CODEWITHSUNDEEP
ruby
farhany
farhany

Reputation: 1541

Read variables from file and replace

I have an Apache2 server side include file. It looks like (actual data redacted):

<!--#set var="FIRST_VAR" value="HI" -->
<!--#set var="SECOND_VAR" value"THERE" -->
<!--#set var="FINAL_VAR" value="HEY/${FIRST_VAR}/${SECOND_VAR}/${FINAL_VAR}" -->

What I'd like to do is to do a regex match with named variables, and then build the nested variables using a hash, perhaps. Example:

@SSI["FIRST_VAR"] = "HI"
@SSI["SECOND_VAR"] = "THERE"
@SSI["FINAL_VAR"] = "HEY/HI/THERE"

I'm at a loss on how to do this properly.

Upvotes: 1

Views: 57

Answers (2)

Cary Swoveland
Cary Swoveland

Reputation: 110725

Suppose you read the file into a string, separate the string into lines and strip whitespace from the ends of each line.

arr =<<_.lines.map(&:strip)
<!--#set var="V1" value="HI" -->
<!--#set var="V2" value="THERE" -->
<!--#set var="FV" value="HEY/${V1}/${V2}" -->
_
  #=> ["<!--#set var=\"V1\" value=\"HI\" -->",
  #    "<!--#set var=\"V2\" value=\"THERE\" -->",
  #    "<!--#set var=\"FV\" value=\"HEY/${V1}/${V2}\" -->"]

I will use the following regular expressions.

r1 = /
     (?<=\svar=\")   # match ' var="' in a positive lookbehind
     .+?             # match one or more characters lazily
     (?=\")          # match '"' in a positive lookahead
     /x              # free-spacing regex definition mode

r2 = /
     (?<=\svalue=\") # match ' value"' in a positive lookbehind
     .+?             # match one or more characters lazily
     (?=\")          # match '"' in a positive lookahead
     /x              # free-spacing regex definition mode

r3 = /
     \/\$\{          # match '/${'
     \w+             # match one or more word characters
     \}              # match '}'
     /x              # free-spacing regex definition mode

r4 = /
     (?<=\/\$\{)     # match '/${' in a positive lookbehind
     .+?             # match any number of character, lazily
     (?=\})          # match `}` in a positive lookahead
     /x              # free-spacing regex definition mode

We can then compute the desired return value as follows.

g = arr.each_with_object({}) do |s,h|
  var   = s[r1].strip
  value = s[r2].strip
  h["\/\$\{#{var}\}"] = value.gsub(r3) { |s| h.key?(s) ? "/#{h[s]}" : "" }
end
  #=> {"/${V1}"=>"HI", "/${V2}"=>"THERE", "/${FV}"=>"HEY/HI/THERE"}

Lastly, modify the keys.

g.each_with_object({}) { |(k,v),h| h[k[r4]] = v }
  #=> {"V1"=>"HI", "V2"=>"THERE", "FV"=>"HEY/HI/THERE"}

The two above expressions could of course be chained. In fact, it could be written as a single line, but I wouldn't recommend it.

Upvotes: 2

max pleaner
max pleaner

Reputation: 26778

If you have a string equal to the text of the file:

txt = <<-TXT
<!--#set var="FIRST_VAR" value="HI" -->
<!--#set var="SECOND_VAR" value="THERE" -->
<!--#set var="FINAL_VAR" value="HEY/${FIRST_VAR}/${SECOND_VAR}/${FINAL_VAR}" -->
TXT

(also note I've added an = after value in the second line)

Then you can build a regex like so:

result = txt.scan /<!--#set var=\"(.+)\" value=\"(.+)\" -->/
# =>
# [["FIRST_VAR", "HI"],
#  ["SECOND_VAR", "THERE"],
#  ["FINAL_VAR", "HEY/${FIRST_VAR}/${SECOND_VAR}/${FINAL_VAR}"]
# ]

The regex is basically the same format as that contained in the file, except quotations are escaped and the match groups are defined using (.+).

Although regex is probably better, you could also take a more brute-force style approach using split:

txt.split("\n").map do |line|
  var, rest = line.split("var=\"")[1].split("\" value=\"")
  value = rest.split("\"")[0]
  [var, value]
end

Upvotes: 2

Related Questions