Seeds of the Word

Module:Citation/CS1/Utilities: Difference between revisions

Browse history interactively
← Older edit
Content deleted Content added
m 1 revision imported
m 1 revision imported
 
(8 intermediate revisions by 2 users not shown)
Line 1:
 
local z = {
error_categorieserror_cats_t = {}; -- for categorizing citations that contain errors
error_ids_t = {}; -- list of error identifiers; used to prevent duplication of certain errors; local to this module
error_ids = {};
error_msgs_t = {}; -- sequence table of error messages
message_tail = {};
maintenance_catsmaint_cats_t = {}; -- for categorizing citations that aren't erroneous per se, but could use a little work
properties_catsprop_cats_t = {}; -- for categorizing citations based on certain properties, language of source for instance
prop_keys_t = {}; -- for adding classes to the citation's <cite> tag
};
 
Line 21:
]]
 
local function is_set ( var )
return not (var == nil or var == '');
end
Line 32:
]]
 
local function in_array ( needle, haystack )
if needle == nil then
return false;
end
for n, v in ipairs( (haystack ) do
if v == needle then
return n;
Line 42:
end
return false;
end
 
 
--[[--------------------------< H A S _ A C C E P T _ A S _ W R I T T E N >------------------------------------
 
When <str> is wholly wrapped in accept-as-written markup, return <str> without markup and true; return <str> and false else
 
with allow_empty = false, <str> must have at least one character inside the markup
with allow_empty = true, <str> the markup frame can be empty like (()) to distinguish an empty template parameter from the specific condition "has no applicable value" in citation-context.
 
After further evaluation the two cases might be merged at a later stage, but should be kept separated for now.
 
]]
 
local function has_accept_as_written (str, allow_empty)
if not is_set (str) then
return str, false;
end
 
local count;
 
if true == allow_empty then
str, count = str:gsub ('^%(%((.*)%)%)$', '%1'); -- allows (()) to be an empty set
else
str, count = str:gsub ('^%(%((.+)%)%)$', '%1');
end
return str, 0 ~= count;
end
 
Line 47 ⟶ 74:
--[[--------------------------< S U B S T I T U T E >----------------------------------------------------------
 
Populates numbered arguments in a message string using an argument table. <args> may be a single string or a
sequence table of multiple strings.
 
]]
 
local function substitute ( msg, args )
return args and mw.message.newRawMessage ( msg, args ):plain() or msg;
end
 
Line 58 ⟶ 86:
--[[--------------------------< E R R O R _ C O M M E N T >----------------------------------------------------
 
Wraps error messages with cssCSS markup according to the state of hidden. <content> may be a single string or a
sequence table of multiple strings.
 
]]
 
local function error_comment ( content, hidden )
return substitute( (hidden and cfg.presentation['hidden-error'] or cfg.presentation['visible-error'], content );
end
 
 
--[[--------------------------< H Y P H E N _ T O _ D A S H >--------------------------------------------------
 
Converts a hyphen to a dash under certain conditions. The hyphen must separate
like items; unlike items are returned unmodified. These forms are modified:
letter - letter (A - B)
digit - digit (4-5)
digit separator digit - digit separator digit (4.1-4.5 or 4-1-4-5)
letterdigit - letterdigit (A1-A5) (an optional separator between letter and
digit is supported – a.1-a.5 or a-1-a-5)
digitletter - digitletter (5a - 5d) (an optional separator between letter and
digit is supported – 5.a-5.d or 5-a-5-d)
 
any other forms are returned unmodified.
 
str may be a comma- or semicolon-separated list
 
]]
 
local function hyphen_to_dash (str)
if not is_set (str) then
return str;
end
 
local accept; -- boolean
 
str = str:gsub ("(%(%(.-%)%))", function(m) return m:gsub(",", "，"):gsub(";", "；") end) -- replace commas and semicolons in accept-as-written markup with similar unicode characters so they'll be ignored during the split
str = str:gsub ('&[nm]dash;', {['&ndash;'] = '–', ['&mdash;'] = '—'}); -- replace &mdash; and &ndash; entities with their characters; semicolon mucks up the text.split
str = str:gsub ('&#45;', '-'); -- replace HTML numeric entity with hyphen character
str = str:gsub ('&nbsp;', ' '); -- replace &nbsp; entity with generic keyboard space character
local out = {};
local list = mw.text.split (str, '%s*[,;]%s*'); -- split str at comma or semicolon separators if there are any
 
for _, item in ipairs (list) do -- for each item in the list
item, accept = has_accept_as_written (item); -- remove accept-this-as-written markup when it wraps all of item
if not accept and mw.ustring.match (item, '^%w*[%.%-]?%w+%s*[%-–—]%s*%w*[%.%-]?%w+$') then -- if a hyphenated range or has endash or emdash separators
if item:match ('^%a+[%.%-]?%d+%s*%-%s*%a+[%.%-]?%d+$') or -- letterdigit hyphen letterdigit (optional separator between letter and digit)
item:match ('^%d+[%.%-]?%a+%s*%-%s*%d+[%.%-]?%a+$') or -- digitletter hyphen digitletter (optional separator between digit and letter)
item:match ('^%d+[%.%-]%d+%s*%-%s*%d+[%.%-]%d+$') or -- digit separator digit hyphen digit separator digit
item:match ('^%d+%s*%-%s*%d+$') or -- digit hyphen digit
item:match ('^%a+%s*%-%s*%a+$') then -- letter hyphen letter
item = item:gsub ('(%w*[%.%-]?%w+)%s*%-%s*(%w*[%.%-]?%w+)', '%1–%2'); -- replace hyphen, remove extraneous space characters
else
item = mw.ustring.gsub (item, '%s*[–—]%s*', '–'); -- for endash or emdash separated ranges, replace em with en, remove extraneous whitespace
end
end
table.insert (out, item); -- add the (possibly modified) item to the output table
end
 
local temp_str = ''; -- concatenate the output table into a comma separated string
temp_str, accept = has_accept_as_written (table.concat (out, ', ')); -- remove accept-this-as-written markup when it wraps all of concatenated out
if accept then
temp_str = has_accept_as_written (str); -- when global markup removed, return original str; do it this way to suppress boolean second return value
return temp_str:gsub("，", ","):gsub("；", ";");
else
return temp_str:gsub("，", ","):gsub("；", ";"); -- else, return assembled temp_str
end
end
 
Line 69 ⟶ 158:
--[=[-------------------------< M A K E _ W I K I L I N K >----------------------------------------------------
 
Makes a wikilink; when botboth link and display text is provided, returns a wikilink in the form [[L|D]]; if only
link is provided (or link and display are the same), returns a wikilink in the form [[L]]; if neither are provided or link is omitted, returns an
provided or link is omitted, returns an empty string.
 
]=]
 
local function make_wikilink (link, display)
if not is_set (link) then return '' end
 
if is_set (display) then
return if table.concatis_set ({'[[',display) and link, '|',~= display, ']]'});then
return table.concat ({'[[', link, '|', display, ']]'});
else
return table.concat ({'[[', link, ']]'});
end
else
return table.concat ({'[[', link, ']]'});
end
end
 
 
--[[--------------------------< S E T _ M E RS RS OA RG E >--------------------------------------------------------------
 
Sets an error conditionmessage and returnsusing the appropriate~/Configuration errorerror_conditions{} message.table along Thewith actualarguments placement of the error messagesupplied in the output isfunction
call, inserts the resulting message in z.error_msgs_t{} sequence table, and returns the error message.
the responsibility of the calling function.
 
<error_id> – key value for appropriate error handler in ~/Configuration error_conditions{} table
<arguments> – may be a single string or a sequence table of multiple strings to be subsititued into error_conditions[error_id].message
<raw> – boolean
true – causes this function to return the error message not wrapped in visible-error, hidden-error span tag;
returns error_conditions[error_id].hidden as a second return value
does not add message to z.error_msgs_t sequence table
false, nil – adds message wrapped in visible-error, hidden-error span tag to z.error_msgs_t
returns the error message wrapped in visible-error, hidden-error span tag; there is no second return value
<prefix> – string to be prepended to <message> -- TODO: remove support for these unused(?) arguments?
<suffix> – string to be appended to <message>
 
TODO: change z.error_cats_t and z.maint_cats_t to have the form cat_name = true? this to avoid dups without having to have an extra table
 
]]
 
local added_maint_cats = {} -- list of maintenance categories that have been added to z.maint_cats_t; TODO: figure out how to delete this table
local function set_error( error_id, arguments, raw, prefix, suffix )
 
local error_state = cfg.error_conditions[ error_id ];
local function set_message (error_id, arguments, raw, prefix, suffix)
local error_state = cfg.error_conditions[error_id];
prefix = prefix or ""'';
suffix = suffix or ""'';
if error_state == nil then
error( (cfg.messages['undefined_error'] .. ': ' .. error_id); -- because missing error handler in Module:Citation/CS1/Configuration
 
elseif is_set( error_state.category ) then
table.insert( elseif z.error_categories,is_set (error_state.category) );then
if error_state.message then -- when error_state.message defined, this is an error message
table.insert (z.error_cats_t, error_state.category);
else
if not added_maint_cats[error_id] then
added_maint_cats[error_id] = true; -- note that we've added this category
table.insert (z.maint_cats_t, substitute (error_state.category, arguments)); -- make cat name then add to table
end
return; -- because no message, nothing more to do
end
end
 
local message = substitute ( error_state.message, arguments );
 
message = table.concat (
Line 124 ⟶ 235:
});
 
z.error_idserror_ids_t[ error_id ] = true;
if z.error_ids_t['err_citation_missing_title'] and -- if missing-title error already noted
if in_array( error_id, { 'bare_url_missing_title', 'trans_missing_title' } )
in_array (error_id, {'err_bare_url_missing_title', 'err_trans_missing_title'}) then -- and this error is one of these
and z.error_ids['citation_missing_title'] then
return '', false; -- don't bother because one flavor of missing title is sufficient
return '', false;
end
message = table.concat ({ prefix, message, suffix });
 
if rawtrue == trueraw then
return message, error_state.hidden; -- return message not wrapped in visible-error, hidden-error span tag
end
 
returnmessage = error_comment( (message, error_state.hidden ); -- wrap message in visible-error, hidden-error span tag
table.insert (z.error_msgs_t, message); -- add it to the messages sequence table
return message; -- and done; return value generally not used but is used as a flag in various functions of ~/Identifiers
end
 
Line 149 ⟶ 262:
alias – one of the list of possible aliases in the aliases lists from Module:Citation/CS1/Configuration
index – for enumerated parameters, identifies which one
enumerated – true/false flag used to choose how enumerated aliases are examined
value – value associated with an alias that has previously been selected; nil if not yet selected
selected – the alias that has previously been selected; nil if not yet selected
Line 167 ⟶ 280:
end
 
if is_set (args[alias]) then -- alias is in the template's argument list
if value ~= nil and selected ~= alias then -- if we have already selected one of the aliases
local skip;
for _, v in ipairs (error_list) do -- spin through the error list to see if we've added this alias
if v == alias then
skip = true;
Line 177 ⟶ 290:
end
if not skip then -- has not been added so
table.insert ( error_list, alias ); -- add error alias to the error list
end
else
Line 190 ⟶ 303:
--[[--------------------------< A D D _ M A I N T _ C A T >------------------------------------------------------
 
Adds a category to z.maintenance_catsmaint_cats_t using names from the configuration file with additional text if any.
To prevent duplication, the added_maint_cats table lists the categories by key that have been added to z.maintenance_catsmaint_cats_t.
 
]]
 
local added_maint_cats = {} -- list of maintenance categories that have been added to z.maintenance_cats
local function add_maint_cat (key, arguments)
if not added_maint_cats [key] then
added_maint_cats [key] = true; -- note that we've added this category
table.insert( (z.maintenance_catsmaint_cats_t, substitute (cfg.maint_cats [key], arguments)); -- make name then add to table
end
end
 
 
--[[--------------------------< A D D _ P R O P _ C A T >--------------------------------------------------------
 
Adds a category to z.prop_cats_t using names from the configuration file with additional text if any.
 
foreign_lang_source and foreign_lang_source_2 keys have a language code appended to them so that multiple languages
may be categorized but multiples of the same language are not categorized.
 
added_prop_cats is a table declared in page scope variables above
 
]]
 
local added_prop_cats = {}; -- list of property categories that have been added to z.prop_cats_t
 
local function add_prop_cat (key, arguments, key_modifier)
local key_modified = key .. ((key_modifier and key_modifier) or ''); -- modify <key> with <key_modifier> if present and not nil
if not added_prop_cats [key_modified] then
added_prop_cats [key_modified] = true; -- note that we've added this category
table.insert (z.prop_cats_t, substitute (cfg.prop_cats [key], arguments)); -- make name then add to table
table.insert (z.prop_keys_t, 'cs1-prop-' .. key); -- convert key to class for use in the citation's <cite> tag
end
end
Line 214 ⟶ 350:
]]
 
local function safe_for_italics ( str )
if not is_set (str) then return str end
 
return str;
if str:sub (1, 1) == "'" then str = "<span></span>" .. str; end
else
if str:sub (-1, -1) == "'" then str = str .. "<span></span>" .. str; end
if str:sub(-1,-1) == "'" then str = str .. "<span></span>"; end
return str:gsub ('\n', ' '); -- Remove newlines as they break italics.
-- Remove newlines as they break italics.
return str:gsub( '\n', ' ' );
end
end
 
Line 236 ⟶ 369:
 
local function wrap_style (key, str)
if not is_set ( str ) then
return "";
elseif in_array( (key, { 'italic-title', 'trans-italic-title' } ) then
str = safe_for_italics ( str );
end
 
return substitute ( cfg.presentation[key], {str} );
end
 
 
--[[--------------------------< M A K E _ S E P _ L I S T >------------------------------------------------------------
 
make a separated list of items using provided separators.
<sep_list> - typically '<comma><space>'
<sep_list_pair> - typically '<space>and<space>'
<sep_list_end> - typically '<comma><space>and<space>' or '<comma><space>&<space>'
 
defaults to cfg.presentation['sep_list'], cfg.presentation['sep_list_pair'], and cfg.presentation['sep_list_end']
if <sep_list_end> is specified, <sep_list> and <sep_list_pair> must also be supplied
 
]]
 
local function make_sep_list (count, list_seq, sep_list, sep_list_pair, sep_list_end)
local list = '';
 
if not sep_list then -- set the defaults
sep_list = cfg.presentation['sep_list'];
sep_list_pair = cfg.presentation['sep_list_pair'];
sep_list_end = cfg.presentation['sep_list_end'];
end
if 2 >= count then
list = table.concat (list_seq, sep_list_pair); -- insert separator between two items; returns list_seq[1] then only one item
elseif 2 < count then
list = table.concat (list_seq, sep_list, 1, count - 1); -- concatenate all but last item with plain list separator
list = table.concat ({list, list_seq[count]}, sep_list_end); -- concatenate last item onto end of <list> with final separator
end
return list;
end
 
Line 259 ⟶ 424:
]]
 
local function select_one ( args, aliases_list, error_condition, index )
local value = nil; -- the value assigned to the selected parameter
local selected = ''; -- the name of the parameter we have chosen
Line 266 ⟶ 431:
if index ~= nil then index = tostring(index); end
 
for _, alias in ipairs ( aliases_list ) do -- for each alias in the aliases list
if alias:match ('#') then -- if this alias can be enumerated
if '1' == index then -- when index is 1 test for enumerated and non-enumerated aliases
value, selected = is_alias_used (args, alias, index, false, value, selected, error_list); -- first test for non-enumerated alias
end
value, selected = is_alias_used (args, alias, index, true, value, selected, error_list); -- test for enumerated alias
else
value, selected = is_alias_used (args, alias, index, false, value, selected, error_list); -- test for non-enumerated alias
end
end
 
if #error_list > 0 and 'none' ~= error_condition then -- for cases where this code is used outside of extract_names()
for i, v in ipairs (error_list) do
local error_str = "";
error_list[i] = wrap_style ('parameter', v);
for _, k in ipairs( error_list ) do
if error_str ~= "" then error_str = error_str .. cfg.messages['parameter-separator'] end
error_str = error_str .. wrap_style ('parameter', k);
end
iftable.insert #(error_list, >wrap_style 1('parameter', thenselected));
set_message (error_condition, {make_sep_list (#error_list, error_list)});
error_str = error_str .. cfg.messages['parameter-final-separator'];
else
error_str = error_str .. cfg.messages['parameter-pair-separator'];
end
error_str = error_str .. wrap_style ('parameter', selected);
table.insert( z.message_tail, { set_error( error_condition, {error_str}, true ) } );
end
Line 302 ⟶ 460:
The str:gsub() returns either A|B froma [[A|B]] or B from [[B]] or B from B (no wikilink markup).
 
In l(), l:gsub() removes the link and pipe (if they exist); the second :gsub() trims white spacewhitespace from the label
if str was wrapped in wikilink markup. Presumably, this is because without wikimarkup in str, there is no match
in the initial gsub, the replacement function l() doesn't get called.
Line 309 ⟶ 467:
 
local function remove_wiki_link (str)
return (str:gsub( ("%[%[([^%[%]]*)%]%]", function(l)
return l:gsub( ("^[^|]*|(.*)$", "%1" ):gsub ("^%s*(.-)%s*$", "%1");
end));
end
Line 317 ⟶ 475:
--[=[-------------------------< I S _ W I K I L I N K >--------------------------------------------------------
 
Determines if str is a wikilink, extracts, and returns the the wikilink type, link text, and display text parts.
If str is a complex wikilink ([[L|D]]):
returns wl_type 2 and D and L from [[L|D]];
Line 325 ⟶ 483:
returns wl_type 0, str as D, and L as empty string.
 
trims leading and trailing white spacewhitespace and pipes from L and D ([[L|]] and [[|D]] are accepted by MediaWiki and
treated like [[D]]; while [[|D|]] is not accepted by MediaWiki, here, we accept it and return D without the pipes).
 
Line 335 ⟶ 493:
 
if not str:match ('^%[%[[^%]]+%]%]$') then -- is str some sort of a wikilink (must have some sort of content)
return 0, str, ''; -- not a wililinkwikilink; return wl_type as 0, str as D, and empty string as L
end
Line 346 ⟶ 504:
D = mw.text.trim (D, '%s|'); -- trim white space and pipe characters
-- L = L and mw.text.trim (L, '%s|');
return wl_type, D, L or '';
end
Line 374 ⟶ 531:
local flag;
while true do
if argument:find ( "'''''", 1, true ) then -- bold italic (5)
argument, flag = argument:gsub ("%'%'%'%'%'", ""); -- remove all instances of it
elseif argument:find ( "''''", 1, true ) then -- italic start and end without content (4)
argument, flag=argument:gsub ("%'%'%'%'", "");
elseif argument:find ( "'''", 1, true ) then -- bold (3)
argument, flag=argument:gsub ("%'%'%'", "");
elseif argument:find ( "''", 1, true ) then -- italic (2)
argument, flag = argument:gsub ("%'%'", "");
else
break;
Line 407 ⟶ 564:
 
return {
is_setadd_maint_cat = is_setadd_maint_cat, -- exported functions
add_prop_cat = add_prop_cat,
in_array = in_array,
substitute = substitute,
error_comment = error_comment,
has_accept_as_written = has_accept_as_written,
set_error = set_error,
hyphen_to_dash = hyphen_to_dash,
select_one = select_one,
in_array = in_array,
add_maint_cat = add_maint_cat,
is_set = is_set,
wrap_style = wrap_style,
safe_for_italics = safe_for_italics,
remove_wiki_link = remove_wiki_link,
is_wikilink = is_wikilink,
make_sep_list = make_sep_list,
make_wikilink = make_wikilink,
remove_wiki_link = remove_wiki_link,
safe_for_italics = safe_for_italics,
select_one = select_one,
set_message = set_message,
set_selected_modules = set_selected_modules,
strip_apostrophe_markup = strip_apostrophe_markup,
substitute = substitute,
wrap_style = wrap_style,
 
z = z, -- exported table
}
Retrieved from "https://en.seminaverbi.bibleget.io/wiki/Module:Citation/CS1/Utilities"