["sc"] = {type = "script"}, ["tr"] = true, ["ts"] = true, ["lit"] = true, } local args = require("Module:parameters").process(parent_args, params) -- Do further processing of the parsed arguments in `args`. ... } The `params` table should have the parameter names as the keys, and a (possibly empty) table of parameter tags as the value. An empty table as the value merely states that the parameter exists, but should not receive any special treatment; if desired, empty tables can be replaced with the value `true` as a perforamnce optimization. Possible parameter tags are listed below: ; {required = true} : The parameter is required; an error is shown if it is not present. The template's page itself is an exception; no error is shown there. ; {default =} : Specifies a default input value for the parameter, if it is absent or empty. This will be processed as though it were the input instead, so (for example) {default = "und"} with the type {"language"} will return a language object for [[:Category:Undetermined language|Undetermined language]] if no language code is provided. When used on list parameters, this specifies a default value for the first item in the list only. Note that it is not possible to generate a default that depends on the value of other parameters. If used together with {required = true}, the default applies only to template pages (see the following entry), as a side effect of the fact that "required" parameters aren't actually required on template pages. This can be used to show an example of the template in action when the template page is visited; however, it is preferred to use `template_default` for this purpose, for clarity. ; {template_default =} : Specifies a default input value for absent or empty parameters only on the template demo invocation (the invocation of the template that is displayed when the template page that implements the template is viewed). Template pages are pages in template space that invoke (through {{tl|#invoke:}}) the module that implements the template and calls [[Module:parameters]]. For example, the page [[Template:en-noun]] implements the {{tl|en-noun}} template, which in turn invokes [[Module:en-headword]], and is a template page for [[Module:en-headword]]. When the template page [[Template:en-noun]] is visited, the {{tl|#invoke:}} of the template's module is expanded as if the template were called without arguments, and the output is inserted at that point into the processed page. This output serves as a sort of demo of the template's functionality. `template_default` can be used to supply default values for use only in this demo. Since the template page may also contain other invocations of the same template (e.g. on the template's documentation page, which is typically transcluded into the template page itself), `template_default` does not apply if there are any arguments passed to the template or if the template is invoked on any other page but its own template page (which is checked by comparing the name of the invoking template to the current pagename). Both `template_default` and `default` can be specified for the same parameter. If this is done, `template_default` applies for the argumentless template invocation on the template page, and `default` in all other circumstances As an example, {{tl|cs-IPA}} uses the equivalent of {[1] = {default = "+", template_default = "příklad"}} to supply a default of {"+"} for mainspace and documentation pages (which tells the module to use the value of the {{para|pagename}} parameter, falling back to the actual pagename), but {"příklad"} (which means "example"), on [[Template:cs-IPA]]. ; {alias_of =} : Treat the parameter as an alias of another. When arguments are specified for this parameter, they will automatically be renamed and stored under the alias name. This allows for parameters with multiple alternative names, while still treating them as if they had only one name. The conversion-related properties of an aliased parameter (e.g. `type`, `set`, `convert`, `sublist`) are taken from the aliasee, and the corrresponding properties set on the alias itself are ignored; but other properties on the alias are taken from the alias's spec and not from the aliasee's spec. This means, for example, that if you create an alias of a list parameter, the alias must also specify the `list` property or it is not a list. (In such a case, a value specified for the alias goes into the first item of the aliasee's list. You cannot make a list alias of a non-list parameter; this causes an error to be thrown.) Similarly, if you specify `separate_no_index` on an aliasee but not on the alias, uses of the unindexed aliasee parameter are stored into the `.default` key, but uses of the unindexed alias are stored into the first numbered key of the aliasee's list. Aliases cannot be required, as this prevents the other name or names of the parameter from being used. Parameters that are aliases and required at the same time cause an error to be thrown. ; {allow_empty = true} : If the argument is an empty string value, it is not converted to {nil}, but kept as-is. The use of `allow_empty` is disallowed if a type has been specified, and causes an error to be thrown. ; {no_trim = true} : Spacing characters such as spaces and newlines at the beginning and end of a positional parameter are not removed. (MediaWiki itself automatically trims spaces and newlines at the edge of named parameters.) The use of `no_trim` is disallowed if a type has been specified, and causes an error to be thrown. ; {type =} : Specifies what value type to convert the argument into. The default is to leave it as a text string. Alternatives are: :; {type = "boolean"} :: The value is treated as a boolean value, either true or false. No value, the empty string, and the strings {"0"}, {"no"}, {"n"}, {"false"}, {"f"} and {"off"} are treated as {false}, all other values are considered {true}. :; {type = "number"} :: The value is converted into a number, and throws an error if the value is not parsable as a number. Input values may be signed (`+` or `-`), and may contain decimal points and leading zeroes. If {allow_hex = true}, then hexadecimal values in the form {"0x100"} may optionally be used instead, which otherwise have the same syntax restrictions (including signs, decimal digits, and leading zeroes after {"0x"}). Hexadecimal inputs are not case-sensitive. Lua's special number values (`inf` and `nan`) are not possible inputs. :; {type = "range"} :: The value is interpreted as a hyphen-separated range of two numbers (e.g. {"2-4"} is interpreted as the range from {2} to {4}). A number input without a hyphen is interpreted as a range from that number to itself (e.g. the input {"1"} is interpreted as the range from {1} to {1}). Any optional flags which are available for numbers will also work for ranges. :; {type = "language"} :: The value is interpreted as a full or [[Wiktionary:Languages#Etymology-only languages|etymology-only language]] code language code (or name, if {method = "name"}) and converted into the corresponding object (see [[Module:languages]]). If the code or name is invalid, then an error is thrown. The additional setting {family = true} can be given to allow [[Wiktionary:Language families|language family codes]] to be considered valid and the corresponding object returned. Note that to distinguish an etymology-only language object from a full language object, use {object:hasType("language", "etymology-only")}. :; {type = "full language"} :: The value is interpreted as a full language code (or name, if {method = "name"}) and converted into the corresponding object (see [[Module:languages]]). If the code or name is invalid, then an error is thrown. Etymology-only languages are not allowed. The additional setting {family = true} can be given to allow [[Wiktionary:Language families|language family codes]] to be considered valid and the corresponding object returned. :; {type = "Wikimedia language"} :: The value is interpreted as a code and converted into a Wikimedia language object. If the code is invalid, then an error is thrown. If {fallback = true} is specified, conventional language codes which are different from their Wikimedia equivalent will also be accepted as a fallback. :; {type = "family"} :: The value is interpreted as a language family code (or name, if {method = "name"}) and converted into the corresponding object (see [[Module:families]]). If the code or name is invalid, then an error is thrown. :; {type = "script"} :: The value is interpreted as a script code (or name, if {method = "name"}) and converted into the corresponding object (see [[Module:scripts]]). If the code or name is invalid, then an error is thrown. :; {type = "title"} :: The value is interpreted as a page title and converted into the corresponding object (see the [[mw:Extension:Scribunto/Lua_reference_manual#Title_library|Title library]]). If the page title is invalid, then an error is thrown; by default, external titles (i.e. those on other wikis) are not treated as valid. Options are: ::; {namespace = n} ::: The default namespace, where {n} is a namespace number; this is treated as {0} (the mainspace) if not specified. ::; {allow_external = true} ::: External titles are treated as valid. ::; {prefix = "namespace override"} (default) ::: The default namespace prefix will be prefixed to the value is already prefixed by a namespace prefix. For instance, the input {"Foo"} with namespace {10} returns {"Template:Foo"}, {"Wiktionary:Foo"} returns {"Wiktionary:Foo"}, and {"Template:Foo"} returns {"Template:Foo"}. Interwiki prefixes cannot act as overrides, however: the input {"fr:Foo"} returns {"Template:fr:Foo"}. ::; {prefix = "force"} ::: The default namespace prefix will be prefixed unconditionally, even if the value already appears to be prefixed. This is the way that {{tl|#invoke:}} works when calling modules from the module namespace ({828}): the input {"Foo"} returns {"Module:Foo"}, {"Wiktionary:Foo"} returns {"Module:Wiktionary:Foo"}, and {"Module:Foo"} returns {"Module:Module:Foo"}. ::; {prefix = "full override"} ::: The same as {prefix = "namespace override"}, except that interwiki prefixes can also act as overrides. For instance, {"el:All topics"} with namespace {14} returns {"el:Category:All topics"}. Due to the limitations of MediaWiki, only the first prefix in the value may act as an override, so the namespace cannot be overridden if the first prefix is an interwiki prefix: e.g. {"el:Template:All topics"} with namespace {14} returns {"el:Category:Template:All topics"}. :; {type = "parameter"} :: The value is interpreted as the name of a parameter, and will be normalized using the method that Scribunto uses when constructing a {frame.args} table of arguments. This means that integers will be converted to numbers, but all other arguments will remain as strings (e.g. {"1"} will be normalized to {1}, but {"foo"} and {"1.5"} will remain unchanged). Note that Scribunto also trims parameter names, following the same trimming method that this module applies by default to all parameter types. :: This type is useful when one set of input arguments is used to construct a {params} table for use in a subsequent {export.process()} call with another set of input arguments; for instance, the set of valid parameters for a template might be defined as {{tl|#invoke:[some module]|args=}} in the template, where {args} is a sublist of valid parameters for the template. :; {type = "qualifier"} :: The value is interpreted as a qualifier and converted into the correct format for passing into `format_qualifiers()` in [[Module:qualifier]] (which currently just means converting it to a one-item list). :; {type = "labels"} :: The value is interpreted as a comma-separated list of labels and converted into the correct format for passing into `show_labels()` in [[Module:labels]] (which is currently a list of strings). Splitting is done on commas not followed by whitespace, except that commas inside of double angle brackets do not count even if not followed by whitespace. This type should be used by for normal labels (typically specified using {{para|l}} or {{para|ll}}) and accent qualifiers (typically specified using {{para|a}} and {{para|aa}}). :; {type = "references"} :: The value is interpreted as one or more references, in the format prescribed by `parse_references()` in [[Module:references]], and converted into a list of objects of the form accepted by `format_references()` in the same module. If a syntax error is found in the reference format, an error is thrown. :; {type = "genders"} :: The value is interpreted as one or more comma-separated gender/number specs, in the format prescribed by [[Module:gender and number]]. Inline modifiers (`<q:...>`, `<qq:...>`, `<l:...>`, `<ll:...>` or `<ref:...>`) may be attached to a gender/number spec. :; {type = "form of tags"} :: The value is interpreted as an ampersand-separated list of grammar tags and converted into the correct format for passing as `tags` into `tagged_inflections()` in [[Module:form of]] (which is currently a list of strings). Splitting is always done by ampersands. This type should be used by for inflection qualifiers that act as grammar tags (typically specified using {{para|infl}}). :; {type = function(val) ... end} :: `type` may be set to a function (or callable table), which must take the argument value as its sole argument, and must output one of the other recognized types. This is particularly useful for lists (see below), where certain values need to be interpreted differently to others. ; {list =} : Treat the parameter as a list of values, each having its own parameter name, rather than a single value. The parameters will have a number at the end, except optionally for the first (but see also {require_index = true}). For example, {list = true} on a parameter named "head" will include the parameters {{para|head}} (or {{para|head1}}), {{para|head2}}, {{para|head3}} and so on. If the parameter name is a number, another number doesn't get appended, but the counting simply continues, e.g. for parameter {3} the sequence is {{para|3}}, {{para|4}}, {{para|5}} etc. List parameters are returned as numbered lists, so for a template that is given the parameters `|head=a|head2=b|head3=c`, the processed value of the parameter {"head"} will be { { "a", "b", "c" }}}. : The value for {list =} can also be a string. This tells the module that parameters other than the first should have a different name, which is useful when the first parameter in a list is a number, but the remainder is named. An example would be for genders: {list = "g"} on a parameter named {1} would have parameters {{para|1}}, {{para|g2}}, {{para|g3}} etc. : If the number is not located at the end, it can be specified by putting {"\1"} at the number position. For example, parameters {{para|f1accel}}, {{para|f2accel}}, ... can be captured by using the parameter name {"f\1accel"}, as is done in [[Module:headword/templates]]. ; {set =} : Require that the value of the parameter be one of the specified values (or omitted, if {required = true} isn't given). Two formats are allowed; either a list of possible values can be supplied, or a table can be supplied where the keys are allowed values and the values are either `true` or a string naming a value found elsewhere in the table as a key. In the latter case, the key is an alias and the value is the canonical value, and if the user uses the alias, it will automatically be mapped to the canonical value. In such a case, the canonical value cannot itself be an alias. The use of `set` is disallowed if {type = "boolean"} and causes an error to be thrown. ; {sublist =} : The value of the parameter is a delimiter-separated list of individual raw values. The resulting field in `args` will be a Lua list (i.e. a table with numeric indices) of the converted values. If {sublist = true} is given, the values will be split on commas (possibly with whitespace on one or both sides of the comma, which is ignored). If {sublist = "comma without whitespace"} is given, the values will be split on commas which are not followed by whitespace, and which aren't preceded by an escaping backslash. Otherwise, the value of `sublist` should be either a Lua pattern specifying the delimiter(s) to split on or a function (or callable table) to do the splitting, which is passed two values (the value to split and a function to signal an error) and should return a list of the split values. ; {convert =} : If given, this specifies a function (or callable table) to convert the raw parameter value into the Lua object used during further processing. The function is passed two arguments, the raw parameter value itself and a function used to signal an error during parsing or conversion, and should return one value, the converted parameter. The error-signaling function contains the name and raw value of the parameter embedded into the message it generates, so these do not need to specified in the message passed into it. If `type` is specified in conjunction with `convert`, the processing by `type` happens first. If `sublist` is given in conjunction with `convert`, the raw parameter value will be split appropriately and `convert` called on each resulting item. ; {allow_hex = true} : When used in conjunction with {type = "number"}, allows hexadecimal numbers as inputs, in the format {"0x100"} (which is not case-sensitive). ; {family = true} : When used in conjunction with {type = "language"}, allows [[Wiktionary:Language families|language family codes]] to be returned. To check if a given object refers to a language family, use {object:hasType("family")}. ; {method = "name"} : When used in conjunction with {type = "language"}, {type = "family"} or {type = "script"}, checks for and parses a language, family or script name instead of a code. ; {allow_holes = true} : This is used in conjunction with list-type parameters. By default, the values are tightly packed in the resulting list. This means that if, for example, an entry specified `head=a|head3=c` but not {{para|head2}}, the returned list will be { {"a", "c"}}}, with the values stored at the indices {1} and {2}, not {1} and {3}. If it is desirable to keep the numbering intact, for example if the numbers of several list parameters correlate with each other (like those of {{tl|affix}}), then this tag should be specified. : If {allow_holes = true} is given, there may be {nil} values in between two real values, which makes many of Lua's table processing functions no longer work, like {#} or {ipairs()}. To remedy this, the resulting table will contain an additional named value, `maxindex`, which tells you the highest numeric index that is present in the table. In the example above, the resulting table will now be { { "a", nil, "c", maxindex = 3}}}. That way, you can iterate over the values from {1} to `maxindex`, while skipping {nil} values in between. ; {disallow_holes = true} : This is used in conjunction with list-type parameters. As mentioned above, normally if there is a hole in the source arguments, e.g. `head=a|head3=c` but not {{para|head2}}, it will be removed in the returned list. If {disallow_holes = true} is specified, however, an error is thrown in such a case. This should be used whenever there are multiple list-type parameters that need to line up (e.g. both {{para|head}} and {{para|tr}} are available and {{para|head3}} lines up with {{para|tr3}}), unless {allow_holes = true} is given and you are prepared to handle the holes in the returned lists. ; {disallow_missing = true} : This is similar to {disallow_holes = true}, but an error will not be thrown if an argument is blank, rather than completely missing. This may be used to tolerate intermediate blank numerical parameters, which sometimes occur in list templates. For instance, `head=a|head2=|head3=c` will not throw an error, but `head=a|head3=c` will. ; {require_index = true} : This is used in conjunction with list-type parameters. By default, the first parameter can have its index omitted. For example, a list parameter named `head` can have its first parameter specified as either {{para|head}} or {{para|head1}}. If {require_index = true} is specified, however, only {{para|head1}} is recognized, and {{para|head}} will be treated as an unknown parameter. {{tl|affixusex}} (and variants {{tl|suffixusex}}, {{tl|prefixusex}}) use this, for example, on all list parameters. ; {separate_no_index = true} : This is used to distinguish between {{para|head}} and {{para|head1}} as different parameters. For example, in {{tl|affixusex}}, to distinguish between {{para|sc}} (a script code for all elements in the usex's language) and {{para|sc1}} (the script code of the first element, used when the first element is prefixed with a language code to indicate that it is in a different language). When this is used, the resulting table will contain an additional named value, `default`, which contains the value for the indexless argument. ; {flatten = true} : This is used in conjunction with list-type parameters when `sublist` or a list-generating type such as {"labels"} or {"genders"} is also specified, and causes the resulting list to be flattened. Not currently compatible with {allow_holes = true}. ; {replaced_by =} : Specifies that the parameter is no longer valid, and has been replaced by some other mechanism. If the value of `replaced_by` is a string, it is the name of the new parameter to use instead. Use the `reason` tag to specify the reason why this change has been made, e.g. {reason = "for consistency with the corresponding parameter in other Romance-language headword templates"}. If the value of `replaced_by` is {false}, there is no replacement parameter. In this case, `instead` should be supplied with a description of what to do instead, e.g. {instead = "use an inline modifier on |2= such as <q:...>, <qq:...>, <l:...> or <ll:...>"}. You can also supply a justification in `reason` if you feel it is appropriate or necessary to do so. ; {reason =} : When used in conjunction with `replaced_by`, specifies the reason for the parameter replacement. ; {instead =} : When used in conjunction with {replaced_by = false}, specifies what to do instead of using the removed parameter. ; {demo = true} : This is used as a way to ensure that the parameter is only enabled on the template's own page (and its documentation page), and in the User: namespace; otherwise, it will be treated as an unknown parameter. This should only be used if special settings are required to showcase a template in its documentation (e.g. adjusting the pagename or disabling categorization). In most cases, it should be possible to do this without using demo parameters, but they may be required if a template/documentation page also contains real uses of the same template as well (e.g. {{tl|shortcut}}), as a way to distinguish them. ; {deprecated = true} : This is for tracking the use of deprecated parameters, including any aliases that are being brought out of use. See [[Wiktionary:Tracking]] for more information. ]==] -- Returns true if the current page is a template or module containing the current {{#invoke}}. -- If the include_documentation argument is given, also returns true if the current page is either page's documentation page. local own_page, own_page_or_documentation local function is_own_page(include_documentation) if own_page == nil then if current_namespace == nil then local current_title = mw_title.getCurrentTitle() current_title_text, current_namespace = current_title.prefixedText, current_title.namespace end local frame = current_namespace == 828 and mw.getCurrentFrame() or current_namespace == 10 and mw.getCurrentFrame():getParent() if frame then local frame_title_text = frame:getTitle() own_page = current_title_text == frame_title_text own_page_or_documentation = own_page or current_title_text == frame_title_text .. "/documentation" else own_page, own_page_or_documentation = false, false end end return include_documentation and own_page_or_documentation or own_page end -------------------------------------- Some helper functions ----------------------------- -- Convert a list in `list` to a string, separating the final element from the preceding one(s) by `conjunction`. If -- `dump_vals` is given, pass all values in `list` through mw.dumpObject() (WARNING: this destructively modifies -- `list`). This is similar to serialCommaJoin() in [[Module:table]] when used with the `dontTag = true` option, but -- internally uses mw.text.listToText(). local function concat_list(list, conjunction, dump_vals) if dump_vals then for k, v in pairs(list) do list[k] = dump(v) end end return list_to_text(list, nil, conjunction) end -- A helper function for use with generating error-signaling functions in the presence of raw value conversion. Format a -- message `msg`, including the processed value `processed` if it is different from the raw value `rawval`; otherwise, -- just return `msg`. local function msg_with_processed(msg, rawval, processed) if rawval == processed then return msg end local processed_type = type(processed) return format("%s (processed value %s)", msg, (processed_type == "string" or processed_type == "number") and processed or dump(processed) ) end -- Separate form of tags with ampersand (&). local function split_tags_on_ampersand(tags) return split(tags, "&") end -------------------------------------- Error handling ----------------------------- local function process_error(fmt, ...) local args = {...} for i, val in ipairs(args) do args[i] = dump(val) end if type(fmt) == "table" then -- hacky signal that we're called from internal_process_error(), and not to omit stack frames return error(format(fmt[1], unpack(args))) end return error(format(fmt, unpack(args)), 3) end local function internal_process_error(fmt, ...) process_error({"Internal error in `params` table: " .. fmt}, ...) end -- Check that a parameter or argument is in the form form Scribunto normalizes input argument keys into (e.g. 1 not "1", "foo" not " foo "). Otherwise, it won't be possible to normalize inputs in the expected way. Unless is_argument is set, also check that the name only contains one placeholder at most, and that strings don't resolve to numeric keys once the placeholder has been substituted. local function validate_name(name, desc, extra_name, is_argument) local normalized = scribunto_parameter_key(name) if name and name == normalized then if is_argument or type(name) ~= "string" then return end local placeholder = find(name, "\1", nil, true) if not placeholder then return elseif find(name, "\1", placeholder + 1, true) then error(format( "Internal error: expected %s to only contain one placeholder, but saw %s", extra_name and (desc .. dump(extra_name)) or desc, dump(name) )) end local first_name = gsub(name, "\1", "1") normalized = scribunto_parameter_key(first_name) if first_name == normalized then return end error(format( "Internal error: %s cannot resolve to numeric parameters once any placeholder has been substituted, but %s resolves to %s", extra_name and (desc .. dump(extra_name)) or desc, dump(name), dump(normalized) )) elseif normalized == nil then error(format( "Internal error: expected %s to be of type string or number, but saw %s", extra_name and (desc .. dump(extra_name)) or desc, type(name) )) end error(format( "Internal error: expected %s to be Scribunto-compatible: %s (a %s) should be %s (a %s)", extra_name and (desc .. dump(extra_name)) or desc, dump(name), type(name), dump(normalized), type(normalized) )) end local function validate_alias_options(...) local invalid = { required = true, default = true, template_default = true, allow_holes = true, disallow_holes = true, disallow_missing = true, } function validate_alias_options(param, name, main_param, alias_of) for k in pairs(param) do if invalid[k] then track("bad alias option") -- internal_process_error( -- "parameter %s cannot have the option %s, as it is an alias of parameter %s.", -- name, option, alias_of -- ) end end -- Soon, aliases will inherit options from the main parameter via __index. Track cases where this would happen. if main_param ~= true then for k in pairs(main_param) do if param[k] == nil and not invalid[k] then if k == "list" then -- these need to be changed to list = false to retain current behaviour track("mismatched list alias option") elseif not (k == "type" or k == "set" or k == "sublist") then -- rarely specified on aliases, as they're effectively inherited already track("mismatched alias option") end end end end end validate_alias_options(...) end -- TODO: give ranges instead of long lists, if possible. --[==[ func: export.params_list_error(params, msg) Given a key-value table of raw parameters `params`, display an error message about all the parameters seen in the table. The parameter names are displayed in sorted order. `msg` should be e.g. {"required"} or {"not used by this template"}. This is used internally to display error messages about required or invalid parameters, and can be used for the same purpose by code that processes its own parameters (e.g. if the `return_unknown` flag is specified to `process`). ]==] local function params_list_error(params, msg) local list, n = {}, 0 for name in sorted_pairs(params) do n = n + 1 list[n] = name end error(format( "Parameter%s %s.", format(n == 1 and " %s is" or "s %s are", concat_list(list, " and ", true)), msg ), 3) end export.params_list_error = params_list_error -- Helper function for use with convert_val_error(). Format a list of possible choices using `concat_list` and -- conjunction "or", displaying "either " before the choices if there's more than one. local function format_choice_list(valid) return (#valid > 1 and "either " or "") .. concat_list(valid, " or ") end -- Signal an error for a value `val` that is not of the right type `valid` (which is either a string specifying a type, or -- a list of possible values, in the case where `set` was used). `name` is the name of the parameter and can be a -- function to signal an error (which is assumed to automatically display the parameter's name and value). `seetext` is -- an optional additional explanatory link to display (e.g. [[WT:LOL]], the list of possible languages and codes). local function convert_val_error(val, name, valid, seetext) if is_callable(name) then if type(valid) == "table" then valid = "choice, must be " .. format_choice_list(valid) end name(format("Invalid %s; the value %s is not valid%s", valid, val, seetext and "; see " .. seetext or "")) else if type(valid) == "table" then valid = format_choice_list(valid) else valid = "a valid " .. valid end error(format("Parameter %s must be %s; the value %s is not valid.%s", dump(name), valid, dump(val), seetext and " See " .. seetext .. "." or "")) end end -- Generate the appropriate error-signaling function given parameter value `val` and name `name`. If `name` is already -- a function, it is just returned; otherwise a function is generated and returned that displays the passed-in messaeg -- along with the parameter's name and value. local function make_parse_err(val, name) if is_callable(name) then return name end return function(msg) error(format("%s: parameter %s=%s", msg, name, val)) end end -------------------------------------- Value conversion ----------------------------- -- For a list parameter `name` and corresponding value `list_name` of the `list` field (which should have the same value -- as `name` if `list = true` was given), generate a pattern to match parameters of the list and store the pattern as a -- key in `patterns`, with corresponding value set to `name`. For example, if `list_name` is "tr", the pattern will -- match "tr" as well as "tr1", "tr2", ..., "tr10", "tr11", etc. If the `list_name` contains a \1 in it, the numeric -- portion goes in place of the \1. For example, if `list_name` is "f\1accel", the pattern will match "faccel", -- "f1accel", "f2accel", etc. Any \1 in `name` is removed before storing into `patterns`. local function save_pattern(name, list_name, patterns) name = type(name) == "string" and gsub(name, "\1", "") or name if find(list_name, "\1", nil, true) then patterns["^" .. gsub(pattern_escape(list_name), "\1", "([1-9]%%d*)") .. "$"] = name else patterns["^" .. pattern_escape(list_name) .. "([1-9]%d*)$"] = name list_name = list_name .. "\1" end validate_name(list_name, "the list field of parameter ", name) return patterns end -- A helper function for use with `sublist`. It is an iterator function for use in a for-loop that returns split -- elements of `val` using `sublist` (a Lua split pattern; boolean `true` to split on commas optionally surrounded by -- whitespace; "comma without whitespace" to split only on commas not followed by whitespace which have not been escaped -- by a backslash; or a function to do the splitting, which is passed two values, the value to split and a function to -- signal an error, and should return a list of the split elements). `name` is the parameter name or error-signaling -- function passed into convert_val(). local function split_sublist(val, name, sublist) if sublist == true then return gsplit(val, "%s*,%s*") -- Split an argument on comma, but not comma followed by whitespace. elseif sublist == "comma without whitespace" then -- If difficult cases, use split_on_comma. if find(val, "\\", nil, true) or match(val, ",%s") then return iterate_list(split_on_comma(val)) end -- Otherwise, use gsplit. return gsplit(val, ",") elseif type(sublist) == "string" then return gsplit(val, sublist) elseif not is_callable(sublist) then error(format('Internal error: expected `sublist` to be of type "string" or "function" or boolean `true`, but saw %s', dump(sublist))) end return iterate_list(sublist(val, make_parse_err(val, name))) end -- For parameter named `name` with value `val` and param spec `param`, if the `set` field is specified, verify that the -- value is one of the one specified in `set`, and throw an error otherwise. `name` is taken directly from the -- corresponding parameter passed into convert_val() and may be a function to signal an error. Optional `param_type` is -- a string specifying the conversion type of `val` and is used for special-casing: If `param_type` is "boolean", an -- internal error is thrown (since `set` cannot be used in conjunction with booleans) and if `param_type` is "number", -- no checking happens because in this case `set` contains numbers and is checked inside the number conversion function -- itself, after converting `val` to a number. Return the canonical value of `val` (which may be different from `val` -- if an alias map is given). local function check_set(val, name, param, param_type) if param_type == "boolean" then error(format('Internal error: cannot use `set` with `type = "%s"`', param_type)) -- Needs to be special cased because the check happens after conversion to numbers. elseif param_type == "number" then return val end local set, map = param.set if sets == nil then map = finalize_set(set, name) sets = {[set] = map} else map = sets[set] if map == nil then map = finalize_set(set, name) sets[set] = map end end local newval = map[val] if newval == true then return val elseif newval ~= nil then return newval end local list = {} for k, v in sorted_pairs(map) do if v == true then insert(list, dump(k)) else insert(list, ("%s (alias of %s)"):format(dump(k), dump(v))) end end -- If the parameter is not required then put "or empty" at the end of the list, to avoid implying the parameter is actually required. if not param.required then insert(list, "empty") end convert_val_error(val, name, list) end local function convert_language(val, name, param, allow_etym) local method, func = param.method if method == nil or method == "code" then func, method = get_language_by_code, "code" elseif method == "name" then func, method = get_language_by_name, "name" else error(format('Internal error: expected `method` for type `language` to be "code", "name" or undefined, but saw %s', dump(method))) end local lang = func(val, nil, allow_etym, param.family) if lang then return lang end local list, links = {"language"}, {"[[WT:LOL]]"} if allow_etym then insert(list, "etymology language") insert(links, "[[WT:LOL/E]]") end if param.family then insert(list, "family") insert(links, "[[WT:LOF]]") end convert_val_error(val, name, concat_list(list, " or ") .. " " .. (method == "name" and "name" or "code"), concat_list(links, " and ")) end local function convert_number(val, allow_hex) -- Call tonumber_extended with the `real_finite` flag, which filters out ±infinity and NaN. -- By default, specify base 10, which prevents 0x hex inputs from being converted. -- If `allow_hex` is set, then don't give a base, which means 0x hex inputs will work. local num = tonumber_extended(val, not allow_hex and 10 or nil, "finite_real") if not num then return num end if match(val, "[eEpP.]") then -- float track("number not an integer") end if find(val, "+", nil, true) then track("number with +") end -- Track various unusual number inputs to determine if it should be restricted to positive integers by default (possibly including 0). if not is_positive_integer(num) then track("number not a positive integer") if num == 0 then track("number is 0") elseif not is_integer(num) then track("number not an integer") end end return num end -- TODO: validate parameter specs separately, as it's making the handler code really messy at the moment. local type_handlers = setmetatable({ ["boolean"] = function(val) return yesno(val, true) end, ["family"] = function(val, name, param) local method, func = param.method if method == nil or method == "code" then func, method = get_family_by_code, "code" elseif method == "name" then func, method = get_family_by_name, "name" else error(format('Internal error: expected `method` for type `family` to be "code", "name" or undefined, but saw %s', dump(method))) end return func(val) or convert_val_error(val, name, "family " .. method, "[[WT:LOF]]") end, ["labels"] = function(val, name, param) -- FIXME: Should be able to pass in a parse_err function. return split_labels_on_comma(val) end, ["form of tags"] = function(val, name, param) return split_tags_on_ampersand(val) end, ["language"] = function(val, name, param) return convert_language(val, name, param, true) end, ["full language"] = convert_language, ["number"] = function(val, name, param) local allow_hex = param.allow_hex if allow_hex and allow_hex ~= true then error(format( 'Internal error: expected `allow_hex` for type `number` to be of type "boolean" or undefined, but saw %s', dump(allow_hex) )) end local num = convert_number(val, allow_hex) if param.set then -- Don't pass in "number" here; otherwise no checking will happen. num = check_set(num, name, param) end if num then return num end convert_val_error(val, name, (allow_hex and "decimal or hexadecimal " or "") .. "number") end, ["range"] = function(val, name, param) local allow_hex = param.allow_hex if allow_hex and allow_hex ~= true then error(format( 'Internal error: expected `allow_hex` for type `range` to be of type "boolean" or undefined, but saw %s', dump(allow_hex) )) end -- Pattern ensures leading minus signs are accounted for. local m1, m2 = match(val, "^(%s*%S.-)%-(%s*%S.*)") if m1 then m1 = convert_number(m1, allow_hex) if m1 then m2 = convert_number(m2, allow_hex) if m2 then return {m1, m2} end end end -- Try `val` if it couldn't be split into a range, and return a range of `val` to `val` if possible. local num = convert_number(val, allow_hex) if num then return {num, num} end convert_val_error(val, name, (allow_hex and "decimal or hexadecimal " or "") .. "number or a hyphen-separated range of two numbers") end, ["parameter"] = function(val, name, param) -- Use the `no_trim` option, as any trimming will have already been done. return scribunto_parameter_key(val, true) end, ["qualifier"] = function(val, name, param) return {val} end, ["references"] = function(val, name, param) return parse_references(val, make_parse_err(val, name)) end, ["genders"] = function(val, name, param) if not val:find("[,<]") then return {{spec = val}} end -- NOTE: We don't pass in allow_space_around_comma. Consistent with other comma-separated types, there shouldn't -- be spaces around the comma. return parse_gender_and_number_spec { spec = val, parse_err = make_parse_err(val, name), allow_multiple = true, } end, ["script"] = function(val, name, param) local method, func = param.method if method == nil or method == "code" then func, method = get_script_by_code, "code" elseif method == "name" then func, method = get_script_by_name, "name" else error(format('Internal error: expected `method` for type `script` to be "code", "name" or undefined, but saw %s', dump(method))) end return func(val) or convert_val_error(val, name, "script " .. method, "[[WT:LOS]]") end, ["string"] = function(val, name, param) -- To be removed as unnecessary. track("string") return val end, -- TODO: add support for resolving to unsupported titles. -- TODO: split this into "page name" (i.e. internal) and "link target" (i.e. external as well), which is more intuitive. ["title"] = function(val, name, param) local namespace = param.namespace if namespace == nil then namespace = 0 else local valid_type = type(namespace) ~= "number" and 'of type "number" or undefined' or not namespaces[namespace] and "a valid namespace number" or nil if valid_type then error(format('Internal error: expected `namespace` for type `title` to be %s, but saw %s', valid_type, dump(namespace))) end end -- Decode entities. WARNING: mw.title.makeTitle must be called with `decoded` (as it doesn't decode) and mw.title.new must be called with `val` (as it does decode, so double-decoding needs to be avoided). local decoded, prefix, title = decode_entities(val), param.prefix -- If the input is a fragment, treat the title as the current title with the input fragment. if sub(decoded, 1, 1) == "#" then -- If prefix is "force", only get the current title if it's in the specified namespace. current_title includes the namespace prefix. if current_namespace == nil then local current_title = mw_title.getCurrentTitle() current_title_text, current_namespace = current_title.prefixedText, current_title.namespace end if not (prefix == "force" and namespace ~= current_namespace) then title = new_title(current_title_text .. val) end elseif prefix == "force" then -- Unconditionally add the namespace prefix (mw.title.makeTitle). title = make_title(namespace, decoded) elseif prefix == "full override" then -- The first input prefix will be used as an override (mw.title.new). This can be a namespace or interwiki prefix. title = new_title(val, namespace) elseif prefix == nil or prefix == "namespace override" then -- Only allow namespace prefixes to override. Interwiki prefixes therefore need to be treated as plaintext (e.g. "el:All topics" with namespace 14 returns "el:Category:All topics", but we want "Category:el:All topics" instead; if the former is really needed, then the input ":el:Category:All topics" will work, as the initial colon overrides the namespace). mw.title.new can take namespace names as well as numbers in the second argument, and will throw an error if the input isn't a valid namespace, so this can be used to determine if a prefix is for a namespace, since mw.title.new will return successfully only if there's either no prefix or the prefix is for a valid namespace (in which case we want the override). local success success, title = pcall(new_title, val, match(decoded, "^.-%f[:]") or namespace) -- Otherwise, get the title with mw.title.makeTitle, which unconditionally adds the namespace prefix, but behaves like mw.title.new if the namespace is 0. if not success then title = make_title(namespace, decoded) end else error(format('Internal error: expected `prefix` for type `title` to be "force", "full override", "namespace override" or undefined, but saw %s', dump(prefix))) end local allow_external = param.allow_external if allow_external == true then return title or convert_val_error(val, name, "Wiktionary or external page title") elseif not allow_external then return title and is_internal_title(title) and title or convert_val_error(val, name, "Wiktionary page title") end error(format('Internal error: expected `allow_external` for type `title` to be of type "boolean" or undefined, but saw %s', dump(allow_external))) end, ["Wikimedia language"] = function(val, name, param) local fallback = param.fallback if fallback == true then return get_wm_lang_by_code_with_fallback(val) or convert_val_error(val, name, "Wikimedia language or language code") elseif not fallback then return get_wm_lang_by_code(val) or convert_val_error(val, name, "Wikimedia language code") end error(format('Internal error: expected `fallback` for type `Wikimedia language` to be of type "boolean" or undefined, but saw %s', dump(fallback))) end, }, { -- TODO: decode HTML entities in all input values. Non-trivial to implement, because we need to avoid any downstream functions decoding the output from this module, which would be double-decoding. Note that "title" has this implemented already, and it needs to have both the raw input and the decoded input to avoid double-decoding by me.title.new, so any implementation can't be as simple as decoding in __call then passing the result to the handler. __call = function(self, val, name, param, param_type, default) local val_type = type(val) -- TODO: check this for all possible parameter types. if val_type == param_type then return val elseif val_type ~= "string" then local expected = "string" if default and (param_type == "boolean" or param_type == "number") then expected = param_type .. " or " .. expected end error(format( "Internal error: %sargument %s has the type %s; expected a %s.", default and (default .. " for ") or "", name, dump(val_type), expected )) end local func = self[param_type] if func == nil then error(format("Internal error: %s is not a recognized parameter type.", dump(param_type))) end return func(val, name, param) end }) --[==[ func: export.convert_val(val, name, param) Convert a parameter value according to the associated specs listed in the `params` table passed to [[Module:parameters]]. `val` is the value to convert for a parameter whose name is `name` (used only in error messages). `param` is the spec (the value part of the `params` table for the parameter). In place of passing in the parameter name, `name` can be a function that throws an error, displaying the specified message along with the parameter name and value. This function processes all the conversion-related fields in `param`, including `type`, `set`, `sublist`, `convert`, etc. It returns the converted value. ]==] local function convert_val(val, name, param, default) local param_type = param.type or "string" -- If param.type is a function, resolve it to a recognized type. if is_callable(param_type) then param_type = param_type(val) end local convert, sublist = param.convert, param.sublist -- `val` might not be a string if it's the default value. if sublist and type(val) == "string" then local retlist, set = {}, param.set if convert then local thisindex, thisval, insval, parse_err = 0 if is_callable(name) then -- We assume the passed-in error function in `name` already shows the parameter name and raw value. function parse_err(msg) name(format("%s: item #%s=%s", msg_with_processed(msg, thisval, insval), thisindex, thisval) ) end else function parse_err(msg) error(format("%s: item #%s=%s of parameter %s=%s", msg_with_processed(msg, thisval, insval), thisindex, thisval, name, val) ) end end for v in split_sublist(val, name, sublist) do thisindex, thisval = thisindex + 1, v if set then v = check_set(v, name, param, param_type) end insert(retlist, convert(type_handlers(v, name, param, param_type, default), parse_err)) end else for v in split_sublist(val, name, sublist) do if set then v = check_set(v, name, param, param_type) end insert(retlist, type_handlers(v, name, param, param_type, default)) end end return retlist elseif param.set then val = check_set(val, name, param, param_type) end local retval = type_handlers(val, name, param, param_type, default) if convert then local parse_err if is_callable(name) then -- We assume the passed-in error function in `name` already shows the parameter name and raw value. if retval == val then -- This is an optimization to avoid creating a closure. The second arm works correctly even -- when retval == val. parse_err = name else function parse_err(msg) name(msg_with_processed(msg, val, retval)) end end else function parse_err(msg) error(format("%s: parameter %s=%s", msg_with_processed(msg, val, retval), name, val)) end end retval = convert(retval, parse_err) end -- If `sublist` is set but the input wasn't a string, return `retval` as a one-item list. if sublist then retval = {retval} end return retval end export.convert_val = convert_val -- used by [[Module:parameter utilities]] local function unknown_param(name, val, args_unknown) track("unknown parameters") args_unknown[name] = val return args_unknown end local function check_string_param_modifier(param_type, name, tag) if param_type and not (param_type == "string" or param_type == "parameter" or is_callable(param_type)) then internal_process_error( "%s cannot be set unless %s is set to %s (the default), %s or a function: parameter %s has the type %s.", tag, "type", "string", "parameter", name, param_type ) end end local function hole_error(params, name, listname, this, nxt, extra) -- `process_error` calls `dump` on values to be inserted into -- error messages, but with numeric lists this causes "numeric" -- to look like the name of the list rather than a description, -- as `dump` adds quote marks. Insert it early to avoid this, -- but add another %s specifier in all other cases, so that -- actual list names will be displayed properly. local offset, specifier, starting_from = 0, "%s", "" local msg = "Item %%d in the list of %s parameters must be given if item %%d is given, because %sthere shouldn't be any gaps due to missing%s parameters." local specs = {} if type(listname) == "string" then specs[2] = listname elseif type(name) == "number" then offset = name - 1 -- To get the original parameter. specifier = "numeric" -- If the list doesn't start at parameter 1, avoid implying -- there can't be any gaps in the numeric parameters if -- some parameter with a lower key is optional. for j = name - 1, 1, -1 do local _param = params[j] if not (_param and _param.required) then starting_from = format("(starting from parameter %d) ", dump(j + 1)) break end end else specs[2] = name end specs[1] = this + offset -- Absolute index for this item. insert(specs, nxt + offset) -- Absolute index for the next item. process_error(format(msg, specifier, starting_from, extra or ""), unpack(specs)) end local function check_disallow_holes(params, val, name, listname, extra) for i = 1, val.maxindex do if val[i] == nil then hole_error(params, name, listname, i, num_keys(val)[i], extra) end end end local function handle_holes(params, val, name) local param = params[name] local disallow_holes = param.disallow_holes -- Iterate up the list, and throw an error if a hole is found. if disallow_holes then check_disallow_holes(params, val, name, param.list, " or empty") end -- Iterate up the list, and throw an error if a hole is found due to a -- missing parameter, treating empty parameters as part of the list. This -- applies beyond maxindex if blank arguments are supplied beyond it, so -- isn't mutually exclusive with `disallow_holes`. local empty = val.empty if param.disallow_missing then if empty then -- Remove `empty` from `val`, so it doesn't get returned. val.empty = nil for i = 1, max(val.maxindex, empty.maxindex) do if val[i] == nil and not empty[i] then local keys = extend(num_keys(val), num_keys(empty)) sort(keys) hole_error(params, name, param.list, i, keys[i]) end end -- If there's no table of empty parameters, the check is identical to -- `disallow_holes`, except that the error message only refers to -- missing parameters, not missing or empty ones. If `disallow_holes` is -- also set, there's no point checking again. elseif not disallow_holes then check_disallow_holes(params, val, name, param.list) end end -- If `allow_holes` is set, there's nothing left to do. if param.allow_holes then -- do nothing -- Otherwise, remove any holes: `pairs` won't work, as it's unsorted, and -- iterating from 1 to `maxindex` times out with inputs like |100000000000=, -- so use num_keys to get a list of numerical keys sorted from lowest to -- highest, then iterate up the list, moving each value in `val` to the -- lowest unused positive integer key. This also avoids the need to create a -- new table. If `disallow_holes` is specified, then there can't be any -- holes in the list, so there's no reason to check again; this doesn't -- apply to `disallow_missing`, however. else if not disallow_holes then local keys, i = num_keys(val), 0 while true do i = i + 1 local key = keys[i] if key == nil then break elseif i ~= key then track("holes compressed") val[i], val[key] = val[key], nil end end end -- Some code depends on only numeric params being present when no holes are -- allowed (e.g. by checking for the presence of arguments using next()), so -- remove `maxindex`. val.maxindex = nil end end local function maybe_flatten(params, val, name) local param = params[name] if param.flatten then if param.allow_holes then process_error("For parameter %s, can't set both `allow_holes` and `flatten`", name) end if not param.sublist and param.type ~= "genders" and param.type ~= "labels" and param.type ~= "references" and param.type ~= "qualifier" and param.type ~= "form of tags" then process_error("For parameter %s, can only set `flatten` along with `sublist` or a list-generating type", name) end -- Do the flattening ourselves rather than calling flatten() in [[Module:table]], which will attempt to -- flatten non-list objects like title objects, and cause an error in the process. -- FIXME: We should do this in-place if possible. local newlist = {} for _, sublist in ipairs(val) do for _, item in ipairs(sublist) do insert(newlist, item) end end val = newlist end return val end -- If both `template_default` and `default` are given, `template_default` takes precedence, but only on the template or -- module page. This means a different default can be specified for the template or module page example. However, -- `template_default` doesn't apply if any args are set, which helps (somewhat) with examples on documentation pages -- transcluded into the template page. HACK: We still run into problems on documentation pages transcluded into the -- template page when pagename= is set. Check this on the assumption that pagename= is fairly standard. local function convert_default_val(name, param, pagename_set, any_args_set, add_empty_sublist) if not pagename_set then local val = param.template_default if val ~= nil and not any_args_set and is_own_page() then return convert_val(val, name, param, "template default") end end local val = param.default if val ~= nil then return convert_val(val, name, param, "default") -- Sublist parameters should return an empty table if not given, but only do -- this if the parameter isn't also a list (in which case it will already -- be an empty table). -- FIXME: do this once all modules that pass in a sublist parameter treat an empty sublist identically to a nil argument; some currently do things based on the fact an argument exists at all. -- elseif add_empty_sublist and param.sublist then --return {} end end --[==[ Process arguments with a given list of parameters. Return a table containing the processed arguments. The `args` parameter specifies the arguments to be processed; they are the arguments you might retrieve from {frame:getParent().args} (the template arguments) or in some cases {frame.args} (the invocation arguments). The `params` parameter specifies a list of valid parameters, and consists of a table. If an argument is encountered that is not in the parameter table, an error is thrown. The structure of the `params` table is as described above in the intro comment. '''WARNING:''' The `params` table is destructively modified to save memory. Nonetheless, different keys can share the same value objects in memory without causing problems. The `return_unknown` parameter, if set to {true}, prevents the function from triggering an error when it comes across an argument with a name that it doesn't recognise. Instead, the return value is a pair of values: the first is the processed arguments as usual, while the second contains all the unrecognised arguments that were left unprocessed. This allows you to do multi-stage processing, where the entire set of arguments that a template should accept is not known at once. For example, an inflection-table might do some generic processing on some arguments, but then defer processing of the remainder to the function that handles a specific inflectional type. ]==] function export.process(args, params, return_unknown) -- Process parameters for specific properties local args_new, args_unknown, any_args_set, required, patterns, list_args, index_list, args_placeholders, placeholders_n = {} -- TODO: memoize the processing of each unique `param` value, since it's common for the same value to be used for many parameter names. for name, param in pairs(params) do validate_name(name, "parameter names") if param ~= true then local spec_type = type(param) if type(param) ~= "table" then internal_process_error( "spec for parameter %s must be a table of specs or the value true, but found %s.", name, spec_type ~= "boolean" and spec_type or param ) end -- Populate required table, and make sure aliases aren't set to required. if param.required then if required == nil then required = {} end required[name] = true end local listname, alias_of = param.list, param.alias_of if alias_of then validate_name(alias_of, "the alias_of field of parameter ", name) if alias_of == name then internal_process_error( "parameter %s cannot be an alias of itself.", name ) end local main_param = params[alias_of] -- Check that the alias_of is set to a valid parameter. if not (main_param == true or type(main_param) == "table") then internal_process_error( "parameter %s is an alias of an invalid parameter.", name ) end validate_alias_options(param, name, main_param, alias_of) -- Aliases can't be lists unless the canonical parameter is also a list. if listname and (main_param == true or not main_param.list) then internal_process_error( "list parameter %s is set as an alias of %s, which is not a list parameter.", name, alias_of ) -- Can't be an alias of an alias. elseif main_param ~= true then local main_alias_of = main_param.alias_of if main_alias_of ~= nil then internal_process_error( "alias_of cannot be set to another alias: parameter %s is set as an alias of %s, which is in turn an alias of %s. Set alias_of for %s to %s.", name, alias_of, main_alias_of, name, main_alias_of ) end end end local replaced_by = param.replaced_by if replaced_by then -- replaced_by can be `false`, which is OK validate_name(replaced_by, "the replaced_by field of parameter ", name) if replaced_by == name then internal_process_error( "parameter %s cannot be replaced by itself.", name ) end local main_param = params[replaced_by] -- Check that the replaced_by is set to a valid parameter. if not (main_param == true or type(main_param) == "table") then internal_process_error( "parameter %s is set to be replaced by an invalid parameter.", name ) end -- Can't be a replaced-by of a replaced-by. if main_param ~= true then local main_replaced_by = main_param.replaced_by if main_replaced_by ~= nil then internal_process_error( "replaced_by cannot be set to another replaced-by parameter: parameter %s is set as replaced by %s, which is in turn replaced by %s. Set replaced_by for %s to %s.", name, replaced_by, main_replaced_by, name, main_replaced_by ) end end if param.instead ~= nil then internal_process_error("the `instead` tag can only be given when `replaced_by` is set to `false`.") end elseif replaced_by == false then if param.instead ~= nil and type(param.instead) ~= "string" then internal_process_error( "the `instead` tag must be a string, but saw %s.", param.instead ) end end if replaced_by ~= nil then if param.reason ~= nil and type(param.reason) ~= "string" then internal_process_error( "the `reason` tag must be a string, but saw %s.", param.reason ) end end if listname then if not alias_of then local key = name if type(name) == "string" then key = gsub(name, "\1", "") end local list_arg = {maxindex = 0} args_new[key] = list_arg if list_args == nil then list_args = {} end list_args[key] = list_arg end local list_type = type(listname) if list_type == "string" then -- If the list property is a string, then it represents the name -- to be used as the prefix for list items. This is for use with lists -- where the first item is a numbered parameter and the -- subsequent ones are named, such as 1, pl2, pl3. patterns = save_pattern(name, listname, patterns or {}) elseif listname ~= true then internal_process_error( "list field for parameter %s must be a boolean, string or undefined, but saw a %s.", name, list_type ) elseif type(name) == "number" then if index_list ~= nil then internal_process_error( "only one numeric parameter can be a list, unless the list property is a string." ) end -- If the name is a number, then all indexed parameters from -- this number onwards go in the list. index_list = name else patterns = save_pattern(name, name, patterns or {}) end if find(name, "\1", nil, true) then if args_placeholders then placeholders_n = placeholders_n + 1 args_placeholders[placeholders_n] = name else args_placeholders, placeholders_n = {name}, 1 end end end end end --Process required changes to `params`. if args_placeholders then for i = 1, placeholders_n do local name = args_placeholders[i] params[gsub(name, "\1", "")], params[name] = params[name], nil end end -- Process the arguments for name, val in pairs(args) do any_args_set = true validate_name(name, "argument names", nil, true) -- Guaranteeing that all values are strings avoids issues with type coercion being inconsistent between functions. local val_type = type(val) if val_type ~= "string" then internal_process_error( "argument %s has the type %s; all arguments must be strings.", name, val_type ) end local orig_name, raw_type, index, canonical = name, type(name) if raw_type == "number" then if index_list and name >= index_list then index = name - index_list + 1 name = index_list end elseif patterns then -- Does this argument name match a pattern? for pattern, pname in next, patterns do index = match(name, pattern) -- It matches, so store the parameter name and the -- numeric index extracted from the argument name. if index then index = tonumber(index) name = pname break end end end local param = params[name] -- If the argument is not in the list of parameters, store it in a separate list. if not param then args_unknown = unknown_param(name, val, args_unknown or {}) elseif param == true then canonical = orig_name val = php_trim(val) if val ~= "" then -- If the parameter is duplicated, throw an error. if args_new[name] ~= nil then process_error( "Parameter %s has been entered more than once. This is probably because a parameter alias has been used.", canonical ) end args_new[name] = val end else if param.replaced_by == false then process_error( ("Parameter %%s has been removed and is no longer valid%s.%s"):format( param.reason and ", " .. param.reason or "", param.instead and " Instead, " .. param.instead .. "." or ""), name ) elseif param.replaced_by then process_error( ("Parameter %%s has been replaced by %%s%s."):format( param.reason and ", " .. param.reason or ""), name, param.replaced_by ) end if param.deprecated then track("deprecated parameter", name) end if param.require_index then -- Disallow require_index for numeric parameter names, as this doesn't make sense. if raw_type == "number" then internal_process_error( "cannot set require_index for numeric parameter %s.", name ) -- If a parameter without the trailing index was found, and -- require_index is set on the param, treat it -- as if it isn't recognized. elseif not index then args_unknown = unknown_param(name, val, args_unknown or {}) end end -- Check that separate_no_index is not being used with a numeric parameter. if param.separate_no_index then if raw_type == "number" then internal_process_error( "cannot set separate_no_index for numeric parameter %s.", name ) elseif type(param.alias_of) == "number" then internal_process_error( "cannot set separate_no_index for parameter %s, as it is an alias of numeric parameter %s.", name, param.alias_of ) end end -- If no index was found, use 1 as the default index. -- This makes list parameters like g, g2, g3 put g at index 1. -- If `separate_no_index` is set, then use 0 as the default instead. if not index and param.list then index = param.separate_no_index and 0 or 1 end -- Normalize to the canonical parameter name. If it's a list, but the alias is not, then determine the index. local raw_name = param.alias_of if raw_name then raw_type = type(raw_name) if raw_type == "number" then name = raw_name local main_param = params[raw_name] if main_param ~= true and main_param.list then if not index then index = param.separate_no_index and 0 or 1 end canonical = raw_name + index - 1 else canonical = raw_name end else name = gsub(raw_name, "\1", "") local main_param = params[name] if not index and main_param ~= true and main_param.list then index = param.separate_no_index and 0 or 1 end if not index or index == 0 then canonical = name elseif name == raw_name then canonical = name .. index else canonical = gsub(raw_name, "\1", index) end end else canonical = orig_name end -- Only recognize demo parameters if this is the current template or module's -- page, or its documentation page. if param.demo and not is_own_page("include_documentation") then args_unknown = unknown_param(name, val, args_unknown or {}) end -- Remove leading and trailing whitespace unless no_trim is true. if param.no_trim then check_string_param_modifier(param.type, name, "no_trim") else val = php_trim(val) end -- Empty string is equivalent to nil unless allow_empty is true. if param.allow_empty then check_string_param_modifier(param.type, name, "allow_empty") elseif val == "" then -- If `disallow_missing` is set, keep track of empty parameters -- via the `empty` field in `arg`, which will be used by the -- `disallow_missing` check. This will be deleted before -- returning. if index and param.disallow_missing then local arg = args_new[name] local empty = arg.empty if empty == nil then empty = {maxindex = 0} arg.empty = empty end empty[index] = true if index > empty.maxindex then empty.maxindex = index end end val = nil end -- Allow boolean false. if val ~= nil then -- Convert to proper type if necessary. local main_param = params[raw_name] if main_param ~= true then val = convert_val(val, orig_name, main_param or param) end -- Mark it as no longer required, as it is present. if required then required[name] = nil end -- Store the argument value. if index then local arg = args_new[name] -- If the parameter is duplicated, throw an error. if arg[index] ~= nil then process_error( "Parameter %s has been entered more than once. This is probably because a list parameter has been entered without an index and with index 1 at the same time, or because a parameter alias has been used.", canonical ) end arg[index] = val -- Store the highest index we find. local maxindex = arg.maxindex if index > maxindex then maxindex = index end if arg[0] ~= nil then arg.default, arg[0] = arg[0], nil if maxindex < 1 then maxindex = 1 end end arg.maxindex = maxindex if not params[name].list then args_new[name] = val -- Don't store index 0, as it's a proxy for the default. elseif index > 0 then arg[index] = val end else -- If the parameter is duplicated, throw an error. if args_new[name] ~= nil then process_error( "Parameter %s has been entered more than once. This is probably because a parameter alias has been used.", canonical ) end if not raw_name then args_new[name] = val else local main_param = params[raw_name] if main_param ~= true and main_param.list then local main_arg = args_new[raw_name] main_arg[1] = val -- Store the highest index we find. if main_arg.maxindex < 1 then main_arg.maxindex = 1 end else args_new[raw_name] = val end end end end end end -- Remove holes in any list parameters if needed. This must be handled -- straight after the previous loop, as any instances of `empty` need to be -- converted to nil. if list_args then for name, val in next, list_args do handle_holes(params, val, name) end end -- If the current page is the template which invoked this Lua instance, then ignore the `require` flag, as it -- means we're viewing the template directly. Required parameters sometimes have a `template_default` key set, -- which gets used in such cases as a demo. -- Note: this won't work on other pages in the Template: namespace (including the /documentation subpage), -- or if the #invoke: is on a page in another namespace. local pagename_set = args_new.pagename -- Handle defaults. for name, param in pairs(params) do if param ~= true then local arg_new = args_new[name] if arg_new == nil then args_new[name] = convert_default_val(name, param, pagename_set, any_args_set, true) elseif param.list and arg_new[1] == nil then local default_val = convert_default_val(name, param, pagename_set, any_args_set) if default_val ~= nil then arg_new[1] = default_val if arg_new.maxindex == 0 then arg_new.maxindex = 1 end end end end end -- Flatten nested lists if called for. This must come after setting the default. if list_args then for name, val in next, list_args do args_new[name] = maybe_flatten(params, val, name) end end -- The required table should now be empty. -- If any parameters remain, throw an error, unless we're on the current template or module's page. if required and next(required) ~= nil and not is_own_page() then params_list_error(required, "required") -- Return the arguments table. -- If there are any unknown parameters, throw an error, unless return_unknown is set, in which case return args_unknown as a second return value. elseif return_unknown then return args_new, args_unknown or {} elseif args_unknown and next(args_unknown) ~= nil then params_list_error(args_unknown, "not used by this template") end return args_new end return export c23kzesggrlowmn06oii7zu1bbzht4t 828 34335 184523 2026-05-21T06:15:23Z P1ayer 1197 创建页面，内容为“local export = {} local math_module = "Module:math" local dump = mw.dumpObject local format = string.format local gsub = string.gsub local match = string.match local php_trim -- defined below local sub = string.sub local tonumber = tonumber local tostring = tostring local type = type do local php_htmlspecialchars_data local function get_php_htmlspecialchars_data() php_htmlspecialchars_data, get_php_htmlspecialchars_data = { ["\""] = "&quot;", ["&"]…” 184523 Scribunto text/plain local export = {} local math_module = "Module:math" local dump = mw.dumpObject local format = string.format local gsub = string.gsub local match = string.match local php_trim -- defined below local sub = string.sub local tonumber = tonumber local tostring = tostring local type = type do local php_htmlspecialchars_data local function get_php_htmlspecialchars_data() php_htmlspecialchars_data, get_php_htmlspecialchars_data = { ["\""] = "&quot;", ["&"] = "&amp;", ["'"] = "&#039;", ["<"] = "&lt;", [">"] = "&gt;", }, nil return php_htmlspecialchars_data end --[==[Lua equivalent of PHP's {{code|php|htmlspecialchars($string)}}, which converts the characters `&"'<>` to HTML entities. If the `quotes` flag is set to {"compat"}, then `'` will not be converted, and if it is set to {"noquotes"}, then neither `"` nor `'` will be converted.]==] function export.php_htmlspecialchars(str, quotes) if quotes == nil or quotes == "quotes" then quotes = "'\"" elseif quotes == "compat" then quotes = "\"" elseif quotes == "noquotes" then quotes = "" else local quotes_type = type(quotes) error('`quotes` must be "quotes", "compat", "noquotes" or nil; received ' .. (quotes_type == "string" and dump(quotes) or "a " .. quotes_type)) end return (gsub(str, "[&<>" .. quotes .. "]", php_htmlspecialchars_data or get_php_htmlspecialchars_data())) end end do local function tonumber_extended(...) tonumber_extended = require(math_module).tonumber_extended return tonumber_extended(...) end -- Normalizes a string for use in comparisons which emulate PHP's equals -- operator, which coerces certain strings to numbers: those within the -- range -2^63 to 2^63 - 1 which don't have decimal points or exponents are -- coerced to integers, while any others are coerced to doubles if possible; -- otherwise, they remain as strings. PHP and Lua have the same precision -- for doubles, but Lua's integer precision range is -2^53 + 1 to 2^53 - 1. -- Any integers within Lua's precision, as well as all doubles, are simply -- coerced to numbers, but PHP integers outside of Lua's precision are -- emulated as normalized strings, with leading 0s and any + sign removed. -- The `not_long` flag is used for the second comparator if the first did -- not get normalized to a long integer, as PHP will only coerce strings to -- integers if it's possible for both comparators. local function php_normalize_string(str, not_long) local num = tonumber_extended(str, nil, true) -- Must be a number that isn't ±infinity, NaN or hexadecimal. if not num or match(str, "^%s*[+-]?0[xX]()") then return str -- If `not_long` is set or `num` is within Lua's precision, return as a -- number. elseif not_long or num < 9007199254740992 and num > -9007199254740992 then return num, "number" end -- Check if it could be a long integer, and return as a double if not. local sign, str_no_0 = match(str, "^%s*([+-]?)0*(%d+)$") if not str_no_0 then return num, "number" end -- Otherwise, check if it's a long integer. 2^63 is 9223372036854775808, -- so slice off the last 15 digits and deal with the two parts -- separately. If the integer value would be too high/low, return as a -- string. local high = tonumber(sub(str_no_0, 1, -16)) if high > 9223 then return str elseif high == 9223 then local low = tonumber(sub(str_no_0, -15)) -- Range is -2^63 to 2^63 - 1 (not symmetrical). if low > 372036854775808 or low == 372036854775808 and sign ~= "-" then return str end end return (sign == "+" and "" or sign) .. str_no_0, "long integer", num end --[==[Lua equivalent of PHP's {{code|php|===}} operator for strings.]==] function export.php_string_equals(str1, str2) if str1 == str2 then return true end local str1, str1_type, str1_num = php_normalize_string(str1) if str1 == str2 then return true elseif str1_type == "long integer" then local str2, str2_type = php_normalize_string(str2) return str2 == (str2_type == "number" and str1_num or str1) elseif str1_type == "number" then return str1 == php_normalize_string(str2, true) end return false end end --[==[Lua equivalent of PHP's {{code|php|trim($string)}}, which trims {"\0"}, {"\t"}, {"\n"}, {"\v"}, {"\r"} and {" "}. This is useful when dealing with template parameters, since the native parser trims them like this.]==] function export.php_trim(str) return match(str, "[^ \t-\v\r%z].*%f[ \t-\v\r%z]") or "" end php_trim = export.php_trim --[==[Lua equivalent of PHP's {{code|php|ltrim($string)}}, which trims {"\0"}, {"\t"}, {"\n"}, {"\v"}, {"\r"} and {" "} from the beginning of the input string.]==] function export.php_ltrim(str) return (gsub(str, "^[ \t-\v\r%z]+", "")) end --[==[Lua equivalent of PHP's {{code|php|rtrim($string)}}, which trims {"\0"}, {"\t"}, {"\n"}, {"\v"}, {"\r"} and {" "} from the end of the input string.]==] function export.php_rtrim(str) return match(str, "^.+%f[ \t-\v\r%z]") or "" end --[==[Takes a template or module parameter name as either a string or number, and returns the Scribunto-normalized form (i.e. the key that that parameter would have in a {frame.args} table). For example, {"1"} (a string) is normalized to {1} (a number), {" foo "} is normalized to {"foo"}, and {1.5} (a number) is normalized to {"1.5"} (a string). Inputs which cannot be normalized (e.g. booleans) return {nil}. Strings are trimmed with {export.php_trim}, unless the `no_trim` flag is set. If it is, then string parameters are not trimmed, but strings may still be converted to numbers if they do not contain whitespace; this is necessary when normalizing keys into the form received by PHP during callbacks, before any trimming occurs (e.g. in the table of arguments when calling {frame:expandTemplates()}). After trimming (if applicable), keys are then converted to numbers if '''all''' of the following are true: # They are integers; i.e. no decimals or leading zeroes (e.g. {"2"}, but not {"2.0"} or {"02"}). # They are ≤ 2{{sup|53}} and ≥ -2{{sup|53}}. # There is no leading sign unless < 0 (e.g. {"2"} or {"-2"}, but not {"+2"} or {"-0"}). # They contain no leading or trailing whitespace (which may be present when the `no_trim` flag is set). Numbers are converted to strings if '''either''': # They are not integers (e.g. {1.5}). # They are > 2{{sup|53}} or < -2{{sup|53}}. When converted to strings, integers ≤ 2{{sup|63}} and ≥ -2{{sup|63}} are formatted as integers (i.e. all digits are given), which is the range of PHP's integer precision, though the actual output may be imprecise since Lua's integer precision is > 2{{sup|53}} to < -2{{sup|53}}. All other numbers use the standard formatting output by {tostring()}.]==] function export.scribunto_parameter_key(key, no_trim) local key_type = type(key) if key_type == "string" then if not no_trim then key = php_trim(key) end if match(key, "^()-?[1-9]%d*$") then local num = tonumber(key) -- Lua integers are only precise to 2^53 - 1, so specifically check -- for 2^53 and -2^53 as strings, since a numerical comparison won't -- work as it can't distinguish 2^53 from 2^53 + 1. return ( num <= 9007199254740991 and num >= -9007199254740991 or key == "9007199254740992" or key == "-9007199254740992" ) and num or key end return key == "0" and 0 or key elseif key_type == "number" then -- No special handling needed for inf or NaN. return key % 1 == 0 and ( key <= 9007199254740992 and key >= -9007199254740992 and key or key <= 9223372036854775808 and key >= -9223372036854775808 and format("%d", key) ) or tostring(key) end return nil end --[==[Takes a template or module parameter value as either a string, number or boolean, and returns the Scribunto-normalized form (i.e. the value that that parameter would have in a {frame.args} table), which is always a string. For example, {"foo"} remains the same, {2} (a number) is normalized to {"2"} (a string), {true} is normalized to {"1"}, and {false} is normalized to {""}. Inputs which cannot be normalized (e.g. tables) return {nil}. By default, returned values are not trimmed, which matches the treatment of unnamed parameters (e.g. `bar` in {{tl|<nowiki/>foo|bar}}). If the `named` flag is set, then returned values will be trimmed, which matches the treatment of named parameters (e.g. `baz` in {{tl|<nowiki/>foo|bar=baz}}).]==] function export.scribunto_parameter_value(value, named) local value_type = type(value) if value_type == "string" then return named and php_trim(value) or value elseif value_type == "number" then return tostring(value) elseif value_type == "boolean" then return value and "1" or "" end return nil end return export 90dvo9iqmpdordtuzghbnnl9rw4476r 828 34336 184524 2026-05-21T06:17:30Z P1ayer 1197 创建页面，内容为“local export = {} local debug_track_module = "Module:debug/track" local table_get_unprotected_metatable = "Module:table/getUnprotectedMetatable" local chain -- defined below local chain_iter -- defined below local format = string.format local gmatch = string.gmatch local ipairs = ipairs local is_callable -- defined below local pairs = pairs local pcall = pcall local rawget = rawget local require = require local select = select local tostring = tostring local…” 184524 Scribunto text/plain local export = {} local debug_track_module = "Module:debug/track" local table_get_unprotected_metatable = "Module:table/getUnprotectedMetatable" local chain -- defined below local chain_iter -- defined below local format = string.format local gmatch = string.gmatch local ipairs = ipairs local is_callable -- defined below local pairs = pairs local pcall = pcall local rawget = rawget local require = require local select = select local tostring = tostring local type = type local unpack = unpack or table.unpack -- Lua 5.2 compatibility local unroll -- defined below local xpcall = xpcall local function debug_track(...) debug_track = require(debug_track_module) return debug_track(...) end local function get_unprotected_metatable(...) get_unprotected_metatable = require(table_get_unprotected_metatable) return get_unprotected_metatable(...) end local function _iterString(iter, i) i = i + 1 local char = iter() if char ~= nil then return i, char end end -- Iterate over UTF-8-encoded codepoints in string. local function iterString(str) return _iterString, gmatch(str, ".[\128-\191]*"), 0 end --[==[ Return {true} if the input is a function or functor (an object which can be called like a function, because it has a {__call} metamethod). Note: if the input is an object with a {__call} metamethod, but this function is not able to find it because the object's metatable is protected with {__metatable}, then it will return {false} by default, or {nil} if the {allow_maybe} flag is set.]==] function export.is_callable(f, allow_maybe) if type(f) == "function" then return true end -- An object is a functor if it has a `__call` metamethod. The only way to truly confirm this is by trying to call it, but that could be expensive or have side effects, so look for a `__call` metamethod instead. If the metatable is protected with `__metatable`, this may not be possible. local mt = get_unprotected_metatable(f) if mt == nil then return false -- `get_unprotected_metatable` returns false if the metatable is protected. elseif mt == false then debug_track("fun/is_callable/protected metatable") if allow_maybe then return nil end return false end -- `__call` metamethods have to be functions, so don't recurse to check it. local __call = rawget(mt, "__call") return __call and type(__call) == "function" or false end is_callable = export.is_callable --[==[ A version of {xpcall} which takes any arguments to be given to {f} as additional arguments after the error handler. This fixes a deficiency in the standard version of {xpcall}, which is not able to handle arguments to be given to {f}, and brings it in line with {pcall}.]==] function export.xpcall(f, err_handler, ...) -- If there are no arguments, just call xpcall() with `f`. if select("#", ...) == 0 then return xpcall(f, err_handler) end -- Any arguments have to be smuggled in via a table, as ... can't be an -- upvalue, and it's not possible to use pcall() to get aroud this, because -- xpcall() calls the error handler before the stack unwinds. local args = {...} return xpcall(function() return f(unpack(args)) end, err_handler) end do local function catch_values(f, success, ...) if success then return success, ... -- Error message will only take this exact form if `f` is not callable, -- because it will contain a traceback if it was thrown further up the -- stack. elseif (...) == format("attempt to call a %s value", type(f)) then return false end return error(...) end --[==[ A special form of {pcall()}, which returns {true} plus the result value(s) if {f} is callable, or {false} if it isn't. Errors that occur within the called function are not protected.]==] function export.try_call(f, ...) local callable = is_callable(f, true) if callable then return true, f(...) elseif callable == false then return false end -- If `callable` is nil, there's a protected metatable, so there's no way to check without doing a protected call. return catch_values(f, pcall(f, ...)) end end --[==[ Takes two or more functions as arguments, and returns a new function which calls each of the input functions in turn. Any arguments given to the returned function are given to the first function, and all other functions receive the output value(s) from the previous function.]==] function export.chain(func1, func2, ...) local function chained_func(...) return func2(func1(...)) end if select("#", ...) == 0 then return chained_func end return chain(chained_func, ...) end chain = export.chain --[==[ Takes the usual for-loop parameters (an iterator, plus an optional state and initial index), and unrolls the iterator by returning every (first) value returned by the iterator. For instance, {unroll(pairs(t))} will return every key in {t}, and {unroll(string.gmatch(s, "%w+"))} will return every word in {s}.]==] function export.unroll(iter, state, k) k = iter(state, k) if k ~= nil then return k, unroll(iter, state, k) end end unroll = export.unroll --[==[ Takes a generator function (i.e. a function that returns an iterator, such as {ipairs}) and one or more additional functions, and returns a new generator function. Any arguments given to the new generator (e.g. an input table) are given to the original generator, and the additional functions are called on each iteration. The first additional function takes the output from the original iterator (i.e. the function returned by the original generator), and any further functions receive the output value(s) from the previous function. This can be used to modify the values returned from an iterator.]==] function export.chainIter(gen, new_iter, ...) if select("#", ...) > 0 then new_iter = chain(new_iter, ...) end return function(...) local orig_iter, state, k = gen(...) -- k has to be the first value returned by orig_iter on the last iteration, not whatever new_iter returned. local function catch_values(...) k = ... if k ~= nil then return new_iter(...) end end return function() return catch_values(orig_iter(state, k)) end, state, k end end chain_iter = export.chainIter do local function catch_values(start, iter, state, k, ...) if start == k or k == nil then return k, ... end return catch_values(start, iter, state, iter(state, k)) end function export.iterateFrom(start, iter, state, k) local first = true return function(state, k) if first then first = false return catch_values(start, iter, state, iter(state, k)) end return iter(state, k) end, state, k end end -- map(function(number) return number ^ 2 end, -- { 1, 2, 3 }) --> { 1, 4, 9 } -- map(function (char) return string.char(string.byte(char) - 0x20) end, -- "abc") --> { "A", "B", "C" } function export.map(func, iterable, isArray) local array = {} for k, v in (type(iterable) == "string" and iterString or (isArray or iterable[1] ~= nil) and ipairs or pairs)(iterable) do array[k] = func(v, k, iterable) end return array end function export.mapIter(func, iter, state, init) -- init could be anything local array, i = {}, 0 for x, y in iter, state, init do i = i + 1 array[i] = func(y, x, state) end return array end do local function iter_tuples(tuples) local i = tuples.i if i > 1 then i = i - 1 tuples.i = i return unpack(tuples[i]) end end -- Takes an iterator function, and returns a new iterator that iterates in reverse, given the same arguments. -- Note: changes to the state during iteration are not taken into account, since all the return values are calculated in advance. function export.reverseIter(func) return function(...) -- Store all returned values as a list of tuples, then iterate in reverse over that list. local tuples, i, iter, state, val1 = {}, 0, func(...) while true do i = i + 1 local vals = {iter(state, val1)} -- Terminates if the first return value is nil, even if other values are non-nil. val1 = vals[1] if val1 == nil then tuples.i = i return iter_tuples, tuples end tuples[i] = vals end end end end function export.forEach(func, iterable, isArray) for k, v in (type(iterable) == "string" and iterString or (isArray or iterable[1] ~= nil) and ipairs or pairs)(iterable) do func(v, k, iterable) end return nil end ------------------------------------------------- -- From http://lua-users.org/wiki/CurriedLua -- reverse(...) : take some tuple and return a tuple of elements in reverse order -- -- e.g. "reverse(1,2,3)" returns 3,2,1 local function reverse(...) -- reverse args by building a function to do it, similar to the unpack() example local function reverseHelper(acc, v, ...) if select("#", ...) == 0 then return v, acc() else return reverseHelper(function() return v, acc() end, ...) end end -- initial acc is the end of the list return reverseHelper(function() return end, ...) end function export.curry(func, numArgs) -- currying 2-argument functions seems to be the most popular application numArgs = numArgs or 2 -- no sense currying for 1 arg or less if numArgs <= 1 then return func end -- helper takes an argTrace function, and number of arguments remaining to be applied local function curryHelper(argTrace, n) if n == 0 then -- kick off argTrace, reverse argument list, and call the original function return func(reverse(argTrace())) else -- "push" argument (by building a wrapper function) and decrement n return function(onearg) return curryHelper(function() return onearg, argTrace() end, n - 1) end end end -- push the terminal case of argTrace into the function first return curryHelper(function() return end, numArgs) end ------------------------------------------------- -- some(function(val) return val % 2 == 0 end, -- { 2, 3, 5, 7, 11 }) --> true function export.some(func, t, isArray) for k, v in ((isArray or t[1] ~= nil) and ipairs or pairs)(t) do if func(v, k, t) then return true end end return false end -- all(function(val) return val % 2 == 0 end, -- { 2, 4, 8, 10, 12 }) --> true function export.all(func, t, isArray) for k, v in ((isArray or t[1] ~= nil) and ipairs or pairs)(t) do if not func(v, k, t) then return false end end return true end function export.filter(func, t, isArray) local new_t = {} if isArray or t[1] ~= nil then -- array local new_i = 0 for i, v in ipairs(t) do if func(v, i, t) then new_i = new_i + 1 new_t[new_i] = v end end else for k, v in pairs(t) do if func(v, k, t) then new_t[k] = v -- or create array? end end end return new_t end function export.fold(func, t, accum) for i, v in ipairs(t) do accum = func(accum, v, i, t) end return accum end ------------------------------- -- Fancy stuff local function capture(...) local vals = {n = select("#", ...), ...} return function() return unpack(vals, 1, vals.n) end end -- Log input and output of function. -- Receives a function and returns a modified form of that function. function export.logReturnValues(func, prefix) return function(...) local inputValues = capture(...) local returnValues = capture(func(...)) if prefix then mw.log(prefix, inputValues()) mw.log(returnValues()) else mw.log(inputValues()) mw.log(returnValues()) end return returnValues() end end export.log = export.logReturnValues -- Convenience function to make all functions in a table log their input and output. function export.logAll(t) for k, v in pairs(t) do if is_callable(v) then t[k] = export.logReturnValues(v, tostring(k)) end end return t end return export mc1wbmetcrpjp1ihyth6j2ajabvak7y 828 34337 184525 2026-05-21T06:19:46Z P1ayer 1197 创建页面，内容为“local _getmetatable = debug.getmetatable -- For testing (and just in case it gets enabled). if _getmetatable ~= nil then -- Avoid debug.getmetatable() throwing an error if 0 arguments are passed, -- for parity with the other function. return function(t) return _getmetatable(t) end end _getmetatable = getmetatable local pcall = pcall local rawget = rawget local setmetatable = setmetatable local type = type --[==[ Attempts to retrieve the input value's…” 184525 Scribunto text/plain local _getmetatable = debug.getmetatable -- For testing (and just in case it gets enabled). if _getmetatable ~= nil then -- Avoid debug.getmetatable() throwing an error if 0 arguments are passed, -- for parity with the other function. return function(t) return _getmetatable(t) end end _getmetatable = getmetatable local pcall = pcall local rawget = rawget local setmetatable = setmetatable local type = type --[==[ Attempts to retrieve the input value's metatable, and returns it if found. If the value does not have a metatable, returns {nil}. If the input value does have a metatable, but that metatable is not possible to retrieve because it is protected with the `__metatable` metamethod, returns {false}. This is a useful way to ensure that functions can reliably distinguish between objects that do not have metamethods, objects with known metamethods, and objects with unknown metamethods.]==] return function(t) local mt = _getmetatable(t) -- If `mt` is nil, there's no metatable. if mt == nil then return nil -- If `mt` is not a table, the real metatable is protected and there's no -- way of retrieving it. elseif type(mt) ~= "table" then return false end -- Try setting `mt` as the metatable with `setmetatable`; if the metatable -- is protected, this will cause an error to be thrown (revealing it as -- protected), and if it isn't, then `mt` must be the real metatable anyway, -- so nothing has changed. Also make a special exception for data loaded via -- mw.loadData(), which sets each metatable at its own __metatable key as a -- way to stop the use of setmetatable() without actually hiding it. This is -- spoofable, but low-risk. return (pcall(setmetatable, t, mt) or rawget(mt, "mw_loadData") == true) and mt or false end oqd72lzzcljldqpmbzxsx5fm947exar