Config files changes - move more logic from scripts to config
(re: #4733)
This commit is contained in:
Bryan Housel
parent
87b45682fb
commit
ef3f1e74c4
12 changed files with 454 additions and 395 deletions
|
|
@ -1,30 +0,0 @@
|
|||
{
|
||||
"keepTags": [
|
||||
"^amenity/bicycle_rental$",
|
||||
"^amenity/car_sharing$",
|
||||
"^amenity/charging_station$",
|
||||
"^amenity/fire_station$",
|
||||
"^amenity/police$",
|
||||
"^amenity/post_(depot|office)$",
|
||||
"^emergency/ambulance_station$"
|
||||
],
|
||||
"discardKeys": ["^amenity/bicycle_rental\\|nextbike( gmbh)?$"],
|
||||
"discardNames": [
|
||||
"^(buzón\\sde\\s)?correos$",
|
||||
"^agencja pocztowa$",
|
||||
"^bureau de poste$",
|
||||
"^correo$",
|
||||
"^government$",
|
||||
"^mail(box)?$",
|
||||
"^pa(r|cz)komat$",
|
||||
"^pac?k(etbox|station)$",
|
||||
"^po(s|š)t(a|e|o)$",
|
||||
"^post\\s?(agentur|filiale|partner|stelle)?$",
|
||||
"^ufficio postale$",
|
||||
"^urząd pocztowy$",
|
||||
"^главпочтамт$",
|
||||
"^по(чт|шт|щ)а$",
|
||||
"^почтовое отделение$",
|
||||
"^郵局$"
|
||||
]
|
||||
}
|
||||
|
|
@ -1,18 +0,0 @@
|
|||
{
|
||||
"keepTags": [
|
||||
"^route/(aerialway|bus|light_rail|subway|train|tram|trolleybus|walking_bus)$"
|
||||
],
|
||||
"discardKeys": [],
|
||||
"discardNames": [
|
||||
"^[ilnr][chimsw][nm]$",
|
||||
"^city$",
|
||||
"^international$",
|
||||
"^local$",
|
||||
"^mtb$",
|
||||
"^municipal$",
|
||||
"^national$",
|
||||
"^rail$",
|
||||
"^regional$",
|
||||
"^urban$"
|
||||
]
|
||||
}
|
||||
|
|
@ -1,130 +1,21 @@
|
|||
{
|
||||
"keepTags": [
|
||||
"^amenity/bank$",
|
||||
"^amenity/bureau_de_change$",
|
||||
"^amenity/cafe$",
|
||||
"^amenity/car_rental$",
|
||||
"^amenity/cinema$",
|
||||
"^amenity/clinic$",
|
||||
"^amenity/dentist$",
|
||||
"^amenity/fast_food$",
|
||||
"^amenity/fuel$",
|
||||
"^amenity/hospital$",
|
||||
"^amenity/ice_cream$",
|
||||
"^amenity/money_transfer$",
|
||||
"^amenity/payment_(centre|terminal)$",
|
||||
"^amenity/pharmacy$",
|
||||
"^amenity/public_bookcase$",
|
||||
"^amenity/restaurant$",
|
||||
"^amenity/social_(centre|facility)$",
|
||||
"^amenity/vending_machine$",
|
||||
"^amenity/veterinary$",
|
||||
"^healthcare/audiologist$",
|
||||
"^healthcare/blood_donation$",
|
||||
"^healthcare/counselling$",
|
||||
"^healthcare/laboratory$",
|
||||
"^healthcare/physiotherapist$",
|
||||
"^leisure/adult_gaming_centre$",
|
||||
"^leisure/amusement_arcade$",
|
||||
"^leisure/bowling_alley$",
|
||||
"^leisure/escape_game$",
|
||||
"^leisure/fitness_centre$",
|
||||
"^office/accountant$",
|
||||
"^office/employment_agency$",
|
||||
"^office/estate_agent$",
|
||||
"^office/financial(_advisor)?$",
|
||||
"^office/insurance$",
|
||||
"^office/tax_advisor$",
|
||||
"^office/telecommunication$",
|
||||
"^shop/(?!(yes|no|banner_shop|gas|convenience;gas|laundry|lottery|mall|vacant))",
|
||||
"^tourism/(h|m)otel$"
|
||||
],
|
||||
"discardKeys": [
|
||||
"^amenity/bank\\|(bank zachodni|bz) wbk$",
|
||||
"^amenity/bank\\|(landesbau|bau|kreis|stadt)?sparkasse.*$",
|
||||
"^amenity/bank\\|bank spółdzielczy$",
|
||||
"^amenity/bank\\|lbs$",
|
||||
"^amenity/bank\\|pko$",
|
||||
"^amenity/bank\\|volks- und raiffeisenbank$",
|
||||
"^amenity/bank\\|volksbank$",
|
||||
"^amenity/bank\\|volksbank( |-)raiffeisenbank$",
|
||||
"^amenity/bank\\|vr( |-)bank$",
|
||||
"^amenity/bar\\|уют$",
|
||||
"^amenity/cafe\\|((eiscaf(e|é)\\s)?(dolomiti|venezia)|kiosque(\\sà)?\\scafé|caf(e|é)\\sexpress|встреча|рандеву|чайхана|жемчужина|виктория|бер(е|ё)зка|лакомка|мечта|оазис|парус|сказка||||||)$",
|
||||
"^amenity/cinema\\|октябрь$",
|
||||
"^amenity/clinic\\|(clinic|health center|health sub center|urgent care|consultorio|фап)$",
|
||||
"^amenity/dentist\\|дантист$",
|
||||
"^amenity/fast_food\\|(ali\\sbaba|antalya|asia(\\s|-)(bistro|imbiss|wok)|(berlin|city)\\sdöner|city\\s(grill|pizza)|kfc/taco\\sbell)$",
|
||||
"^amenity/fast_food\\|(marmaris|pizza\\s(house|time))$",
|
||||
"^amenity/fuel\\|gnv|gpl|posto|spbu|جایگاه\\scng$",
|
||||
"^amenity/hospital\\|(clinic|health center|(government )?hospital( municipal)?)$",
|
||||
"^amenity/hospital\\|(центральная районная )?больница$",
|
||||
"^amenity/hospital\\|фап$",
|
||||
"^amenity/hospital\\|центральна районна лікарня$",
|
||||
"^amenity/hospital\\|црб$",
|
||||
"^amenity/ice_cream\\|gelateria|venezia$",
|
||||
"^amenity/pharmacy\\|(adler|bahnhof|brunnen|burg|bären|einhorn|engel|hirsch|hubertus|kur|linden|löwen)(-| )?apotheke$",
|
||||
"^amenity/pharmacy\\|(centrum|farm(a|á)cia\\s(central(e)?|comunale))$",
|
||||
"^amenity/pharmacy\\|(marien|markt|mohren|neue|park|rathaus|rats|rosen|schloss|sonnen|stadt|stern)(-| )?apotheke$",
|
||||
"^amenity/pharmacy\\|(айболит|арніка|фармация|фармація)$",
|
||||
"^amenity/pharmacy\\|pharmacie\\s(centrale|comunale|de\\s(l'hôtel\\sde\\sville|la\\s(gare|mairie|poste))|du\\s(centre|marché|parc))$",
|
||||
"^amenity/restaurant\\|((gasthaus|gasthof|schwarzer)?\\s?(adler|löwen))$",
|
||||
"^amenity/restaurant\\|(ali\\sbaba|adria|akropolis|amici|aroma|asia|athen|athos|bahnhof|bamboo|bären|bella\\s(vista|napoli))",
|
||||
"^amenity/restaurant\\|(belvedere|brasserie|canteen|cantina|capri|carpe\\sdiem|casa\\s?(blanca|mia)|casino)",
|
||||
"^amenity/restaurant\\|(china\\s(buffet|city|garden|house|king|palace|star|town|wok)|(da|la|las|el)\\s)",
|
||||
"^amenity/restaurant\\|(delphi|deutsches\\shaus|dionysos|dolce\\svita|dorfkrug|europa|frohsinn)",
|
||||
"^amenity/restaurant\\|(golden\\sdragon|great\\swall)$",
|
||||
"^amenity/restaurant\\|(grüner\\sbaum|hellas|hirsch(en)?|hong\\skong|happy garden)",
|
||||
"^amenity/restaurant\\|(indian\\spalace|ichiban|il\\sgiardino|italia|jade\\sgarden|jägerhof)",
|
||||
"^amenity/restaurant\\|(kirchenwirt|koliba|kreta|kreuz|krone|lamm|le\\s(bistrot|patio))",
|
||||
"^amenity/restaurant\\|(l'escale|lemongrass|linde|lindenhof|little\\s(italy|india)|los\\s(amigos|arcos))",
|
||||
"^amenity/restaurant\\|(lotus|löwen|maharaja|mamma\\smia|mandarin|marco\\spolo|marmaris|mediterraneo|mensa|milano|mykonos|mythos)",
|
||||
"^amenity/restaurant\\|(namaste|napoli|new\\schina|oasis|ochsen|olympia|osaka|paradise|parrilla|peking|pho|pinocchio)",
|
||||
"^amenity/restaurant\\|(pizza\\shouse|pizz(a|e)ria|pizza\\sking|pomodoro|portofino|poseidon|ratskeller|rhodos|ristorante\\s)",
|
||||
"^amenity/restaurant\\|(roma|rose|rössl(e|i)|saffron|saigon|sakura|san\\smarco|santorini|schützenhaus|shalimar|shanghai|sonne|sphinx)",
|
||||
"^amenity/restaurant\\|(sportheim|sternen|syrtaki|taj\\smahal|taste\\sof\\sindia|the\\skitchen|toscana|tokyo|traube)",
|
||||
"^amenity/restaurant\\|(venezia|waldschänke|wok|zorbas|(zum|zur)\\s)",
|
||||
"^amenity/restaurant\\|don camillo",
|
||||
"^amenity/restaurant\\|永和豆漿$",
|
||||
"^amenity/social_facility\\|dom pomocy społecznej$",
|
||||
"^amenity/social_facility\\|maison de retraite$",
|
||||
"^amenity/social_facility\\|safe haven$",
|
||||
"^amenity/vending_machine\\|(schlauchomat|tobaccoland)$",
|
||||
"^shop/(convenience|kiosk)\\|(777|bp|esso|exxon|mobil|shell)$",
|
||||
"^shop/bakery\\|(backshop|schneider|schäfer|каравай)$",
|
||||
"^shop/bakery\\|(bäckerei\\s)?müller",
|
||||
"^shop/butcher\\|butcher$",
|
||||
"^shop/clothes\\|трикотаж$",
|
||||
"^shop/convenience\\|(24 часа|августина|альянс|апельсин|ассорти|бер(е|ё)зка|бриз|валентина|весна|визит|ветеран|водолей|гастроном|гермес|гурман|диана|добрый|дружба|домашний|елена)$",
|
||||
"^shop/convenience\\|(almac[eé]n|despensa)$",
|
||||
"^shop/convenience\\|(boutique d'alimentation|sup[eé]rette)$",
|
||||
"^shop/convenience\\|(milk bar|юлия)$",
|
||||
"^shop/convenience\\|(калинка|каштан|квартал|колобок|колос(ок)?|каравай|кооператор|корзинка|кулинария|лакомка|ласточка|лидер|луч|любава|любимый|людмила|люкс)$",
|
||||
"^shop/convenience\\|(меркурий|мечта|мини маркет|мираж|белорусские продукты|продукты 24|натали|надежда|ника)$",
|
||||
"^shop/convenience\\|(оазис|околица|олимп|парус|пеликан|престиж|подсолнух|радуга|росинка|ромашка|родник|рассвет|русь)$",
|
||||
"^shop/convenience\\|(сельпо|семейный|сказка|солнышко|спутник|солнечный|светлана|удача|ивушка)$",
|
||||
"^shop/convenience\\|(юбилейный|южный|эконом|шанс|чайка|центральный|хороший|хуторок|фаворит|фортуна|татьяна|тройка|эдем|уют)$",
|
||||
"^shop/copyshop\\|xerox$",
|
||||
"^shop/cosmetics\\|(vente de )?cosm[eé]tiques?$",
|
||||
"^shop/fishing\\|рыболов$",
|
||||
"^shop/florist\\|(мир цветов|орхидея)$",
|
||||
"^shop/frame\\|rumah\\spenduduk$",
|
||||
"^shop/hairdresser\\|(figaro|haarwerk|стиль|локон|шарм|fodrászat)$",
|
||||
"^shop/hardware\\|(всё для дома|1000 мелочей|хозяюшка)$",
|
||||
"^shop/jewelry\\|585$",
|
||||
"^shop/jewelry\\|jubiler$",
|
||||
"^shop/money_transfer|mobil(e)?\\smoney$",
|
||||
"^shop/null|null$",
|
||||
"^shop/optician\\|(die brille|vision plus|оптика|[oó]ptica)$",
|
||||
"^shop/supermarket\\|(famiglia cooperativa|семейный)$",
|
||||
"^shop/tailor\\|couture\\s?(dame|homme|mixte)?$",
|
||||
"^shop/ticket\\|ticket office$",
|
||||
"^tourism/(h|m)otel\\|budget inn$",
|
||||
"^tourism/hotel\\|(h(o|ô)tel\\s)?((zur\\s)?post|de france|de la gare|строитель)$",
|
||||
"^tourism/hotel\\|(h(o|ô)tel\\s)?(astoria|belvedere|central|city|continental|europa)\\s?(h(o|ô)tel)?$",
|
||||
"^tourism/hotel\\|(h(o|ô)tel\\s)?(grand|imperial|krone|palace|panorama|park|plaza|royal|victoria)\\s?(h(o|ô)tel)?$"
|
||||
],
|
||||
"discardNames": [
|
||||
"genericWords": [
|
||||
"^(buzón\\sde\\s)?correos$",
|
||||
"^agencja pocztowa$",
|
||||
"^bureau de poste$",
|
||||
"^correo$",
|
||||
"^government$",
|
||||
"^mail(box)?$",
|
||||
"^pa(r|cz)komat$",
|
||||
"^pac?k(etbox|station)$",
|
||||
"^po(s|š)t(a|e|o)$",
|
||||
"^post\\s?(agentur|filiale|partner|stelle)?$",
|
||||
"^ufficio postale$",
|
||||
"^urząd pocztowy$",
|
||||
"^главпочтамт$",
|
||||
"^по(чт|шт|щ)а$",
|
||||
"^почтовое отделение$",
|
||||
"^郵局$",
|
||||
"^((drummed|independent) )?(cng|filling|fuel|gas|lpg|petrol)(\\s(station|pump))?$",
|
||||
"^((le|the) )?salo(o)?n( de)?( coiffure| fryzjerski| beauté)?( homme| dame)?$",
|
||||
"^((горячий|свежий) )?хлеб(ница)?$",
|
||||
|
|
@ -166,6 +57,7 @@
|
|||
"^atelier de couture$",
|
||||
"^atm$",
|
||||
"^auto serwis$",
|
||||
"^automat na jizdenky$"
|
||||
"^bakery$",
|
||||
"^bakkal$",
|
||||
"^bank$",
|
||||
|
|
@ -199,6 +91,7 @@
|
|||
"^building$",
|
||||
"^bäckerei$",
|
||||
"^bücherschrank$",
|
||||
"^butcher$",
|
||||
"^büfé$",
|
||||
"^cabinet dentaire$",
|
||||
"^caf(e|é) de(s)? (la )?(place|sports|gare|mairie|poste|paris)$",
|
||||
|
|
@ -318,6 +211,7 @@
|
|||
"^magazin",
|
||||
"^maison$",
|
||||
"^market$",
|
||||
"^null$"
|
||||
"^mercado( municipal)?$",
|
||||
"^mercerie$",
|
||||
"^mini\\s?(abc|market|mart|mercado|pump)",
|
||||
|
|
@ -406,6 +300,7 @@
|
|||
"^tacos$",
|
||||
"^tailleur$",
|
||||
"^tailor$",
|
||||
"^ticket office$",
|
||||
"^taller( mec(á|a)nico)?$",
|
||||
"^tanatorio$",
|
||||
"^tankstelle$",
|
||||
|
|
@ -568,7 +463,6 @@
|
|||
"^自販機$",
|
||||
"^郵便ポスト$",
|
||||
"^銀行$",
|
||||
"^食堂$",
|
||||
"automat na jizdenky"
|
||||
"^食堂$"
|
||||
]
|
||||
}
|
||||
|
|
@ -1,57 +1,59 @@
|
|||
{
|
||||
"Q945463": {
|
||||
"note": "Bonita - https://github.com/osmlab/name-suggestion-index/issues/2652",
|
||||
"wikidata": "Q892598",
|
||||
"wikipedia": "de:Bonita (Unternehmen)"
|
||||
},
|
||||
"Q7319360": {
|
||||
"note": "Rexall - https://github.com/osmlab/name-suggestion-index/issues/4669",
|
||||
"wikidata": "Q39045377",
|
||||
"wikipedia": "en:Rexall (Canada)"
|
||||
},
|
||||
"Q62663379": {
|
||||
"note": "Unichem - Wikidata redirect",
|
||||
"wikidata": "Q7884722",
|
||||
"wikipedia": "en:Unichem (NZ)"
|
||||
},
|
||||
"Q6117880": {
|
||||
"note": "Jackson's Stores - https://github.com/osmlab/name-suggestion-index/issues/2791",
|
||||
"wikidata": "Q64617393",
|
||||
"wikipedia": ""
|
||||
},
|
||||
"Q56599145": {
|
||||
"note": "Steers - https://github.com/osmlab/name-suggestion-index/issues/3154",
|
||||
"wikidata": "Q3056765",
|
||||
"wikipedia": "en:Steers"
|
||||
},
|
||||
"Q466043": {
|
||||
"note": "Api - https://github.com/osmlab/name-suggestion-index/issues/4196",
|
||||
"wikidata": "Q646807",
|
||||
"wikipedia": "it:Italiana Petroli (azienda 2019)"
|
||||
},
|
||||
"Q21527177": {
|
||||
"note": "CPC Corporation - Wikidata redirect",
|
||||
"wikidata": "Q698024",
|
||||
"wikipedia": "en:CPC Corporation"
|
||||
},
|
||||
"Q28729658": {
|
||||
"note": "Colin's Jeans - https://github.com/osmlab/name-suggestion-index/issues/3154",
|
||||
"wikidata": "Q18015543",
|
||||
"wikipedia": "be:Colin’s"
|
||||
},
|
||||
"Q18720442": {
|
||||
"note": "But - Wikidata redirect",
|
||||
"wikidata": "Q2877537",
|
||||
"wikipedia": "fr:But (entreprise)"
|
||||
},
|
||||
"Q1126518": {
|
||||
"note": "ConocoPhillips - https://github.com/osmlab/name-suggestion-index/issues/3154",
|
||||
"wikidata": "Q214763",
|
||||
"wikipedia": "en:ConocoPhillips"
|
||||
},
|
||||
"Q1296860": {
|
||||
"note": "Gulf Oil LP - https://github.com/osmlab/name-suggestion-index/issues/3154",
|
||||
"wikidata": "Q5617505",
|
||||
"wikipedia": "en:Gulf Oil LP"
|
||||
"replacements": {
|
||||
"Q945463": {
|
||||
"note": "Bonita - https://github.com/osmlab/name-suggestion-index/issues/2652",
|
||||
"wikidata": "Q892598",
|
||||
"wikipedia": "de:Bonita (Unternehmen)"
|
||||
},
|
||||
"Q7319360": {
|
||||
"note": "Rexall - https://github.com/osmlab/name-suggestion-index/issues/4669",
|
||||
"wikidata": "Q39045377",
|
||||
"wikipedia": "en:Rexall (Canada)"
|
||||
},
|
||||
"Q62663379": {
|
||||
"note": "Unichem - Wikidata redirect",
|
||||
"wikidata": "Q7884722",
|
||||
"wikipedia": "en:Unichem (NZ)"
|
||||
},
|
||||
"Q6117880": {
|
||||
"note": "Jackson's Stores - https://github.com/osmlab/name-suggestion-index/issues/2791",
|
||||
"wikidata": "Q64617393",
|
||||
"wikipedia": ""
|
||||
},
|
||||
"Q56599145": {
|
||||
"note": "Steers - https://github.com/osmlab/name-suggestion-index/issues/3154",
|
||||
"wikidata": "Q3056765",
|
||||
"wikipedia": "en:Steers"
|
||||
},
|
||||
"Q466043": {
|
||||
"note": "Api - https://github.com/osmlab/name-suggestion-index/issues/4196",
|
||||
"wikidata": "Q646807",
|
||||
"wikipedia": "it:Italiana Petroli (azienda 2019)"
|
||||
},
|
||||
"Q21527177": {
|
||||
"note": "CPC Corporation - Wikidata redirect",
|
||||
"wikidata": "Q698024",
|
||||
"wikipedia": "en:CPC Corporation"
|
||||
},
|
||||
"Q28729658": {
|
||||
"note": "Colin's Jeans - https://github.com/osmlab/name-suggestion-index/issues/3154",
|
||||
"wikidata": "Q18015543",
|
||||
"wikipedia": "be:Colin’s"
|
||||
},
|
||||
"Q18720442": {
|
||||
"note": "But - Wikidata redirect",
|
||||
"wikidata": "Q2877537",
|
||||
"wikipedia": "fr:But (entreprise)"
|
||||
},
|
||||
"Q1126518": {
|
||||
"note": "ConocoPhillips - https://github.com/osmlab/name-suggestion-index/issues/3154",
|
||||
"wikidata": "Q214763",
|
||||
"wikipedia": "en:ConocoPhillips"
|
||||
},
|
||||
"Q1296860": {
|
||||
"note": "Gulf Oil LP - https://github.com/osmlab/name-suggestion-index/issues/3154",
|
||||
"wikidata": "Q5617505",
|
||||
"wikipedia": "en:Gulf Oil LP"
|
||||
}
|
||||
}
|
||||
}
|
||||
165
config/trees.json
Normal file
165
config/trees.json
Normal file
|
|
@ -0,0 +1,165 @@
|
|||
{
|
||||
"trees": {
|
||||
|
||||
"brands": {
|
||||
"emoji": "🍔",
|
||||
"mainTag": "brand:wikidata",
|
||||
"subclassOf": "Q4830453",
|
||||
"keepKV": [
|
||||
"^amenity/bank$",
|
||||
"^amenity/bureau_de_change$",
|
||||
"^amenity/cafe$",
|
||||
"^amenity/car_rental$",
|
||||
"^amenity/cinema$",
|
||||
"^amenity/clinic$",
|
||||
"^amenity/dentist$",
|
||||
"^amenity/fast_food$",
|
||||
"^amenity/fuel$",
|
||||
"^amenity/hospital$",
|
||||
"^amenity/ice_cream$",
|
||||
"^amenity/money_transfer$",
|
||||
"^amenity/payment_(centre|terminal)$",
|
||||
"^amenity/pharmacy$",
|
||||
"^amenity/public_bookcase$",
|
||||
"^amenity/restaurant$",
|
||||
"^amenity/social_(centre|facility)$",
|
||||
"^amenity/vending_machine$",
|
||||
"^amenity/veterinary$",
|
||||
"^healthcare/audiologist$",
|
||||
"^healthcare/blood_donation$",
|
||||
"^healthcare/counselling$",
|
||||
"^healthcare/laboratory$",
|
||||
"^healthcare/physiotherapist$",
|
||||
"^leisure/adult_gaming_centre$",
|
||||
"^leisure/amusement_arcade$",
|
||||
"^leisure/bowling_alley$",
|
||||
"^leisure/escape_game$",
|
||||
"^leisure/fitness_centre$",
|
||||
"^office/accountant$",
|
||||
"^office/employment_agency$",
|
||||
"^office/estate_agent$",
|
||||
"^office/financial(_advisor)?$",
|
||||
"^office/insurance$",
|
||||
"^office/tax_advisor$",
|
||||
"^office/telecommunication$",
|
||||
"^shop/(?!(yes|no|banner_shop|gas|convenience;gas|laundry|lottery|mall|vacant))",
|
||||
"^tourism/(h|m)otel$"
|
||||
],
|
||||
"discardKVN": [
|
||||
"^amenity/bank\\|(bank zachodni|bz) wbk$",
|
||||
"^amenity/bank\\|(landesbau|bau|kreis|stadt)?sparkasse.*$",
|
||||
"^amenity/bank\\|(bank spółdzielczy|lbs|pko)$",
|
||||
"^amenity/bank\\|volks- und raiffeisenbank$",
|
||||
"^amenity/bank\\|volksbank$",
|
||||
"^amenity/bank\\|volksbank( |-)raiffeisenbank$",
|
||||
"^amenity/bank\\|vr( |-)bank$",
|
||||
"^amenity/bar\\|уют$",
|
||||
"^amenity/cafe\\|((eiscaf(e|é)\\s)?(dolomiti|venezia)|kiosque(\\sà)?\\scafé|caf(e|é)\\sexpress|встреча|рандеву|чайхана|жемчужина|виктория|бер(е|ё)зка|лакомка|мечта|оазис|парус|сказка||||||)$",
|
||||
"^amenity/cinema\\|октябрь$",
|
||||
"^amenity/clinic\\|(clinic|health center|health sub center|urgent care|consultorio|фап)$",
|
||||
"^amenity/dentist\\|дантист$",
|
||||
"^amenity/fast_food\\|(ali\\sbaba|antalya|asia(\\s|-)(bistro|imbiss|wok)|(berlin|city)\\sdöner|city\\s(grill|pizza)|kfc/taco\\sbell)$",
|
||||
"^amenity/fast_food\\|(marmaris|pizza\\s(house|time))$",
|
||||
"^amenity/fuel\\|gnv|gpl|posto|spbu|جایگاه\\scng$",
|
||||
"^amenity/hospital\\|(clinic|health center|(government )?hospital( municipal)?)$",
|
||||
"^amenity/hospital\\|(центральная районная )?больница$",
|
||||
"^amenity/hospital\\|(фап|центральна районна лікарня|црб)$",
|
||||
"^amenity/ice_cream\\|gelateria|venezia$",
|
||||
"^amenity/pharmacy\\|(adler|bahnhof|brunnen|burg|bären|einhorn|engel|hirsch|hubertus|kur|linden|löwen)(-| )?apotheke$",
|
||||
"^amenity/pharmacy\\|(centrum|farm(a|á)cia\\s(central(e)?|comunale))$",
|
||||
"^amenity/pharmacy\\|(marien|markt|mohren|neue|park|rathaus|rats|rosen|schloss|sonnen|stadt|stern)(-| )?apotheke$",
|
||||
"^amenity/pharmacy\\|(айболит|арніка|фармация|фармація)$",
|
||||
"^amenity/pharmacy\\|pharmacie\\s(centrale|comunale|de\\s(l'hôtel\\sde\\sville|la\\s(gare|mairie|poste))|du\\s(centre|marché|parc))$",
|
||||
"^amenity/restaurant\\|((gasthaus|gasthof|schwarzer)?\\s?(adler|löwen))$",
|
||||
"^amenity/restaurant\\|(ali\\sbaba|adria|akropolis|amici|aroma|asia|athen|athos|bahnhof|bamboo|bären|bella\\s(vista|napoli))",
|
||||
"^amenity/restaurant\\|(belvedere|brasserie|canteen|cantina|capri|carpe\\sdiem|casa\\s?(blanca|mia)|casino)",
|
||||
"^amenity/restaurant\\|(china\\s(buffet|city|garden|house|king|palace|star|town|wok)|(da|la|las|el)\\s)",
|
||||
"^amenity/restaurant\\|(delphi|deutsches\\shaus|dionysos|dolce\\svita|dorfkrug|europa|frohsinn)",
|
||||
"^amenity/restaurant\\|(golden\\sdragon|great\\swall)$",
|
||||
"^amenity/restaurant\\|(grüner\\sbaum|hellas|hirsch(en)?|hong\\skong|happy garden)",
|
||||
"^amenity/restaurant\\|(indian\\spalace|ichiban|il\\sgiardino|italia|jade\\sgarden|jägerhof)",
|
||||
"^amenity/restaurant\\|(kirchenwirt|koliba|kreta|kreuz|krone|lamm|le\\s(bistrot|patio))",
|
||||
"^amenity/restaurant\\|(l'escale|lemongrass|linde|lindenhof|little\\s(italy|india)|los\\s(amigos|arcos))",
|
||||
"^amenity/restaurant\\|(lotus|löwen|maharaja|mamma\\smia|mandarin|marco\\spolo|marmaris|mediterraneo|mensa|milano|mykonos|mythos)",
|
||||
"^amenity/restaurant\\|(namaste|napoli|new\\schina|oasis|ochsen|olympia|osaka|paradise|parrilla|peking|pho|pinocchio)",
|
||||
"^amenity/restaurant\\|(pizza\\shouse|pizz(a|e)ria|pizza\\sking|pomodoro|portofino|poseidon|ratskeller|rhodos|ristorante\\s)",
|
||||
"^amenity/restaurant\\|(roma|rose|rössl(e|i)|saffron|saigon|sakura|san\\smarco|santorini|schützenhaus|shalimar|shanghai|sonne|sphinx)",
|
||||
"^amenity/restaurant\\|(sportheim|sternen|syrtaki|taj\\smahal|taste\\sof\\sindia|the\\skitchen|toscana|tokyo|traube)",
|
||||
"^amenity/restaurant\\|(venezia|waldschänke|wok|zorbas|(zum|zur)\\s)",
|
||||
"^amenity/restaurant\\|don camillo",
|
||||
"^amenity/restaurant\\|永和豆漿$",
|
||||
"^amenity/social_facility\\|dom pomocy społecznej$",
|
||||
"^amenity/social_facility\\|maison de retraite$",
|
||||
"^amenity/social_facility\\|safe haven$",
|
||||
"^amenity/vending_machine\\|(schlauchomat|tobaccoland)$",
|
||||
"^shop/(convenience|kiosk)\\|(777|bp|esso|exxon|mobil|shell)$",
|
||||
"^shop/bakery\\|(backshop|schneider|schäfer|каравай)$",
|
||||
"^shop/bakery\\|(bäckerei\\s)?müller",
|
||||
"^shop/clothes\\|трикотаж$",
|
||||
"^shop/convenience\\|(24 часа|августина|альянс|апельсин|ассорти|бер(е|ё)зка|бриз|валентина|весна|визит|ветеран|водолей|гастроном|гермес|гурман|диана|добрый|дружба|домашний|елена)$",
|
||||
"^shop/convenience\\|(almac[eé]n|despensa)$",
|
||||
"^shop/convenience\\|(boutique d'alimentation|sup[eé]rette)$",
|
||||
"^shop/convenience\\|(milk bar|юлия)$",
|
||||
"^shop/convenience\\|(калинка|каштан|квартал|колобок|колос(ок)?|каравай|кооператор|корзинка|кулинария|лакомка|ласточка|лидер|луч|любава|любимый|людмила|люкс)$",
|
||||
"^shop/convenience\\|(меркурий|мечта|мини маркет|мираж|белорусские продукты|продукты 24|натали|надежда|ника)$",
|
||||
"^shop/convenience\\|(оазис|околица|олимп|парус|пеликан|престиж|подсолнух|радуга|росинка|ромашка|родник|рассвет|русь)$",
|
||||
"^shop/convenience\\|(сельпо|семейный|сказка|солнышко|спутник|солнечный|светлана|удача|ивушка)$",
|
||||
"^shop/convenience\\|(юбилейный|южный|эконом|шанс|чайка|центральный|хороший|хуторок|фаворит|фортуна|татьяна|тройка|эдем|уют)$",
|
||||
"^shop/copyshop\\|xerox$",
|
||||
"^shop/cosmetics\\|(vente de )?cosm[eé]tiques?$",
|
||||
"^shop/fishing\\|рыболов$",
|
||||
"^shop/florist\\|(мир цветов|орхидея)$",
|
||||
"^shop/frame\\|rumah\\spenduduk$",
|
||||
"^shop/hairdresser\\|(figaro|haarwerk|стиль|локон|шарм|fodrászat)$",
|
||||
"^shop/hardware\\|(всё для дома|1000 мелочей|хозяюшка)$",
|
||||
"^shop/jewelry\\|(585|jubiler)$",
|
||||
"^shop/money_transfer|mobil(e)?\\smoney$",
|
||||
"^shop/optician\\|(die brille|vision plus|оптика|[oó]ptica)$",
|
||||
"^shop/supermarket\\|(famiglia cooperativa|семейный)$",
|
||||
"^shop/tailor\\|couture\\s?(dame|homme|mixte)?$",
|
||||
"^tourism/(h|m)otel\\|budget inn$",
|
||||
"^tourism/hotel\\|(h(o|ô)tel\\s)?((zur\\s)?post|de france|de la gare|строитель)$",
|
||||
"^tourism/hotel\\|(h(o|ô)tel\\s)?(astoria|belvedere|central|city|continental|europa)\\s?(h(o|ô)tel)?$",
|
||||
"^tourism/hotel\\|(h(o|ô)tel\\s)?(grand|imperial|krone|palace|panorama|park|plaza|royal|victoria)\\s?(h(o|ô)tel)?$"
|
||||
]
|
||||
},
|
||||
|
||||
"operators": {
|
||||
"emoji": "💼",
|
||||
"mainTag": "operator:wikidata",
|
||||
"subclassOf": "Q43229",
|
||||
"keepKV": [
|
||||
"^amenity/bicycle_rental$",
|
||||
"^amenity/car_sharing$",
|
||||
"^amenity/charging_station$",
|
||||
"^amenity/fire_station$",
|
||||
"^amenity/police$",
|
||||
"^amenity/post_(depot|office)$",
|
||||
"^emergency/ambulance_station$"
|
||||
],
|
||||
"discardKVN": [
|
||||
"^amenity/bicycle_rental\\|nextbike( gmbh)?$"
|
||||
]
|
||||
},
|
||||
|
||||
"transit": {
|
||||
"emoji": "🚇",
|
||||
"mainTag": "network:wikidata",
|
||||
"subclassOf": "Q178512",
|
||||
"keepKV": [
|
||||
"^route/(aerialway|bus|light_rail|subway|train|tram|trolleybus|walking_bus)$"
|
||||
],
|
||||
"discardKVN": [
|
||||
"\\|[ilnr][chimsw][nm]$",
|
||||
"\\|city$",
|
||||
"\\|international$",
|
||||
"\\|local$",
|
||||
"\\|mtb$",
|
||||
"\\|municipal$",
|
||||
"\\|national$",
|
||||
"\\|rail$",
|
||||
"\\|regional$",
|
||||
"\\|urban$"
|
||||
]
|
||||
}
|
||||
}
|
||||
|
|
@ -1,40 +0,0 @@
|
|||
{
|
||||
"title": "filter_brands.json",
|
||||
"description": "name-suggestion-index filters",
|
||||
"type": "object",
|
||||
"additionalProperties": false,
|
||||
"required": ["keepTags", "discardKeys", "discardNames"],
|
||||
"properties": {
|
||||
|
||||
"keepTags": {
|
||||
"description": "(required) An array of regular expressions to match OSM tag-pairs to keep.",
|
||||
"type": "array",
|
||||
"uniqueItems": true,
|
||||
"items": {
|
||||
"type": "string",
|
||||
"format": "regex"
|
||||
}
|
||||
},
|
||||
|
||||
"discardKeys": {
|
||||
"description": "(required) An array of regular expressions to match keys to discard.",
|
||||
"type": "array",
|
||||
"uniqueItems": true,
|
||||
"items": {
|
||||
"type": "string",
|
||||
"format": "regex"
|
||||
}
|
||||
},
|
||||
|
||||
"discardNames": {
|
||||
"description": "(required) An array of regular expressions to match names to discard.",
|
||||
"type": "array",
|
||||
"uniqueItems": true,
|
||||
"items": {
|
||||
"type": "string",
|
||||
"format": "regex"
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
20
schema/genericWords.json
Normal file
20
schema/genericWords.json
Normal file
|
|
@ -0,0 +1,20 @@
|
|||
{
|
||||
"title": "genericWords.json",
|
||||
"description": "JSON Schema for config/genericWords.json",
|
||||
"type": "object",
|
||||
"additionalProperties": false,
|
||||
"required": ["genericWords"],
|
||||
"properties": {
|
||||
|
||||
"genericWords": {
|
||||
"description": "(required) An array of regular expressions to match generic words to discard.",
|
||||
"type": "array",
|
||||
"uniqueItems": true,
|
||||
"items": {
|
||||
"type": "string",
|
||||
"format": "regex"
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
|
|
@ -1,6 +1,6 @@
|
|||
{
|
||||
"title": "items.json",
|
||||
"description": "name-suggestion-index items",
|
||||
"description": "JSON Schema for name-suggestion-index items",
|
||||
"type": "object",
|
||||
"additionalProperties": false,
|
||||
"patternProperties": {
|
||||
|
|
@ -183,4 +183,4 @@
|
|||
}
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
|
|
@ -1,33 +1,45 @@
|
|||
{
|
||||
"title": "replacements.json",
|
||||
"description": "name-suggestion-index replacements",
|
||||
"description": "JSON Schema for config/replacements.json",
|
||||
"$comment": "The contents of this file should contain simple QID replacements that can be done in bulk.",
|
||||
"type": "object",
|
||||
"additionalProperties": false,
|
||||
"patternProperties": {
|
||||
"required": ["replacements"],
|
||||
"properties": {
|
||||
|
||||
"^Q\\d+$": {
|
||||
"replacements": {
|
||||
"description": "(required) Legacy Wikidata QID identifiers that should be replaced with new QID identifiers.",
|
||||
"type": "object",
|
||||
"additionalProperties": false,
|
||||
"required": ["wikidata"],
|
||||
"properties": {
|
||||
"patternProperties": {
|
||||
|
||||
"wikidata": {
|
||||
"description": "(required) The replacement Wikidata QID.",
|
||||
"type": "string"
|
||||
},
|
||||
"^Q\\d+$": {
|
||||
"description": "(required) The legacy Wikidata QID.",
|
||||
"type": "object",
|
||||
"additionalProperties": false,
|
||||
"required": ["wikidata"],
|
||||
"properties": {
|
||||
|
||||
"wikipedia": {
|
||||
"description": "(optional) The replacement Wikipedia value.",
|
||||
"type": "string"
|
||||
},
|
||||
"wikidata": {
|
||||
"description": "(required) The replacement Wikidata QID.",
|
||||
"type": "string"
|
||||
},
|
||||
|
||||
"note": {
|
||||
"description": "(optional) An optional note - can contain anything about the item.",
|
||||
"type": "string"
|
||||
"wikipedia": {
|
||||
"description": "(optional) The replacement Wikipedia value.",
|
||||
"type": "string"
|
||||
},
|
||||
|
||||
"note": {
|
||||
"description": "(optional) An optional note - can contain anything about the item.",
|
||||
"type": "string"
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
}
|
||||
64
schema/trees.json
Normal file
64
schema/trees.json
Normal file
|
|
@ -0,0 +1,64 @@
|
|||
{
|
||||
"title": "trees.json",
|
||||
"description": "JSON Schema for config/trees.json",
|
||||
"type": "object",
|
||||
"additionalProperties": false,
|
||||
"required": ["trees"],
|
||||
"properties": {
|
||||
|
||||
"trees": {
|
||||
"type": "object",
|
||||
"additionalProperties": false,
|
||||
"patternProperties": {
|
||||
|
||||
"^\\S+$": {
|
||||
"description": "(required) The name of the tree (e.g. 'brands')",
|
||||
"type": "object",
|
||||
"additionalProperties": false,
|
||||
"required": ["emoji", "mainTag", "subclassOf", "keepKV", "discardKVN"],
|
||||
"properties": {
|
||||
|
||||
"emoji": {
|
||||
"description": "(required) An emoji used for logging and output (e.g. '🍔')",
|
||||
"type": "string"
|
||||
},
|
||||
|
||||
"mainTag": {
|
||||
"description": "(required) An OSM Tag that this tree uses as the main tag (e.g. 'brand:wikidata')",
|
||||
"type": "string",
|
||||
"pattern": "^\\S+$"
|
||||
},
|
||||
|
||||
"subclassOf": {
|
||||
"description": "(required) A Wikidata QID that items in this tree should all be instanceOf/subclassOf (e.g. 'Q4830453')",
|
||||
"type": "string",
|
||||
"pattern": "^Q\\d+$"
|
||||
},
|
||||
|
||||
"keepKV": {
|
||||
"description": "(required) An array of regular expressions that match k/v tuples to keep.",
|
||||
"type": "array",
|
||||
"uniqueItems": true,
|
||||
"items": {
|
||||
"type": "string",
|
||||
"format": "regex"
|
||||
}
|
||||
},
|
||||
|
||||
"discardKVN": {
|
||||
"description": "(required) An array of regular expressions that match k/v|n tuples to discard.",
|
||||
"type": "array",
|
||||
"uniqueItems": true,
|
||||
"items": {
|
||||
"type": "string",
|
||||
"format": "regex"
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
|
@ -20,15 +20,12 @@ console.log(colors.blue('-'.repeat(70)));
|
|||
console.log(colors.blue('🗂 Build index'));
|
||||
console.log(colors.blue('-'.repeat(70)));
|
||||
|
||||
let _config = {};
|
||||
loadConfig();
|
||||
|
||||
let _collected = {};
|
||||
loadCollected();
|
||||
|
||||
let _replacements = {};
|
||||
loadReplacements();
|
||||
|
||||
let _filters = {};
|
||||
loadFilters();
|
||||
|
||||
let _discard = {};
|
||||
let _keep = {};
|
||||
runFilters();
|
||||
|
|
@ -46,6 +43,64 @@ saveIndex();
|
|||
console.log('');
|
||||
|
||||
|
||||
|
||||
//
|
||||
// Load, validate, cleanup config files
|
||||
//
|
||||
function loadConfig() {
|
||||
['trees', 'replacements', 'genericWords'].forEach(which => {
|
||||
const schema = require(`../schema/${which}.json`);
|
||||
const file = `config/${which}.json`;
|
||||
const contents = fs.readFileSync(file, 'utf8');
|
||||
let data;
|
||||
try {
|
||||
data = JSON5.parse(contents);
|
||||
} catch (jsonParseError) {
|
||||
console.error(colors.red(`Error - ${jsonParseError.message} reading:`));
|
||||
console.error(' ' + colors.yellow(file));
|
||||
process.exit(1);
|
||||
}
|
||||
|
||||
// check JSON schema
|
||||
validate(file, data, schema);
|
||||
|
||||
// Clean and sort the files for consistency, save them that way.
|
||||
if (which === 'trees') {
|
||||
Object.keys(data.trees).forEach(t => {
|
||||
let tree = data.trees[t];
|
||||
let cleaned = {
|
||||
emoji: tree.emoji,
|
||||
mainTag: tree.mainTag,
|
||||
subclassOf: tree.subclassOf,
|
||||
keepKV: tree.keepKV.map(s => s.toLowerCase()).sort(),
|
||||
discardKVN: tree.discardKVN.map(s => s.toLowerCase()).sort()
|
||||
}
|
||||
tree = cleaned;
|
||||
});
|
||||
|
||||
} else if (which === 'replacements') {
|
||||
Object.keys(data).forEach(qid => {
|
||||
let replacement = data.replacements[qid];
|
||||
let cleaned = {
|
||||
note: replacement.note,
|
||||
wikidata: replacement.wikidata,
|
||||
wikipedia: replacement.wikipedia
|
||||
}
|
||||
replacement = cleaned;
|
||||
});
|
||||
|
||||
} else if (which === 'genericWords') {
|
||||
data.genericWords = data.genericWords.map(s => s.toLowerCase()).sort();
|
||||
}
|
||||
|
||||
// Lowercase and sort the files for consistency, save them that way.
|
||||
fs.writeFileSync(file, stringify(sort(data)));
|
||||
|
||||
_config[which] = data;
|
||||
});
|
||||
}
|
||||
|
||||
|
||||
//
|
||||
// Load lists of tags collected from OSM from `dist/collected/*`
|
||||
//
|
||||
|
|
@ -67,67 +122,6 @@ function loadCollected() {
|
|||
}
|
||||
|
||||
|
||||
//
|
||||
// Load, validate, cleanup the replacement file
|
||||
//
|
||||
function loadReplacements() {
|
||||
const replacementsSchema = require('../schema/replacements.json');
|
||||
|
||||
const file = 'config/replacements.json';
|
||||
const contents = fs.readFileSync(file, 'utf8');
|
||||
let data;
|
||||
try {
|
||||
data = JSON5.parse(contents);
|
||||
} catch (jsonParseError) {
|
||||
console.error(colors.red(`Error - ${jsonParseError.message} reading:`));
|
||||
console.error(' ' + colors.yellow(file));
|
||||
process.exit(1);
|
||||
}
|
||||
|
||||
// check JSON schema
|
||||
validate(file, data, replacementsSchema);
|
||||
|
||||
// // Lowercase and sort the files for consistency, save them that way.
|
||||
fs.writeFileSync(file, stringify(sort(data)));
|
||||
|
||||
_replacements = data;
|
||||
}
|
||||
|
||||
|
||||
//
|
||||
// Load, validate, and cleanup filter files
|
||||
//
|
||||
function loadFilters() {
|
||||
const filtersSchema = require('../schema/filters.json');
|
||||
|
||||
['brands', 'operators', 'transit'].forEach(tree => {
|
||||
const file = `config/filter_${tree}.json`;
|
||||
const contents = fs.readFileSync(file, 'utf8');
|
||||
let data;
|
||||
try {
|
||||
data = JSON5.parse(contents);
|
||||
} catch (jsonParseError) {
|
||||
console.error(colors.red(`Error - ${jsonParseError.message} reading:`));
|
||||
console.error(' ' + colors.yellow(file));
|
||||
process.exit(1);
|
||||
}
|
||||
|
||||
// check JSON schema
|
||||
validate(file, data, filtersSchema);
|
||||
|
||||
// Lowercase and sort the files for consistency, save them that way.
|
||||
data = {
|
||||
keepTags: data.keepTags.map(s => s.toLowerCase()).sort(),
|
||||
discardKeys: data.discardKeys.map(s => s.toLowerCase()).sort(),
|
||||
discardNames: data.discardNames.map(s => s.toLowerCase()).sort()
|
||||
};
|
||||
fs.writeFileSync(file, stringify(data));
|
||||
|
||||
_filters[tree] = data;
|
||||
});
|
||||
}
|
||||
|
||||
|
||||
//
|
||||
// Filter the tags collected into _keep and _discard lists
|
||||
//
|
||||
|
|
@ -145,8 +139,8 @@ function runFilters() {
|
|||
transit: ['network']
|
||||
};
|
||||
|
||||
['brands', 'operators', 'transit'].forEach(tree => {
|
||||
let filters = _filters[tree];
|
||||
['brands', 'operators', 'transit'].forEach(t => {
|
||||
let tree = _config.trees[t];
|
||||
let discard = _discard[tree] = {};
|
||||
let keep = _keep[tree] = {};
|
||||
|
||||
|
|
@ -161,12 +155,12 @@ function runFilters() {
|
|||
}
|
||||
});
|
||||
|
||||
// Filter by keepTags (move from discard -> keep)
|
||||
filters.keepTags.forEach(s => {
|
||||
// Filter by keepKV (move from discard -> keep)
|
||||
tree.keepKV.forEach(s => {
|
||||
const re = new RegExp(s, 'i');
|
||||
for (const kvn in discard) {
|
||||
const tag = kvn.split('|', 2)[0];
|
||||
if (re.test(tag)) {
|
||||
const kv = kvn.split('|', 2)[0];
|
||||
if (re.test(kv)) {
|
||||
keep[kvn] = discard[kvn];
|
||||
delete discard[kvn];
|
||||
}
|
||||
|
|
@ -174,7 +168,7 @@ function runFilters() {
|
|||
});
|
||||
|
||||
// Filter by discardKeys (move from keep -> discard)
|
||||
filters.discardKeys.forEach(s => {
|
||||
tree.discardKVN.forEach(s => {
|
||||
const re = new RegExp(s, 'i');
|
||||
for (const kvn in keep) {
|
||||
if (re.test(kvn)) {
|
||||
|
|
@ -185,7 +179,7 @@ function runFilters() {
|
|||
});
|
||||
|
||||
// filter by discardNames (move from keep -> discard)
|
||||
filters.discardNames.forEach(s => {
|
||||
_config.genericWords.forEach(s => {
|
||||
const re = new RegExp(s, 'i');
|
||||
for (let kvn in keep) {
|
||||
const name = kvn.split('|', 2)[1];
|
||||
|
|
@ -429,7 +423,7 @@ function mergeItems() {
|
|||
const matchTag = osmkey.match(/^(\w+):wikidata$/);
|
||||
if (matchTag) { // Look at '*:wikidata' tags
|
||||
const wd = tags[osmkey];
|
||||
const replace = _replacements[wd]; // If it matches a QID in the replacement list...
|
||||
const replace = _config.replacements[wd]; // If it matches a QID in the replacement list...
|
||||
|
||||
if (replace && replace.wikidata !== undefined) { // replace or delete `*:wikidata` tag
|
||||
if (replace.wikidata) {
|
||||
|
|
@ -474,22 +468,11 @@ function mergeItems() {
|
|||
// checkItems()
|
||||
// Checks all the items for several kinds of issues
|
||||
//
|
||||
function checkItems(tree) {
|
||||
function checkItems(t) {
|
||||
console.log('');
|
||||
console.log('🏗 ' + colors.yellow(`Checking ${tree}...`));
|
||||
|
||||
const icon = {
|
||||
brands: '🍔',
|
||||
operators: '💼',
|
||||
transit: '🚅'
|
||||
}[tree];
|
||||
|
||||
const wdTag = {
|
||||
brands: 'brand:wikidata',
|
||||
operators: 'operator:wikidata',
|
||||
transit: 'network:wikidata'
|
||||
}[tree];
|
||||
console.log('🏗 ' + colors.yellow(`Checking ${t}...`));
|
||||
|
||||
const tree = _config.trees[t];
|
||||
const oddChars = /[\s=!"#%'*{},.\/:?\(\)\[\]@\\$\^*+<>«»~`’\u00a1\u00a7\u00b6\u00b7\u00bf\u037e\u0387\u055a-\u055f\u0589\u05c0\u05c3\u05c6\u05f3\u05f4\u0609\u060a\u060c\u060d\u061b\u061e\u061f\u066a-\u066d\u06d4\u0700-\u070d\u07f7-\u07f9\u0830-\u083e\u085e\u0964\u0965\u0970\u0af0\u0df4\u0e4f\u0e5a\u0e5b\u0f04-\u0f12\u0f14\u0f85\u0fd0-\u0fd4\u0fd9\u0fda\u104a-\u104f\u10fb\u1360-\u1368\u166d\u166e\u16eb-\u16ed\u1735\u1736\u17d4-\u17d6\u17d8-\u17da\u1800-\u1805\u1807-\u180a\u1944\u1945\u1a1e\u1a1f\u1aa0-\u1aa6\u1aa8-\u1aad\u1b5a-\u1b60\u1bfc-\u1bff\u1c3b-\u1c3f\u1c7e\u1c7f\u1cc0-\u1cc7\u1cd3\u200b-\u200f\u2016\u2017\u2020-\u2027\u2030-\u2038\u203b-\u203e\u2041-\u2043\u2047-\u2051\u2053\u2055-\u205e\u2cf9-\u2cfc\u2cfe\u2cff\u2d70\u2e00\u2e01\u2e06-\u2e08\u2e0b\u2e0e-\u2e16\u2e18\u2e19\u2e1b\u2e1e\u2e1f\u2e2a-\u2e2e\u2e30-\u2e39\u3001-\u3003\u303d\u30fb\ua4fe\ua4ff\ua60d-\ua60f\ua673\ua67e\ua6f2-\ua6f7\ua874-\ua877\ua8ce\ua8cf\ua8f8-\ua8fa\ua92e\ua92f\ua95f\ua9c1-\ua9cd\ua9de\ua9df\uaa5c-\uaa5f\uaade\uaadf\uaaf0\uaaf1\uabeb\ufe10-\ufe16\ufe19\ufe30\ufe45\ufe46\ufe49-\ufe4c\ufe50-\ufe52\ufe54-\ufe57\ufe5f-\ufe61\ufe68\ufe6a\ufe6b\ufeff\uff01-\uff03\uff05-\uff07\uff0a\uff0c\uff0e\uff0f\uff1a\uff1b\uff1f\uff20\uff3c\uff61\uff64\uff65]+/g;
|
||||
|
||||
let warnMatched = matcher.getWarnings();
|
||||
|
|
@ -503,7 +486,7 @@ function checkItems(tree) {
|
|||
let total = 0; // total items
|
||||
let totalWd = 0; // total items with wikidata
|
||||
|
||||
const paths = Object.keys(_cache.path).filter(tkv => tkv.split('/')[0] === tree);
|
||||
const paths = Object.keys(_cache.path).filter(tkv => tkv.split('/')[0] === t);
|
||||
const display = (val) => `${val.displayName} (${val.id})`;
|
||||
|
||||
paths.forEach(tkv => {
|
||||
|
|
@ -519,7 +502,7 @@ function checkItems(tree) {
|
|||
const tags = item.tags;
|
||||
|
||||
total++;
|
||||
if (tags[wdTag]) totalWd++;
|
||||
if (tags[tree.mainTag]) totalWd++;
|
||||
|
||||
// check tags
|
||||
Object.keys(tags).forEach(osmkey => {
|
||||
|
|
@ -587,7 +570,7 @@ function checkItems(tree) {
|
|||
// // - The items have the same locationSet (or the one without wikidata is worldwide)
|
||||
// const name = tags.name || tags.brand;
|
||||
// const stem = stemmer(name) || name;
|
||||
// const itemwd = tags[wdTag];
|
||||
// const itemwd = tags[tree.mainTag];
|
||||
// const itemls = loco.validateLocationSet(item.locationSet).id;
|
||||
|
||||
// if (!seenName[stem]) seenName[stem] = new Set();
|
||||
|
|
@ -596,7 +579,7 @@ function checkItems(tree) {
|
|||
// if (seenName[stem].size > 1) {
|
||||
// seenName[stem].forEach(other => {
|
||||
// if (other.id === item.id) return; // skip self
|
||||
// const otherwd = other.tags[wdTag];
|
||||
// const otherwd = other.tags[tree.mainTag];
|
||||
// const otherls = loco.validateLocationSet(other.locationSet).id;
|
||||
|
||||
// // pick one of the items without a wikidata tag to be the "duplicate"
|
||||
|
|
@ -690,7 +673,7 @@ function checkItems(tree) {
|
|||
const pctWd = total > 0 ? (totalWd * 100 / total).toFixed(1) : 0;
|
||||
|
||||
console.log('');
|
||||
console.info(colors.blue.bold(`${icon} ${tree}/* completeness:`));
|
||||
console.info(colors.blue.bold(`${tree.emoji} ${t}/* completeness:`));
|
||||
console.info(colors.blue.bold(` ${total} total`));
|
||||
console.info(colors.blue.bold(` ${totalWd} (${pctWd}%) with a '${wdTag}' tag`));
|
||||
console.info(colors.blue.bold(` ${totalWd} (${pctWd}%) with a '${tree.mainTag}' tag`));
|
||||
}
|
||||
|
|
@ -2,20 +2,27 @@ const featureCollection = require('../dist/featureCollection.json');
|
|||
const LocationConflation = require('@ideditor/location-conflation');
|
||||
const loco = new LocationConflation(featureCollection);
|
||||
|
||||
// validate the filter files
|
||||
const filtersSchema = require('../schema/filters.json');
|
||||
// validate the config files
|
||||
const validate = require('../lib/validate');
|
||||
|
||||
const filterBrands = require('../config/filter_brands.json');
|
||||
validate('config/filter_brands.json', filterBrands, filtersSchema);
|
||||
const filterOperators = require('../config/filter_operators.json');
|
||||
validate('config/filter_operators.json', filterOperators, filtersSchema);
|
||||
const filterTransit = require('../config/filter_brands.json');
|
||||
validate('config/filter_transit.json', filterTransit, filtersSchema);
|
||||
validate(
|
||||
'config/generic_words.json',
|
||||
require('../config/generic_words.json'),
|
||||
require('../schema/generic_words.json')
|
||||
);
|
||||
validate(
|
||||
'config/replacements.json',
|
||||
require('../config/replacements.json'),
|
||||
require('../schema/replacements.json')
|
||||
);
|
||||
validate(
|
||||
'config/trees.json',
|
||||
require('../config/trees.json'),
|
||||
require('../schema/trees.json')
|
||||
);
|
||||
|
||||
// reading a fileTree will also validate its contents
|
||||
const fileTree = require('../lib/file_tree');
|
||||
let _cache = {};
|
||||
fileTree.read('brands', _cache, loco);
|
||||
fileTree.read('operators', _cache, loco);
|
||||
fileTree.read('transit', _cache, loco);
|
||||
fileTree.read('transit', _cache, loco);
|
||||
Loading…
Add table
Reference in a new issue