Question

I would like to extract some information from a string in Ruby by only reading the String once (O(n) time complexity).

Here is an example:

The string looks like this: -location here -time 7:30pm -activity biking

I have a Ruby object I want to populate with this info. All the keywords are known, and they are all optional.

def ActivityInfo
  _attr_reader_ :location, :time, :activity

  def initialize(str)
    @location, @time, @activity = DEFAULT_LOCATION, DEFAULT_TIME, DEFAULT_ACTIVITY

    # Here is how I was planning on implementing this
    current_string = ""
    next_parameter = nil # A reference to keep track of which parameter the current string is refering to
    words = str.split
    while !str.empty?
      word = str.shift
      case word
      when "-location"
        if !next_parameter.nil?
          next_parameter.parameter = current_string # Set the parameter value to the current_string
          current_string = ""
        else
        next_parameter = @location
      when "-time"
        if !next_parameter.nil?
          next_parameter.parameter = current_string
          current_string = ""
        else
        next_parameter = @time
      when "-activity"
        if !next_parameter.nil?
          next_parameter.parameter = current_string
          current_string = ""
        else
        next_parameter = @time
      else
        if !current_string.empty?
          current_string += " "
        end
        current_string += word
      end
    end
  end   
end

So basically I just don’t know how to make a variable be the reference of another variable or method, so that I can then set it to a specific value. Or maybe there is just another more efficient way to achieve this?

Thanks!

Answer 1

The string looks suspiciously like a command-line, and there are some good Ruby modules to parse those, such as optparse.

Assuming it’s not, here’s a quick way to parse the commands in your sample into a hash:

cmd = '-location here -time 7:30pm -activity biking'
Hash[*cmd.scan(/-(\w+) (\S+)/).flatten]

Which results in:

{
    "location" => "here",
        "time" => "7:30pm",
    "activity" => "biking"
}

Expanding it a bit farther:

class ActivityInfo
  def initialize(h)
    @location = h['location']
    @time     = h['time'    ]
    @activity = h['activity']
  end
end
act = ActivityInfo.new(Hash[*cmd.scan(/-(\w+) (\S+)/).flatten])

Which sets act to an instance of ActivityInfo looking like:

#<ActivityInfo:0x101142df8
    @activity = "biking",
    @location = "here",
    @time = "7:30pm"
>

—

The OP asked how to deal with situations where the commands are not flagged with - or are multiple words. These are equivalent, but I prefer the first stylistically:

irb(main):003:0> cmd.scan(/-((?:location|time|activity)) \s+ (\S+)/x)
[
    [0] [
        [0] "location",
        [1] "here"
    ],
    [1] [
        [0] "time",
        [1] "7:30pm"
    ],
    [2] [
        [0] "activity",
        [1] "biking"
    ]
]

irb(main):004:0> cmd.scan(/-(location|time|activity) \s+ (\S+)/x)
[
    [0] [
        [0] "location",
        [1] "here"
    ],
    [1] [
        [0] "time",
        [1] "7:30pm"
    ],
    [2] [
        [0] "activity",
        [1] "biking"
    ]
]

If the commands are multiple words, such as “at location”:

irb(main):009:0> cmd = '-at location here -time 7:30pm -activity biking'
"-at location here -time 7:30pm -activity biking"
irb(main):010:0> 
irb(main):011:0* cmd.scan(/-((?:at \s location|time|activity)) \s+ (\S+)/x)
[
    [0] [
        [0] "at location",
        [1] "here"
    ],
    [1] [
        [0] "time",
        [1] "7:30pm"
    ],
    [2] [
        [0] "activity",
        [1] "biking"
    ]
]

If you need even more flexibility look at Ruby’s strscan module. You can use that to tear apart a string and find the commands and their parameters.