instaparse example 1 (clojure sexp grammer)

core.clj

(ns parser.core
  (:require [clojure.java.io :as io]
            [clojure.string :as str]
            [clojure.edn :as edn]
            [instaparse.core :as insta]))

;; TODO comment and newline support
;;  newline = '\n' | '\r\n'
;;  <comment-start> = ';'
;;  <comment> = comment-start #'.*' newline
;;     :
(def sexp
  (insta/parser
   "S = ws* sexpr ws*
    whitespace = ' ' | '\t'
    <comma> = ','
    <ws> = <whitespace> | <comma>
    <lpar> = <'('>
    <rpar> = <')'>
    <map-start> = <'{'>
    <map-end> = <'}'>
    <set-start> = <'#{'>
    <set-end> = <'}'>
    <array-start> = <'['>
    <array-end> = <']'>
    integer = #'[+-]?\\d+'
    double = #'[+-]?\\d+\\.\\d*'
    <number> = ( integer | double )
    <dquote> = '\"'
    <not-dquote> = #'[^\"]'
    string = <dquote> not-dquote* <dquote>
    <identifier-class> = #'[\\p{Alpha}\\-_$!%&<>=][\\d\\p{Alpha}\\-_$!%&<>=]*'
    identifier = identifier-class
    keyword = <':'> identifier-class
    quoted = <'\\''> ws* sexpr
    <sexpr> = string | identifier | number | keyword | list | array | map | set | quoted
    list = lpar ( sexpr | ws )* rpar
    array = array-start ( sexpr | ws )* array-end
    map-key = sexpr
    map-val = sexpr
    map = map-start ws* ( map-key ws* map-val ws* )* map-end
    set = set-start ( sexpr | ws )* set-end
"))
;; (sexp " (   -999.99 888 [77.7 -66. ]   a-b_c3$d {:d, 999, :e-e9_$e fff} #{x -0} ' (de-fg hi$ ) j \"string string\" \"\") ")
;; =>
;; [:S
;;  [:list
;;   [:double "-999.99"]
;;   [:integer "888"]
;;   [:array [:double "77.7"] [:double "-66."]]
;;   [:identifier "a-b_c3$d"]
;;   [:map
;;    [:map-key [:keyword "d"]]
;;    [:map-val [:integer "999"]]
;;    [:map-key [:keyword "e-e9_$e"]]
;;    [:map-val [:identifier "fff"]]]
;;   [:set [:identifier "x"] [:integer "-0"]]
;;   [:quoted [:list [:identifier "de-fg"] [:identifier "hi$"]]]
;;   [:identifier "j"]
;;   [:string "s" "t" "r" "i" "n" "g" " " "s" "t" "r" "i" "n" "g"]
;;   [:string]]]

コメントと全角文字列のIDへの対応が残ってる...。