address-formatting/convert.lisp

(ql:quickload '(:dexador :nyaml ))

(defun dehash (value)
  (typecase value
    (list (mapcar 'dehash value))
    (string value)
    (array (map 'list 'dehash value))
    (hash-table (loop for k being the hash-keys in value using (hash-value v)
                      collect (cons k (dehash v))))
    (t value)))

(defparameter +base-uri+ "https://github.com/OpenCageData/address-formatting/raw/master/conf/")

(defun write-sexp (path value)
  (with-open-file (s path
                     :direction :output
                     :if-exists :supersede
                     :if-does-not-exist :create
                     :external-format :utf-8)
      (write value :stream s)))

(defun convert (file &optional multi)
  (let* ((url (concatenate 'string +base-uri+ file))
         (name (subseq file (1+ (or (position #\/ file :from-end t) -1))
                       (position #\. file :from-end t)))
         (data (nyaml:parse (dex:request url) :multi-document-p t))
         (alist (dehash (cdr data))))
    (write-sexp (format nil "~a.sexp" name) (if multi alist (car alist)))
    (format t "done~%")))

(defparameter +abbrev-uri+ "https://github.com/OpenCageData/address-formatting/raw/master/conf/abbreviations/~a.yaml")
(defparameter +langs+ '("ca" "cs" "da" "de" "en" "es" "et" "eu" "fi" "fr" "hu" "it" "nl"
                        "no" "pl" "pt" "ro" "ru" "se" "sk" "tr" "uk" "vn"))
(defun convert-abbrevs ()
  (write-sexp "abbrevs.sexp"
              (loop for lang in +langs+
                    for uri = (format nil +abbrev-uri+ lang)
                    do (format t "~a.." lang)
                    collect (cons lang (dehash (nyaml:parse (dex:request uri))))))
  (terpri))

(defun convert-all ()
  (convert-abbrevs)
  (loop for file in '("countries/worldwide.yaml"
                      "country2lang.yaml"
                      "county_codes.yaml"
                      "state_codes.yaml")
        do (format t "Processing ~a.." file)
        do (convert file))
  (convert "components.yaml" t))