make country.lang() and lang.country() functions work with real count…

…ry names (Ekaterina Zalivina)

NAMESPACE

@@ -67,6 +67,7 @@ importFrom(stats,coef)
 importFrom(stats,complete.cases)
 importFrom(stats,density)
 importFrom(stats,glm)
+importFrom(stats,na.omit)
 importFrom(stats,sd)
 importFrom(stringdist,stringdist)
 importFrom(utils,download.file)

NEWS

@@ -3,6 +3,7 @@ lingtypology 1.1.9
 - update the `autotyp.feature()` function after the new release 1.0.0
 - make it possible to have just one group in the `control`
 - update glottolog to v. 4.6
+- make `country.lang()` and `lang.country()` functions work with real country names (Ekaterina Zalivina)
 
 lingtypology 1.1.8 (12.10.2021)
 - fix `country.lang()` functions, close #70

R/countries.R

@@ -0,0 +1,15 @@
+#' Catalogue of countries
+#'
+#' Catalogue of countries, ISO-codes and official languages
+#'
+#' @format A data frame with 189 rows and 5 variables:
+#' \describe{
+#'   \item{alpha3}{ISO 3166-3 code of the country}
+#'   \item{alpha2}{ISO 3166-2 code of the country}
+#'   \item{country_name}{Country name}
+#'   \item{additional_names}{Additional names of the country}
+#'   \item{official_languages}{Official languages}
+#' }
+#'
+
+"countries"

R/country.lang.R

@@ -2,23 +2,40 @@
 #'
 #' Takes any vector of languages and returns countries where those languages are used as ISO 3166-1 alpha-2 codes.
 #' @param x A character vector of the languages (can be written in lower case)
+#' @param full_name A logical value, whether return ISO 3166-2 codes or full names.
 #' @author George Moroz <[email protected]>
 #' @seealso \code{\link{aff.lang}}, \code{\link{area.lang}}, \code{\link{gltc.lang}}, \code{\link{iso.lang}}, \code{\link{lat.lang}}, \code{\link{long.lang}}, \code{\link{subc.lang}}, \code{\link{url.lang}}
 #' @examples
 #' country.lang('Korean')
 #' country.lang(c('Korean', 'Polish'))
 #' @export
+#'
+#'
+
 
-country.lang <- function(x) {
+country.lang <- function(x, full_name = TRUE) {
   if (typeof(x) == "list") {
     x <- unlist(x)
   }
   glottolog <- lingtypology::glottolog
-  vapply(x, function(y) {
+  countries <- lingtypology::countries
+
+  result <- vapply(x, function(y) {
     ifelse(
-      is.glottolog(y, response = TRUE) == TRUE,
+      isTRUE(is.glottolog(y, response = TRUE)),
       glottolog[tolower(glottolog$language) %in% tolower(y), ]$countries,
       NA_character_
     )
   }, character(1))
+
+  if(full_name){
+    vapply(result, function(y) {
+      ifelse(
+        !is.na(y),
+        paste0(unique(countries[countries$alpha2 %in% unlist(strsplit(y, ";")),"country_name"]),
+               collapse = ";"),
+        NA_character_
+      )
+    }, character(1))
+  }
 }

R/lang.country.R

@@ -9,46 +9,31 @@
 #' lang.country('AD')
 #' lang.country(c('AD', 'AE'))
 #' @export
+#' @importFrom stats na.omit
+#'
 
-lang.country <- function(x, list = FALSE) {
+lang.country <- function(x, list = TRUE) {
   if (typeof(x) == "list") {
     x <- unlist(x)
   }
-  countries <- c('AD', 'AE', 'AF', 'AG', 'AI', 'AL', 'AM', 'AO', 'AR', 'AS',
-                 'AT', 'AU', 'AW', 'AX', 'AZ', 'BA', 'BB', 'BD', 'BE', 'BF',
-                 'BG', 'BH', 'BI', 'BJ', 'BL', 'BM', 'BN', 'BO', 'BQ', 'BR',
-                 'BS', 'BT', 'BW', 'BY', 'BZ', 'CA', 'CC', 'CD', 'CF', 'CG',
-                 'CH', 'CI', 'CK', 'CL', 'CM', 'CN', 'CO', 'CR', 'CU', 'CV',
-                 'CW', 'CX', 'CY', 'CZ', 'DE', 'DJ', 'DK', 'DM', 'DO', 'DZ',
-                 'EC', 'EE', 'EG', 'EH', 'ER', 'ES', 'ET', 'FI', 'FJ', 'FK',
-                 'FM', 'FO', 'FR', 'GA', 'GB', 'GD', 'GE', 'GF', 'GG', 'GH',
-                 'GI', 'GL', 'GM', 'GN', 'GP', 'GQ', 'GR', 'GS', 'GT', 'GU',
-                 'GW', 'GY', 'HK', 'HN', 'HR', 'HT', 'HU', 'ID', 'IE', 'IL',
-                 'IM', 'IN', 'IO', 'IQ', 'IR', 'IS', 'IT', 'JE', 'JM', 'JO',
-                 'JP', 'KE', 'KG', 'KH', 'KI', 'KM', 'KN', 'KP', 'KR', 'KW',
-                 'KY', 'KZ', 'LA', 'LB', 'LC', 'LI', 'LK', 'LR', 'LS', 'LT',
-                 'LU', 'LV', 'LY', 'MA', 'MC', 'MD', 'ME', 'MF', 'MG', 'MH',
-                 'MK', 'ML', 'MM', 'MN', 'MO', 'MP', 'MQ', 'MR', 'MS', 'MT',
-                 'MU', 'MV', 'MW', 'MX', 'MY', 'MZ', 'NA', 'NC', 'NE', 'NF',
-                 'NG', 'NI', 'NL', 'NO', 'NP', 'NR', 'NU', 'NZ', 'OM', 'PA',
-                 'PE', 'PF', 'PG', 'PH', 'PK', 'PL', 'PM', 'PN', 'PR', 'PS',
-                 'PT', 'PW', 'PY', 'QA', 'RE', 'RO', 'RS', 'RU', 'RW', 'SA',
-                 'SB', 'SC', 'SD', 'SE', 'SG', 'SH', 'SI', 'SJ', 'SK', 'SL',
-                 'SM', 'SN', 'SO', 'SR', 'SS', 'ST', 'SV', 'SX', 'SY', 'SZ',
-                 'TC', 'TD', 'TF', 'TG', 'TH', 'TJ', 'TK', 'TL', 'TM', 'TN',
-                 'TO', 'TR', 'TT', 'TV', 'TW', 'TZ', 'UA', 'UG', 'UM', 'US',
-                 'UY', 'UZ', 'VA', 'VC', 'VE', 'VG', 'VI', 'VN', 'VU', 'WF',
-                 'WS', 'YE', 'YT', 'ZA', 'ZM', 'ZW')
   glottolog <- lingtypology::glottolog
+  countries <- lingtypology::countries
   result <- lapply(x, function(y) {
-    if(toupper(y) %in% countries == TRUE){
+    if(!is.na(y)){
+      if(nchar(y) > 2){
+        y <- unique(countries[countries$country_name %in% y, "alpha2"],
+                    countries[countries$additional_names %in% y, "alpha2"])
+      }}
+
+    if(toupper(y) %in% stats::na.omit(countries$alpha2) == TRUE){
       glottolog[grepl(toupper(y), glottolog$countries),]$language
     } else{
       NA_character_
     }})
-  if (list == FALSE) {
+  if(isFALSE(list)) {
     unlist(result)
   } else {
+    names(result) <- x
     result
   }
 }

database_creation/country.data.csv

@@ -0,0 +1,190 @@
+alpha3,alpha2,country_name,additional_names,official_languages
+ABW,AW,Aruba,Country of Aruba,NA
+AFG,AF,Afganistan,Islamic Republic of Afghanistan,"Dari, Southern Pashto, Central Pashto, Northern Pashto"
+AGO,AO,Angola,Republic of Angola,Portuguese
+AIA,AI,Anguilla,NA,NA
+ALA,AX,Åland,Åland Islands,Swedish
+ALB,AL,Albania,Republic of Albania,Northern Tosk Albanian
+AND,AD,Andorra,Principality of Andorra,Catalan
+AND,AD,Andorra,Principality of the Valleys of Andorra,Catalan
+ARE,AE,United Arab Emirates,the Emirates,Standard Arabic
+ARE,AE,United Arab Emirates,Emirates,Standard Arabic
+ARE,AE,United Arab Emirates,UAE,Standard Arabic
+ARG,AR,Argentina,Argentine Republic,Spanish
+ARG,AR,Argentina,the Argentine,Spanish
+ARM,AM,Armenia,Republic of Armenia,Armenian
+ASM,AS,American Samoa,NA,NA
+ATG,AG,Antigua and Barbuda,NA,English
+AUS,AU,Australia,Commonwealth of Australia,English
+AUT,AT,Austria,Republic of Austria,German
+AZE,AZ,Azerbaijan,Republic of Azerbaijan,Azerbaijani
+BDI,BI,Burundi,Republic of Burundi,"Rundi, Rundi-Kitwa, French, English"
+BEL,BE,Belgium,Kingdom of Belgium,"Dutch, French, German"
+BEN,BJ,Benin,Republic of Benin,French
+BFA,BF,Burkina Faso,NA,French
+BGD,BD,Bangladesh,People's Republic of Bangladesh,Bengali
+BGR,BG,Bulgaria,Republic of Bulgaria,Bulgarian
+BHR,BH,Bahrain,Kingdom of Bahrain,Standard Arabic
+BHS,BS,Bahamas,Commonwealth of the Bahamas,English
+BIH,BA,Bosnia and Herzegovina,B&H,"Bosnian, Croatian, Serbian"
+BLR,BY,Belarus,Republic of Belarus,"Belarusian, Russian"
+BLR,BY,Belarus,Byelorussia,"Belarusian, Russian"
+BLZ,BZ,Belize,NA,English
+BOL,BO,Bolivia,Plurinational State of Bolivia,"Spanish, Central Aymara, Southern Aymara, Araona, Baure, Chiquitano, Canichana, Cavinena, Cayubaba, Chacobo, Moseten-Chimane, Ese Ejja, Eastern Bolivian Guarani, Western Bolivian Guarani, Pauserna, Guarayu, Itonama, Leco, Callawalla, Machinere, Reyesano, Trinitario, Ignaciano, Moseten-Chimane, Itene, Movima, Chacobo, Pacahuara, Karipuna (Pano-Tacanan), Shinabo, Puquina, North Bolivian Quechua, South Bolivian Quechua, Siriono, Jora, Yuqui, Tacana, Toromono, Chipaya, Uru, Zamuco, Ayoreo, Yuracare, Shanenawa, Sharanahua, Yaminahua, Yawanawa, Yora, Wichi Lhamtes Nocten"
+BRA,BR,Brazil,Federative Republic of Brazil,Portuguese
+BRB,BB,Barbados,NA,English
+BRN,BN,Brunei,Nation of Brunei,"English, Standard Malay"
+BRN,BN,Brunei,the Abode of Peace,"English, Standard Malay"
+BTN,BT,Bhutan,Kingdom of Bhutan,Dzongkha
+BWA,BW,Botswana,Republic of Botswana,English
+CAF,CF,Central African Republic,CAR,French
+CAF,CF,Central African Republic,RCA,French
+CAF,CF,Central African Republic,Centrafrique,French
+CAN,CA,Canada,NA,"English, French"
+CHE,CH,Switzerland,Swiss Confederation,"German, French, Italian, Romansh"
+CHL,CL,Chile,Republic of Chile,Spanish
+CHN,CN,China,People's Republic of China,Mandarin Chinese
+CHN,CN,China,PRC,Mandarin Chinese
+CIV,CI,Ivory Coast,Cote d'Ivoire,French
+CMR,CM,Cameroon,Republic of Cameroon,"English, French"
+COD,CD,Congo-Kinshasa,Democratic Republic of the Congo,French
+COD,CD,Congo-Kinshasa, Congo-Kinshasa,French
+COD,CD,Congo-Kinshasa,DR Congo,French
+COD,CD,Congo-Kinshasa,the DRC,French
+COD,CD,Congo-Kinshasa,the DROC,French
+COD,CD,Congo-Kinshasa,the Congo,French
+COG,CG,Congo-Brazzaville,Republic of the Congo,French
+COG,CG,Congo-Brazzaville,Congo-Brazzaville,French
+COL,CO,Colombia,Republic of Colombia,Spanish
+COM,KM,Comoros,Union of the Comoros,"French, Standard Arabic, Maore Comorian"
+CPV,CV,Cape Verde,Cabo Verde,Portuguese
+CRI,CR,Costa Rica,Republic of Costa Rica,Spanish
+CUB,CU,Cuba,Republic of Cuba,Spanish
+CYP,CY,Cyprus,Republic of Cyprus,"Modern Greek, Turkish"
+CZE,CZ,Czech,Czech Republic,"Czech, Slovak"
+CZE,CZ,Czech,Czechia,"Czech, Slovak"
+DEU,DE,Germany,Federal Republic of Germany,German
+DJI,DJ,Djibouti,Republic of Djibouti,"Standard Arabic, French"
+DMA,DM,Dominica,Commonwealth of Dominica,English
+DNK,DK,Denmark,Kingdom of Denmark,Danish
+DOM,DO,Dominican Republic,NA,Spanish
+DZA,DZ,Algeria,People's Democratic Republic of Algeria,"Standard Arabic, Algerian Arabic, Western Algerian Berber"
+ECU,EC,Ecuador,Republic of Ecuador,Spanish
+EGY,EG,Egypt,Arab Republic of Egypt,"Egyptian Arabic, Standard Arabic"
+ERI,ER,Eritrea,State of Eritrea,"Standard Arabic, English, Tigrinya"
+EST,EE,Estonia,Republic of Estonia,Estonian
+ETH,ET,Ethiopia,Federal Democratic Republic of Ethiopia,Amharic
+FIN,FI,Finland,Republic of Finland,"Finnish, Swedish"
+FJI,FJ,Fiji,Republic of Fiji,"English, Fijian, Fiji Hindi"
+FRA,FR,France,French Republic,French
+GAB,GA,Gabon,Gabonese Republic,French
+GBR,GB,United Kingdom,United Kingdom of Great Britain and Northern Ireland,English
+GBR,GB,United Kingdom,UK,English
+GEO,GE,Georgia,Republic of Georgia,Georgian
+GHA,GH,Ghana,Republic of Ghana,English
+GIN,GN,Guinea,Republic of Guinea,French
+GIN,GN,Guinea,Guinea-Conakry,French
+GMB,GM,Gambia,Republic of the Gambia,English
+GMB,GM,Gambia,The Gambia,English
+GNB,GW,Guinea-Bissau,Republic of Guinea-Bissau,Portuguese
+GNQ,GQ,Equatorial Guinea,Republic of Equatorial Guinea,"Spanish, French, Portuguese"
+GRC,GR,Greece,Hellenic Republic,Modern Greek
+GRC,GR,Greece,the Hellenic Republic,Modern Greek
+GRD,GD,Grenada,NA,English
+GTM,GT,Guatemala,Republic of Guatemala,Spanish
+GUY,GY,Guyana,Co-operative Republic of Guyana,English
+HKG,HK,Hong Kong,HKSAR,"Yue Chinese, Mandarin Chinese, English"
+HND,HN,Honduras,Republic of Honduras,Spanish
+HRV,HR,Croatia,Republic of Croatia,Croatian
+HTI,HT,Haiti,Republic of Haiti,"Haitian, French"
+HUN,HU,Hungary,NA,Hungarian
+IDN,ID,Indonesia,Republic of Indonesia,"Indonesian, Indonesian Sign Language"
+IND,IN,India,Republic of India,"English, Hindi"
+IRL,IE,Ireland,Republic of Ireland,"Irish, English"
+IRN,IR,Iran,Islamic Republic of Iran,Western Farsi
+IRN,IR,Iran,Persia,Western Farsi
+IRQ,IQ,Iraq,NA,Standard Arabic
+ISL,IS,Iceland,Republic of Iceland,"Icelandic, Icelandic Sign Language"
+ISR,IL,Israel,State of Israel,"Hebrew, Standard Arabic"
+ITA,IT,Italy,Italian Republic,Italian
+JAM,JM,Jamaica,NA,English
+JOR,JO,Jordan,Hashemite Kingdom of Jordan,Standard Arabic
+JPN,JP,Japan,NA,Japanese
+KAZ,KZ,Kazakhstan,Republic of Kazakhstan,"Kazakh, Russian"
+KEN,KE,Kenya,Republic of Kenya,"English, Swahili"
+KGZ,KG,Kyrgyzstan,Kyrgyz Republic,"Kirghiz, Russian"
+KHM,KH,Cambodia,Kingdom of Cambodia,"Central Khmer, Northern Khmer"
+KIR,KI,Kiribati,Republic of Kiribati,"Gilbertese, English"
+KNA,KN,Saint Kitts and Nevis,Federation of Saint Kitts and Nevis,English
+KNA,KN,Saint Kitts and Nevis,Federation of Saint Christopher and Nevis,English
+KNA,KN,Saint Kitts and Nevis,SKN,English
+KNA,KN,Saint Kitts and Nevis,SKAN,English
+KOR,KR,South Korea,Republic of Korea,Korean
+KOR,KR,South Korea,ROK,Korean
+KWT,KW,Kuwait,State of Kuwait,Standard Arabic
+LAO,LA,Laos,Lao People's Democratic Republic,Lao
+LBN,LB,Lebanon,The Lebanese Republic,Standard Arabic
+LBN,LB,Lebanon,Lebanese Republic,Standard Arabic
+LBR,LR,Liberia,Republic of Liberia,English
+LBY,LY,Libya,NA,"Libyan Arabic, Standard Arabic"
+LIE,LI,Liechtenstein,Principality of Liechtenstein,German
+LKA,LK,Sri Lanka,Democratic Socialist Republic of Sri Lanka,"Sinhala, Tamil"
+LSO,LS,Lesotho,Kingdom of Lesotho,"Southern Sotho, English"
+LTU,LT,Lithuania,Republic of Lithuania,Lithuanian
+LUX,LU,Luxembourg,Grand Duchy of Luxembourg,"French, German, Luxembourgish"
+LVA,LV,Latvia,Republic of Latvia,Latvian
+MAR,MA,Morocco,Kingdom of Morocco,"Standard Arabic, Berber"
+MDA,MD,Moldova,Republic of Moldova,Romanian
+MDG,MG,Madagascar,Republic of Madagascar,"Plateau Malagasy, French"
+MDV,MV,Maldives,Republic of the Maldives,Dhivehi
+MEX,MX,Mexico,United Mexican States,Spanish
+MKD,MK,Macedonia,Republic of Macedonia,Macedonian
+MKD,MK,Macedonia,Republic of North Macedonia,Macedonian
+MMR,MM,Myanmar,Republic of the Union of Myanmar,Central Moroccan Berber
+MRT,MR,Mauritania,Islamic Republic of Mauritania,Standard Arabic
+MWI,MW,Malawi,Republic of Malawi,English
+MYS,MY,Malaysia,Federation of Malaysia,Standard Malay
+NLD,NL,Netherlands,Kingdom of the Netherlands,Dutch
+NLD,NL,Netherlands,the Netherlands,Dutch
+NLD,NL,Netherlands,Holland,Dutch
+NOR,NO,Norway,Kingdom of Norway,Norwegian
+NPL,NP,Nepal,Federal Democratic Republic Of Nepal,Nepali
+NZL,NZ,New Zealand,Realm of New Zealand,"English, Maori, New Zealand Sign Language"
+PAK,PK,Pakistan,Islamic Republic of Pakistan,"Urdu, English"
+PER,PE,Peru,Republic of Peru,"Spanish, Quechua, Aymara"
+PHL,PH,Philippines,Republic of the Philippines,"Filipino, English"
+PLW,PW,Palau,Republic of Palau,"English, Palauan"
+POL,PL,Poland,Republic of Poland,Polish
+PRI,PR,Puerto Rico,Commonwealth of Puerto Rico,"Spanish, English"
+PRK,KP,North Korea,Democratic People's Republic of Korea,Korean
+PRK,KP,North Korea,DPRK,Korean
+PRT,PT,Portugal,Portuguese Republic,Portuguese
+RUS,RU,Russia,Russian Federation,Russian
+RWA,RW,Rwanda,Republic of Rwanda,"Kinyarwanda, English, French"
+SAU,SA,Saudi Arabia,Kingdom of Saudi Arabia,Standard Arabic
+SAU,SA,Saudi Arabia,KSA,Standard Arabic
+SDN,SD,Sudan,Republic of the Sudan,"Arabic, English"
+SGP,SG,Singapore,Republic of Singapore,"English, Standard Malay, Mandarin Chinese, Tamil"
+SLV,SV,El Salvador,Republic of El Salvador,Spanish
+SRB,RS,Serbia,Republic of Serbia,Serbian
+SUR,SR,Suriname,Surinam,Dutch
+SUR,SR,Suriname,Republic of Suriname,Dutch
+TCD,TD,Chad,Republic of Chad,"Chadian Arabic, Standard Arabic, French"
+TCD,TD,Chad,Tchad,"Chadian Arabic, Standard Arabic, French"
+TGO,TG,Togo,Togolese Republic,French
+THA,TH,Thailand,Kingdom of Thailand,Thai
+TLS,TL,East Timor,Democratic Republic of Timor-Leste,"Portuguese, Tetum"
+TTO,TT,Trinidad and Tobago,Republic of Trinidad and Tobago,English
+TUR,TR,Turkey,Republic of Turkey,Turkish
+TWN,TW,Taiwan,Republic of China,Mandarin Chinese
+TWN,TW,Taiwan,ROC,Mandarin Chinese
+TZA,TZ,Tanzania,United Republic of Tanzania,Swahili
+USA,US,United States,United States of America,English
+USA,US,United States,USA,English
+VNM,VN,Vietnam,Socialist Republic of Vietnam,Vietnamese
+VNM,VN,Vietnam,SRV,Vietnamese
+VNM,VN,Vietnam,Viet Nam,Vietnamese
+ZMB,ZM,Zambia,Republic of Zambia,English
+ZWE,ZW,Zimbabwe,Republic of Zimbabwe,"English, Nyanja, Sena, Kalanga, Tsoa, Nambya, Ndau, Zimbabwean Ndebele, Tsonga, Zimbabwe Sign Language, Southern Sotho, Tonga (Zambia), Tswana, Venda, Xhosa"
+NA,NA,Abkhazia,NA,"Abkhaz, Russian"
+NA,NA,Kosovo,Republic of Kosovo,"Northern Tosk Albanian, Turkish"