User:Quarl/___location canonicalize.js: Difference between revisions

Revision as of 08:17, 31 January 2006

Latest revision as of 06:06, 9 February 2006

(28 intermediate revisions by the same user not shown)

Line 1:

// [[User:Quarl/location_canonicalize.js]] - canonicalizes ___location WikiLinks

// as per [[Wikipedia:WikiProject Location Format]]

// Example: [[Seattle, Washington]] becomes [[Seattle, Washington|Seattle]], [[Washington]], [[USA]].

// requiresdepends: wikipage.js, util.js, addlilinkwikitabs.js, wikiedit.js, autoedit.js

// quarl 2006-01-22 initial version

// quarl 2006-02-08 refactored to autoedit.js

//<pre><nowiki>

location_canonicalizevar locz = new Objectautoedit();

'locz',

'LocZ', 'ca-locz', 'Canonicalize ___location wikilinks',

'Location canonicalization');

location_canonicalizelocz.runinitData = function() {

var CountryData = function(states, link_country, regexp_country) {

location_canonicalize.initData();

this.states = states;

wikiPage.getEditorAsync(location_canonicalize.edit);

this.link_country = link_country.match(/\[/) ? link_country : '[['+link_country+']]';

}

location_canonicalize.edit = function(editor) {

// make changes

var result = '';

var input = editor.wpTextbox1;

var changes = [];

// special case the first Infobox, if there is one

if (input.match(/^{{Infobox(?:.|\n)*?\n}}/i)) {

// var left = RegExp.leftContext;

var infobox = RegExp.lastMatch;

var right = RegExp.rightContext;

// treat the infobox separately, so that USA links get added to main

// article.

result = (location_canonicalize.canonicalizeString(infobox, changes) +

location_canonicalize.canonicalizeString(right, changes));

} else {

result = location_canonicalize.canonicalizeString(input, changes);

}

if (changes.length) {

editor.wpTextbox1 = result;

editor.wpSummary = '___location canonicalization: ' + changes.join('; ');

editor.wpMinoredit = true;

editor.submit('wpDiff');

} else {

alert("No changes to make!");

}

}

location_canonicalize.canonicalizeString = function(input, changes) {

var result = '';

while (input.match(/\[\[ *(?:([^|\]]+?) *\| *)?([^\]]+?) *\]\]/)) {

var left = RegExp.leftContext;

var wfull = RegExp.lastMatch;

var wlink = RegExp.$1;

var wtext = RegExp.$2;

var right = RegExp.rightContext;

result += left;

var r = location_canonicalize.wikilink((wlink||wtext), wtext, result, right);

if (r) {

var new_wfull = r.wfull;

result = r.left;

right = r.right;

changes.push(wfull + ' → ' + new_wfull);

} else {

var new_wfull = wfull;

}

result += new_wfull;

input = right;

}

result += input;

return result;

}

location_canonicalize.initData = function() {

if (this.init) return;

this.init = true;

var initRegexp = function(states, link_country, regexp_country) {

d = new Object();

d.states = states;

d.link_country = link_country.match(/\[/) ? link_country : '[['+link_country+']]';

var regexp_country = regexp_country || '\\[\\['+link_country+'\\]\\]';

dthis.regexp_country = new RegExp(regexp_country);

dthis.regexp_country_sq = new RegExp('^, *'+regexp_country);

dthis.regexp_substate = (

new RegExp('^([^,]+), *(' + dthis.states.join('|') + ')$'));

dthis.regexp_state = (

new RegExp('^(?:' + this.states.join('|') + ')$'));

}

this.countries = [

initRegexpnew CountryData( // USA

['Alabama', 'Alaska', 'Arizona', 'Arkansas', 'California', 'Colorado',

'Connecticut', 'Delaware', 'Florida', 'Georgia', 'Hawaii', 'Idaho',

Line 101 ⟶ 39:

'Oklahoma', 'Oregon', 'Pennsylvania', 'Rhode Island', 'South Carolina',

'South Dakota', 'Tennessee', 'Texas', 'Utah', 'Vermont', 'Virginia',

'Washington', 'West Virginia', 'Wisconsin', 'Wyoming' ],

'Washington, DC', 'Washington, D.C.' // not strictly a state, but needs to be qualified with country also

],

'[[United States|USA]]',

'\\[\\[(?:United[ _]States(?:[ _][^|\\\]]+?)?|USA)(?:\\|[^|\\\]]+?)?\\]\\]'),

initRegexpnew CountryData( // Canada

['British Columbia', 'Alberta', 'Saskatchewan', 'Manitoba',

'Ontario', 'Quebec', 'New Brunswick', 'Nova Scotia',

'Prince Edward Island', 'Newfoundland and Labrador'],

'Canada'),

new CountryData( // England

['Bedfordshire', 'Berkshire', 'City of Bristol',

'Buckinghamshire', 'Cambridgeshire', 'Cheshire',

'Cornwall', 'Cumbria', 'Derbyshire', 'Devon', 'Dorset',

'Durham', 'East Riding of Yorkshire', 'East Sussex', 'Essex',

'Gloucestershire', 'Greater London', 'Greater Manchester',

'Hampshire', 'Herefordshire', 'Hertfordshire', 'Isle of Wight',

'Kent', 'Lancashire', 'Leicestershire', 'Lincolnshire',

'City of London', 'Merseyside', 'Norfolk', 'Northamptonshire',

'Northumberland', 'North Yorkshire', 'Nottinghamshire',

'Oxfordshire', 'Rutland', 'Shropshire', 'Somerset',

'South Yorkshire', 'Staffordshire', 'Suffolk', 'Surrey',

'Tyne and Wear', 'Warwickshire', 'West Midlands', 'West Sussex',

'West Yorkshire', 'Wiltshire', 'Worcestershire'],

'England'),

];

}

location_canonicalizelocz.wikilinksplitText = function(wlink, wtext, left, rightinput) {

var inputs = [];

// special case for hat link, if there is one

if (input.match(/^: *''.*/)) {

var infobox = RegExp.lastMatch;

var right = RegExp.rightContext;

inputs.push(infobox);

input = right;

}

// special case the first Infobox, if there is one

if (input.match(/^(?:{{Infobox(?:.|\n)*?\n}}|{\|(?:.|\n)*?\n\|})/i)) {

// var left = RegExp.leftContext;

var infobox = RegExp.lastMatch;

var right = RegExp.rightContext;

// treat the infobox separately, so that USA links get added to main

// article.

inputs.push(infobox);

input = right;

}

inputs.push(input);

return inputs;

}

locz.buildRegExp = function() {

return /\[\[ *(?:([^|\]]+?) *\| *)?([^\]]+?) *\]\]/;

}

locz.replaceRegExp = function(d, m) {

var wlink = m[1] || m[2];

var wtext = m[2];

// non-main namespace - usually a category

if (wtext.match(/:/)) return null;

if (wlink != wtext) return;

Line 122 ⟶ 114:

var c = this.countries[i];

var changes = 0;

var wfull;

if (wtext.match(c.regexp_substate)) {

var city = RegExp.$1, state = RegExp.$2;

wfull = '[[' + wtext + '|' + city + ']]';

// only add link to state and country if we haven't yetlink mentionedit themyet.

if (d.left.match('\\[\\['+state+'\\]\\]')) {

wfull += ', ' + state;

} else {

wfull += ', [['+state+']]';

}

++changes;

} else if (wtext.match(c.regexp_state)) {

// state link -- just need to add country link as necessary

Line 139 ⟶ 133:

if (!wfull) continue;

if (d.left.match(c.regexp_country)) {

// Already mentioned country. Delete redundant subsequent

// country links

right =if (d.right.replacematch(c.regexp_country_sq, '');) {

d.right = RegExp.rightContext;

// only count as a change if we actually delete it!

++changes;

}

} else {

// Haven't mentioned country earlier

if (d.right.match(c.regexp_country_sq)) {

// it's thereright already,after the current link; good.

} else {

// not there; add it.

wfull += ', ' + c.link_country;

++changes;

}

}

if (changes) {

return { wfull: wfull, left: left, right: right};

d.text = wfull;

}

return;

}

return null;

}

location_canonicalizelocz.load_load = function() {

locz.qAutoEdit();

if (wikiPage.nsSpecialP) return;

locz.addTab();

addTab('javascript:location_canonicalize.run()', 'locz', 'ca-locz', 'Canonicalize ___location wikilinks');

}

addOnloadHook(location_canonicalizelocz.load_load);

//</nowiki></pre>