API tools faq
paste
devinteske

Parsing JSON with awk

devinteske
Apr 18th, 2016
482
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
Awk 7.57 KB | None | 0 0
raw download clone embed print report
  1. #!/usr/bin/awk -f
  2. BEGIN {
  3.     nkeys[depth = 0] = keynum = 0
  4.     building_array = building_hash = building_string = 0
  5.     if (!(valid_chars = ENVIRON["VALID_VARNAME_CHARS"])) {
  6.         for (c = 48; c < 58; c++)
  7.             valid_chars = valid_chars sprintf("%c", c)
  8.         for (c = 65; c < 91; c++)
  9.             valid_chars = valid_chars sprintf("%c", c)
  10.         for (c = 97; c < 123; c++)
  11.             valid_chars = valid_chars sprintf("%c", c)
  12.     }
  13.     if (filter_object) filter_object = "^" filter_object "$"
  14.     if (filter_value) filter_value = "^" filter_value "$"
  15. }
  16. ############################################### FUNCTIONS
  17. function dprint(msg) { if (debug) print ": DEBUG1", msg }
  18. function dprint2(msg) { if (debug >= 2) print ": DEBUG2", msg }
  19. function dprint3(msg) { if (debug >= 3) print ": DEBUG3", msg }
  20. function lprint(line) { print (local ? "local " : "") line }
  21. function trim_match()
  22. {
  23.     match_text = substr($0, RSTART, RLENGTH)
  24.     $0 = substr($0, RSTART + RLENGTH)
  25.     return match_text
  26. }
  27. function trim_keyword()
  28. {
  29.     keylen = length(keynum > 0 ? keynum : key[depth,nkeys[depth]])
  30.     keynum = 0
  31.     if (!keylen) return
  32.     x = keyword
  33.     keyword = substr(keyword, 0, length(keyword) - keylen - 1)
  34.         # NB: The "-1" is for keyword separator "_"
  35.     dprint2(sprintf("TRIM KEY ([%s] => [%s])", x, keyword))
  36. }
  37. function objsafe(name)
  38. {
  39.     gsub(("[^" valid_chars "]"), "_", name)
  40.     return name
  41. }
  42. function json_print(object, value)
  43. {
  44.     if (object !~ filter_object) return
  45.     if (value !~ filter_value) return
  46.     gsub(/'/, "&\\\\&&", value)
  47.     object = objsafe(object)
  48.     lprint(object "='" value "'")
  49. }
  50. function json_print_type(object, type)
  51. {
  52.     if (!keyword) return
  53.     if (object !~ filter_object) return
  54.     if (type)
  55.         lprint(object "_type=" type)
  56.     else if (building_array)
  57.         lprint(object "_type=array")
  58.     else if (building_hash)
  59.         lprint(object "_type=hash")
  60.     else
  61.         lprint(object "_type=scalar")
  62. }
  63. function json_print_keys(object, depth)
  64. {
  65.     keys = ""
  66.     for (k = 1; k <= nkeys[depth]; k++) {
  67.         if (printed[key[depth,k]]) continue
  68.         printed[key[depth,k]] = 1
  69.         keys = keys " " key[depth,k]
  70.     }
  71.     json_print(object, substr(keys, 2))
  72. }
  73. function json_unset_value(object)
  74. {
  75.     lprint(objsafe(object) "_value=")
  76. }
  77. function json_filtered_unset_value(object)
  78. {
  79.     if (object !~ filter_object) return
  80.     return json_unset_value(object)
  81. }
  82. function json_objname()
  83. {
  84.     objname = ""
  85.     if (building_array) {
  86.         objname = keyword "_" building_array++
  87.         dprint3(sprintf("RETURN ARRAY OBJNAME [%s]", objname))
  88.     } else if (building_hash) {
  89.         objname = keyword "_" building_hash++
  90.         dprint3(sprintf("RETURN HASH OBJNAME [%s]", objname))
  91.     } else if (building_string) {
  92.         objname = keyword
  93.         dprint3(sprintf("RETURN STRING OBJNAME [%s]", objname))
  94.     } else if (depth <= 1 && keyword) {
  95.         objname = keyword
  96.         dprint3(sprintf("RETURN ROOT OBJNAME [%s]", objname))
  97.     }
  98.  
  99.     if (objname) return objname
  100.  
  101.     # NB: If non-NULL current-depth key, increment for separator
  102.     if ((keylen = length(key[depth,nkeys[depth]]))) keylen++
  103.  
  104.     objname = substr(keyword, 0, length(keyword) - keylen) "_" nkeys[depth]
  105.     dprint3(sprintf("RETURN OBJNAME [%s]", objname))
  106.     return objname
  107. }
  108. ############################################### MAIN LOOP
  109. { while ($0) { # Loop until done processing everything on this line
  110.     if (building_string) {
  111.         while (match($0, /^[^"]*\\"/)) value = value trim_match()
  112.         if (!match($0, /^[^"]*"/)) { # No ending quote
  113.             value = value $0
  114.             next # Continue reading on next line
  115.         }
  116.         objname = json_objname()
  117.         building_string = 0
  118.         value = value substr($0, RSTART, RLENGTH - 1)
  119.         trim_match()
  120.         sub(/^[[:space:]]*,[[:space:]]*/, "")
  121.         json_print(objname, value)
  122.         json_print_type(objname)
  123.         trim_keyword()
  124.         if (depth <= 1) json_print_type(keyword "_" nk)
  125.     }
  126.     ################################### OPENING PATTERNS
  127.     else if (match($0, /^[[:space:]]*{[[:space:]]*/)) {
  128.         building_hash = (depth > 0)
  129.         if (depth > 1) {
  130.             x = keyword
  131.             nk = nkeys[depth]
  132.             keyword = keyword "_" nk
  133.             keynum = nk
  134.             dprint2(sprintf("APPEND HASH ([%s] => [%s])",
  135.                 x, keyword))
  136.         }
  137.         nkeys[++depth] = 0
  138.         trim_match()
  139.     } else if (keyword && match($0, /^[[:space:]]*\[/)) {
  140.         building_array = 1
  141.         trim_match()
  142.     }
  143.     ################################### OBJECTS
  144.     else if (match($0, /^[[:space:]]*"[^"]+"[[:space:]]*:[[:space:]]*/)) {
  145.         nkeys[depth]++
  146.         nk = nkeys[depth]
  147.         key[depth,nk] = trim_match()
  148.         sub(/^[[:space:]]*"/, "", key[depth,nk])
  149.         sub(/"[[:space:]]*:[[:space:]]*$/, "", key[depth,nk])
  150.         kp = key[depth,nk]
  151.         dprint(sprintf("START OBJECT key[%u,%u]=\"%s\"",
  152.             depth, nk, kp))
  153.         if (keyword && !building_hash)
  154.             json_print(keyword "_" nk, objsafe(kp))
  155.         if (!building_hash) {
  156.             x = keyword
  157.             keyword = keyword (keyword ? "_" : "") kp
  158.             dprint2(sprintf("APPEND KEY ([%s] => [%s])",
  159.                 x, keyword))
  160.         }
  161.     }
  162.     ################################### PROPERTIES
  163.     else if (keyword && match($0, /^[[:space:]]*"/)) {
  164.         value = ""
  165.         trim_match()
  166.         while (match($0, /^[^"]*\\"/)) value = value trim_match()
  167.         if (!match($0, /^[^"]*"/)) {
  168.             building_string = 1
  169.             value = $0
  170.             next
  171.         }
  172.         value = value substr($0, RSTART, RLENGTH - 1)
  173.         trim_match()
  174.         sub(/^[[:space:]]*,[[:space:]]*/, "")
  175.  
  176.         object = json_objname()
  177.         nk = nkeys[depth]
  178.         if (depth <= 1) {
  179.             json_print(object, value)
  180.             if (!building_array) json_print_type(object)
  181.         } else {
  182.             json_print(object, key[depth,nk])
  183.             json_print(object "_value", value)
  184.         }
  185.         if (building_hash) {
  186.             if (!value) json_filtered_unset_value(object)
  187.         } else if (!building_array) {
  188.             trim_keyword()
  189.             json_print_type(keyword "_" nk)
  190.         }
  191.     }
  192.     else if (keyword && match($0, \
  193.         /^[[:space:]]*[^[:space:],}\]]+[[:space:]]*/ \
  194.     )) {
  195.         value = trim_match()
  196.         sub(/^[[:space:]]*/, "", value)
  197.         sub(/[[:space:]]*$/, "", value)
  198.         sub(/^[[:space:]]*,[[:space:]]*/, "")
  199.         json_print(keyword, value)
  200.         json_print_type(keyword)
  201.         trim_keyword()
  202.         if (!building_array) json_print_type(keyword "_" nkeys[depth])
  203.     }
  204.     ################################### CLOSING PATTERNS
  205.     else if (match($0, /^[[:space:]]*\][[:space:]]*/)) {
  206.         json_print(keyword "_len", --building_array)
  207.         json_print_type(keyword, "array")
  208.         nk = nkeys[depth]
  209.         items = building_array
  210.         building_array = 0
  211.         trim_keyword()
  212.         trim_match()
  213.         sub(/^[[:space:]]*,[[:space:]]*/, "")
  214.         json_print_type(keyword "_" nk, "array")
  215.         dprint(sprintf("END ARRAY key[%u,%u]=\"%s\"" \
  216.             " holding %u object%s", depth, nk,
  217.             key[depth,nk], items, items != 1 ? "s" : ""))
  218.     }
  219.     else if (match($0, /^[[:space:]]*}[[:space:]]*/)) {
  220.         nk = nkeys[depth]
  221.         if (keyword && depth >= 1) {
  222.             hashname = key[depth-1,nkeys[depth-1]]
  223.             json_print(keyword, hashname)
  224.             json_print(keyword "_len", nk)
  225.             json_print_type(keyword, "hash")
  226.         }
  227.         building_hash = 0
  228.         depth-- # NB: Done prior to calling trim_keyword()
  229.         trim_keyword()
  230.         trim_match()
  231.         sub(/^[[:space:]]*,[[:space:]]*/, "")
  232.         building_hash = (depth > 1)
  233.         if (depth == 1) {
  234.             objname = keyword "_" nkeys[depth]
  235.             json_print_type(objname, "hash")
  236.         }
  237.         if (depth < 0) {
  238.             json_print_keys(keyword, depth+1)
  239.             dprint(sprintf("END HASH key[%u,%u]=\"%s\"" \
  240.                 " holding %u object%s", depth+2,
  241.                 nkeys[depth+2],
  242.                 key[depth+2,nkeys[depth+2]],
  243.                 nkeys[depth+3],
  244.                 nkeys[depth+3] != 1 ? "s" : ""))
  245.         } else if (kp = key[depth,nkeys[depth]]) {
  246.             dprint(sprintf("END HASH key[%u,%u]=\"%s\"" \
  247.                 " holding %u object%s", depth,
  248.                 nkeys[depth], kp, nkeys[depth+1],
  249.                 nkeys[depth+1] != 1 ? "s" : ""))
  250.         }
  251.     }
  252. } }
  253. ############################################### END
  254. END {
  255.     json_print_keys(keyword "_keys", 1)
  256.     dprint(sprintf("END HASH holding %u object%s", nkeys[1],
  257.         nkeys[1] != 1 ? "s" : ""))
  258. }
Add Comment
Please, Sign In to add comment
Public Pastes
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!