Module:Citation/CS1/Utilities

local z = {
 error_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_msgs_t = {};               -- sequence table of error messages
 maint_cats_t = {};               -- for categorizing citations that aren't erroneous per se, but could use a little work
 prop_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
};


--[[--------------------------< F O R W A R D   D E C L A R A T I O N S >--------------------------------------
]]

local cfg;                  -- table of tables imported from selected Module:Citation/CS1/Configuration


--[[--------------------------< I S _ S E T >------------------------------------------------------------------

Returns true if argument is set; false otherwise. Argument is 'set' when it exists (not nil) or when it is not an empty string.

]]

local function is_set (var)
 return not (var == nil or var == '');
end


--[[--------------------------< I N _ A R R A Y >--------------------------------------------------------------

Whether needle is in haystack

]]

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;
  end
 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


--[[--------------------------< 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


--[[--------------------------< E R R O R _ C O M M E N T >----------------------------------------------------

Wraps error messages with CSS 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


--[=[-------------------------< M A K E _ W I K I L I N K >----------------------------------------------------

Makes a wikilink; when both 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 empty string.

]=]

local function make_wikilink (link, display)
 if not is_set (link) then return '' end

 if is_set (display) and link ~= display then   
  return table.concat ({'[[', link, '|', display, ']]'});   
 else
  return table.concat ({'[[', link, ']]'});
 end
end


--[[--------------------------< S E T _ M E S S A G E >----------------------------------------------------------

Sets an error message using the ~/Configuration error_conditions{} table along with arguments supplied in the function
call, inserts the resulting message in z.error_msgs_t{} sequence table, and returns the error message.

<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_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
  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 (
  {
  message,
  ' (',
  make_wikilink (
   table.concat (
    {
    cfg.messages['help page link'],
    '#',
    error_state.anchor
    }),
   cfg.messages['help page label']),
  ')'
  });

 z.error_ids_t[error_id] = true;
 if z.error_ids_t['err_citation_missing_title'] and       -- if missing-title error already noted
  in_array (error_id, {'err_bare_url_missing_title', 'err_trans_missing_title'}) then  -- and this error is one of these
   return '', false;             -- don't bother because one flavor of missing title is sufficient
 end
 
 message = table.concat ({prefix, message, suffix});

 if true == raw then
  return message, error_state.hidden;          -- return message not wrapped in visible-error, hidden-error span tag
 end  

 message = 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


--[[-------------------------< I S _ A L I A S _ U S E D >-----------------------------------------------------

This function is used by select_one() to determine if one of a list of alias parameters is in the argument list
provided by the template.

Input:
 args – pointer to the arguments table from calling template
 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
 error_list – list of aliases that are duplicates of the alias already selected

Returns:
 value – value associated with alias we selected or that was previously selected or nil if an alias not yet selected
 selected – the alias we selected or the alias that was previously selected or nil if an alias not yet selected

]]

local function is_alias_used (args, alias, index, enumerated, value, selected, error_list)
 if enumerated then               -- is this a test for an enumerated parameters?
  alias = alias:gsub ('#', index);          -- replace '#' with the value in index
 else
  alias = alias:gsub ('#', '');           -- remove '#' if it exists
 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;
     break;              -- has been added so stop looking 
    end
   end
   if not skip then             -- has not been added so
    table.insert (error_list, alias);        -- add error alias to the error list
   end
  else
   value = args[alias];            -- not yet selected an alias, so select this one
   selected = alias;
  end
 end
 return value, selected;              -- return newly selected alias, or previously selected alias
end


--[[--------------------------< A D D _ M A I N T _ C A T >------------------------------------------------------

Adds a category to z.maint_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.maint_cats_t.

]]

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.maint_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


--[[--------------------------< S A F E _ F O R _ I T A L I C S >----------------------------------------------

Protects a string that will be wrapped in wiki italic markup '' ... ''

Note: We cannot use <i> for italics, as the expected behavior for italics specified by ''...'' in the title is that
they will be inverted (i.e. unitalicized) in the resulting references.  In addition, <i> and '' tend to interact
poorly under Mediawiki's HTML tidy.

]]

local function safe_for_italics (str)
 if not is_set (str) then return str end

 if str:sub (1, 1) == "'" then 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.
end


--[[--------------------------< W R A P _ S T Y L E >----------------------------------------------------------

Applies styling to various parameters.  Supplied string is wrapped using a message_list configuration taking one
argument; protects italic styled parameters.  Additional text taken from citation_config.presentation - the reason
this function is similar to but separate from wrap_msg().

]]

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


--[[--------------------------< S E L E C T _ O N E >----------------------------------------------------------

Chooses one matching parameter from a list of parameters to consider.  The list of parameters to consider is just
names.  For parameters that may be enumerated, the position of the numerator in the parameter name is identified
by the '#' so |author-last1= and |author1-last= are represented as 'author-last#' and 'author#-last'.

Because enumerated parameter |<param>1= is an alias of |<param>= we must test for both possibilities.


Generates an error if more than one match is present.

]]

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
 local error_list = {};

 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
   error_list[i] = wrap_style ('parameter', v);
  end
  table.insert (error_list, wrap_style ('parameter', selected));
  set_message (error_condition, {make_sep_list (#error_list, error_list)});
 end
 
 return value, selected;
end


--[=[-------------------------< R E M O V E _ W I K I _ L I N K >----------------------------------------------

Gets the display text from a wikilink like [[A|B]] or [[B]] gives B

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 whitespace 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.

]=]

local function remove_wiki_link (str)
 return (str:gsub ("%[%[([^%[%]]*)%]%]", function(l)
  return l:gsub ("^[^|]*|(.*)$", "%1" ):gsub ("^%s*(.-)%s*$", "%1");
 end));
end


--[=[-------------------------< I S _ W I K I L I N K >--------------------------------------------------------

Determines if str is a wikilink, extracts, and returns 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]];
if str is a simple wikilink ([[D]])
 returns wl_type 1 and D from [[D]] and L as empty string;
if not a wikilink:
 returns wl_type 0, str as D, and L as empty string.

trims leading and trailing whitespace 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).

]=]

local function is_wikilink (str)
 local D, L
 local wl_type = 2;               -- assume that str is a complex wikilink [[L|D]]

 if not str:match ('^%[%[[^%]]+%]%]$') then         -- is str some sort of a wikilink (must have some sort of content)
  return 0, str, '';              -- not a wikilink; return wl_type as 0, str as D, and empty string as L
 end
 
 L, D = str:match ('^%[%[([^|]+)|([^%]]+)%]%]$');       -- get L and D from [[L|D]] 

 if not is_set (D) then              -- if no separate display
  D = str:match ('^%[%[([^%]]*)|*%]%]$');         -- get D from [[D]] or [[D|]]
  wl_type = 1; 
 end
 
 D = mw.text.trim (D, '%s|');            -- trim white space and pipe characters 
 return wl_type, D, L or '';
end


--[[--------------------------< S T R I P _ A P O S T R O P H E _ M A R K U P >--------------------------------

Strip wiki italic and bold markup from argument so that it doesn't contaminate COinS metadata.
This function strips common patterns of apostrophe markup.  We presume that editors who have taken the time to
markup a title have, as a result, provided valid markup. When they don't, some single apostrophes are left behind.

Returns the argument without wiki markup and a number; the number is more-or-less meaningless except as a flag
to indicate that markup was replaced; do not rely on it as an indicator of how many of any kind of markup was
removed; returns the argument and nil when no markup removed

]]

local function strip_apostrophe_markup (argument)
 if not is_set (argument) then
  return argument, nil;             -- no argument, nothing to do
 end

 if nil == argument:find ( "''", 1, true ) then        -- Is there at least one double apostrophe?  If not, exit.
  return argument, nil;
 end

 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;
  end
 end

 return argument, flag;              -- done
end


--[[--------------------------< S E T _ S E L E C T E D _ M O D U L E S >--------------------------------------

Sets local cfg table to same (live or sandbox) as that used by the other modules.

]]

local function set_selected_modules (cfg_table_ptr)
 cfg = cfg_table_ptr;
 
end


--[[--------------------------< E X P O R T S >----------------------------------------------------------------
]]

return {
 add_maint_cat = add_maint_cat,            -- exported functions
 add_prop_cat = add_prop_cat,
 error_comment = error_comment,
 has_accept_as_written = has_accept_as_written,
 hyphen_to_dash = hyphen_to_dash,
 in_array = in_array,
 is_set = is_set,
 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
 }