Bharatpedia

Module:Citation/CS1/Utilities: Difference between revisions

Module Talk

Module:Citation/CS1/Utilities (view source)

Revision as of 14:39, 30 September 2023

5,392 bytes added ,  30 September 2023
no edit summary
>Trappist the monk
(sync from sandbox;)
No edit summary
 
(2 intermediate revisions by the same user not shown)
Line 1: Line 1:
local z = {
local z = {
error_categories = {}; -- for categorizing citations that contain errors
error_cats_t = {}; -- for categorizing citations that contain errors
error_ids = {};
error_ids_t = {}; -- list of error identifiers; used to prevent duplication of certain errors; local to this module
message_tail = {};
error_msgs_t = {}; -- sequence table of error messages
maintenance_cats = {}; -- for categorizing citations that aren't erroneous per se, but could use a little work
maint_cats_t = {}; -- for categorizing citations that aren't erroneous per se, but could use a little work
properties_cats = {}; -- for categorizing citations based on certain properties, language of source for instance
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
};
};


Line 13: Line 13:


local cfg; -- table of tables imported from selected Module:Citation/CS1/Configuration
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
if haystack and type(haystack) == "table" then
    for n, v in ipairs(haystack) do
        if v == needle then
            return n;
        end
    end
end
return false;
end




Line 27: Line 59:


local function has_accept_as_written (str, allow_empty)
local function has_accept_as_written (str, allow_empty)
if not is_set (str) then
return str, false;
end
local count;
local count;
if true == allow_empty then
if true == allow_empty then
str, count = str:gsub ('^%(%((.*)%)%)$', '%1'); -- allows (()) to be an empty set
str, count = str:gsub ('^%(%((.*)%)%)$', '%1'); -- allows (()) to be an empty set
Line 37: Line 74:




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


Returns true if argument is set; false otherwise. Argument is 'set' when it exists (not nil) or when it is not an empty string.
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 is_set (var)
local function substitute (msg, args)
return not (var == nil or var == '');
return args and mw.message.newRawMessage (msg, args):plain() or msg;
end
end




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


Whether needle is in haystack
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 in_array (needle, haystack)
local function error_comment (content, hidden)
if needle == nil then
return substitute (hidden and cfg.presentation['hidden-error'] or cfg.presentation['visible-error'], content);
return false;
end
for n, v in ipairs (haystack) do
if v == needle then
return n;
end
end
return false;
end
end




--[[--------------------------< S U B S T I T U T E >----------------------------------------------------------
--[[--------------------------< 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.


Populates numbered arguments in a message string using an argument table.
str may be a comma- or semicolon-separated list


]]
]]


local function substitute (msg, args)
local function hyphen_to_dash (str)
return args and mw.message.newRawMessage (msg, args):plain() or msg;
if not is_set (str) then
end
return str;
end


local accept; -- boolean


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


Wraps error messages with CSS markup according to the state of hidden.
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
local function error_comment (content, hidden)
if accept then
return substitute (hidden and cfg.presentation['hidden-error'] or cfg.presentation['visible-error'], content);
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
end


Line 110: Line 179:
--[[--------------------------< S E T _ M E S S A G E >----------------------------------------------------------
--[[--------------------------< S E T _ M E S S A G E >----------------------------------------------------------


Sets an error condition and returns the appropriate error message. The actual placement of the error message in the output is
Sets an error message using the ~/Configuration error_conditions{} table along with arguments supplied in the function
the responsibility of the calling 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_categories and z.maintenance_cats to have the form cat_name = true; to avoid dups without having to have an extra cat
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.maintenance_cats; TODO: figure out how to delete this 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 function set_message (error_id, arguments, raw, prefix, suffix)
Line 129: Line 210:
elseif is_set (error_state.category) then
elseif is_set (error_state.category) then
if error_state.message then -- when error_state.message defined, this is an error message
if error_state.message then -- when error_state.message defined, this is an error message
table.insert (z.error_categories, error_state.category);
table.insert (z.error_cats_t, error_state.category);
else
else
if not added_maint_cats[error_id] then
if not added_maint_cats[error_id] then
added_maint_cats[error_id] = true; -- note that we've added this category
added_maint_cats[error_id] = true; -- note that we've added this category
table.insert (z.maintenance_cats, substitute (error_state.category, arguments)); -- make cat name then add to table
table.insert (z.maint_cats_t, substitute (error_state.category, arguments)); -- make cat name then add to table
end
end
return; -- because no message, nothing more to do
return; -- because no message, nothing more to do
Line 156: Line 237:
});
});


z.error_ids[error_id] = true;
z.error_ids_t[error_id] = true;
if z.error_ids['err_citation_missing_title'] and -- if missing-title error already noted
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
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
return '', false; -- don't bother because one flavor of missing title is sufficient
Line 164: Line 245:
message = table.concat ({prefix, message, suffix});
message = table.concat ({prefix, message, suffix});


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


return error_comment (message, error_state.hidden);
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
end


Line 199: Line 282:
end
end


if is_set (args[alias]) then -- alias is in the template's argument list
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
if value ~= nil and selected ~= alias then -- if we have already selected one of the aliases
local skip;
local skip;
Line 222: Line 305:
--[[--------------------------< A D D _ M A I N T _ C A T >------------------------------------------------------
--[[--------------------------< A D D _ M A I N T _ C A T >------------------------------------------------------


Adds a category to z.maintenance_cats using names from the configuration file with additional text if any.
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.maintenance_cats.
To prevent duplication, the added_maint_cats table lists the categories by key that have been added to z.maint_cats_t.


]]
]]
Line 230: Line 313:
if not added_maint_cats [key] then
if not added_maint_cats [key] then
added_maint_cats [key] = true; -- note that we've added this category
added_maint_cats [key] = true; -- note that we've added this category
table.insert (z.maintenance_cats, substitute (cfg.maint_cats [key], arguments)); -- make name then add to table
table.insert (z.maint_cats_t, substitute (cfg.maint_cats [key], arguments)); -- make name then add to table
end
end
end
end
Line 237: Line 320:
--[[--------------------------< A D D _ P R O P _ C A T >--------------------------------------------------------
--[[--------------------------< A D D _ P R O P _ C A T >--------------------------------------------------------


Adds a category to z.properties_cats using names from the configuration file with additional text if any.
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
foreign_lang_source and foreign_lang_source_2 keys have a language code appended to them so that multiple languages
Line 246: Line 329:
]]
]]


local added_prop_cats = {}; -- list of property categories that have been added to z.properties_cats
local added_prop_cats = {}; -- list of property categories that have been added to z.prop_cats_t


local function add_prop_cat (key, arguments)
local function add_prop_cat (key, arguments, key_modifier)
if not added_prop_cats [key] then
local key_modified = key .. ((key_modifier and key_modifier) or ''); -- modify <key> with <key_modifier> if present and not nil
added_prop_cats [key] = true; -- note that we've added this category
key = key:gsub ('(foreign_lang_source_?2?)%a%a%a?[%a%-]*', '%1'); -- strip lang code from keyname
if not added_prop_cats [key_modified] then
table.insert (z.properties_cats, substitute (cfg.prop_cats [key], arguments)); -- make name then add to table
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
end
end
Line 273: Line 358:
if str:sub (-1, -1) == "'" then str = str .. "<span></span>"; end
if str:sub (-1, -1) == "'" then str = str .. "<span></span>"; end
-- Remove newlines as they break italics.
return str:gsub ('\n', ' '); -- Remove newlines as they break italics.
return str:gsub ('\n', ' ');
 
end
end


Line 366: Line 449:
end
end
table.insert (error_list, wrap_style ('parameter', selected));
table.insert (error_list, wrap_style ('parameter', selected));
table.insert (z.message_tail, {set_message (error_condition, {make_sep_list (#error_list, error_list)}, true)});
set_message (error_condition, {make_sep_list (#error_list, error_list)});
end
end
Line 451: Line 534:
while true do
while true do
if argument:find ("'''''", 1, true) then -- bold italic (5)
if argument:find ("'''''", 1, true) then -- bold italic (5)
argument, flag = argument:gsub ("%'%'%'%'%'", ""); -- remove all instances of it
argument, flag = argument:gsub ("%'%'%'%'%'", ""); -- remove all instances of it
elseif argument:find ("''''", 1, true) then -- italic start and end without content (4)
elseif argument:find ("''''", 1, true) then -- italic start and end without content (4)
argument, flag=argument:gsub ("%'%'%'%'", "");
argument, flag=argument:gsub ("%'%'%'%'", "");
elseif argument:find ("'''", 1, true) then -- bold (3)
elseif argument:find ("'''", 1, true) then -- bold (3)
argument, flag=argument:gsub ("%'%'%'", "");
argument, flag=argument:gsub ("%'%'%'", "");
elseif argument:find ("''", 1, true) then -- italic (2)
elseif argument:find ("''", 1, true) then -- italic (2)
Line 487: Line 570:
error_comment = error_comment,
error_comment = error_comment,
has_accept_as_written = has_accept_as_written,
has_accept_as_written = has_accept_as_written,
hyphen_to_dash = hyphen_to_dash,
in_array = in_array,
in_array = in_array,
is_set = is_set,
is_set = is_set,
Retrieved from "https://en.bharatpedia.org/wiki/Special:MobileDiff/10354...431501"