How do I count the number of instances of particular words in a paragraph?

Question

I'd like to count the number of times a set of words appear in each paragraph in a text file. I am able to count the number of times a set of words appears in an entire text.

It has been suggested to me that my code is really buggy, so I'll just ask what I would like to do, and if you want, you can look at the code I have at the bottom.

So, given that "frequency_count.txt" has the words "apple pear grape melon kiwi" in it, I want to know how often "apple" shows up in each paragraph of a separate file "test_essay.txt", how often pear shows up, etc., and then for these numbers to be printed out in a series of lines of numbers, each corresponding to a paragraph.

For instance:

apple, pear, grape, melon, kiwi
3,5,2,7,8
2,3,1,6,7
5,6,8,2,3

Where each line corresponds to one of the paragraphs.

I am very, very new to Ruby, so thank you for your patience.

output_file = '/Users/yirenlu/Quora-Personal-Analytics/weka_input6.csv'
o = File.open(output_file, "r+")

common_words = '/Users/yirenlu/Quora-Personal-Analytics/frequency_count.txt'
c = File.open(common_words, "r")

c.each_line{|$line1|
    words1 = $line1.split
    words1.each{|w1|
        the_file = '/Users/yirenlu/Quora-Personal-Analytics/test_essay.txt'
        f = File.open(the_file, "r")
        rows = File.readlines("/Users/yirenlu/Quora-Personal-Analytics/test_essay.txt")
        text = rows.join
        paragraph = text.split(/\n\n/)
        paragraph.each{|p|
            h = Hash.new
            puts "this is each paragraph"
            p.each_line{|line|
                puts "this is each line"
                words = line.split
                words.each{|w|
                    if w1 == w
                        if h.has_key?(w)
                            h[w1] = h[w1] + 1
                        else
                            h[w1] = 1
                        end
                        $x = h[w1]
                    end
                }
            }
            o.print "#{$x},"
        }
    }
    o.print "\n"
    o.print "#{$line1}"
}

Answer 1

If you're used to PHP or Perl you may be under the impression that a variable like $line1 is local, but this is a global. Use of them is highly discouraged and the number of instances where they are strictly required is very short. In most cases you can just omit the $ and use variables that way with proper scoping.

This example also suffers from nearly unreadable indentation, though perhaps that was an artifact of the cut-and-paste procedure.

Generally what you want for counters is to create a hash with a default of zero, then add to that as required:

# Create a hash where the default values for each key is 0
counter = Hash.new(0)

# Add to the counters where required
counter['foo'] += 1
counter['bar'] += 2

puts counter['foo']
# => 1
puts counter['baz']
# => 0

You basically have what you need, but everything is all muddled and just needs to be organized better.

Answer 2

Here are two one-liners to calculate frequencies of words in a string.

The first one is a bit easier to understand, but it's less effective:

txt.scan(/\w+/).group_by{|word| word.downcase}.map{|k,v| [k, v.size]}
# => [['word1', 1], ['word2', 5], ...]

The second solution is:

txt.scan(/\w+/).inject(Hash.new(0)) { |hash, w| hash[w.downcase] += 1; hash}
# => {'word1' => 1, 'word2' => 5, ...}

Answer 3

This could be shorter and easier to read if you use:

1. The CSV library.
2. A more functional approach using map and blocks.

require 'csv'

common_words = %w(apple pear grape melon kiwi)
text = File.open("test_essay.txt").read

def word_frequency(words, text)
  words.map { |word| text.scan(/\b#{word}\b/).length }
end

CSV.open("file.csv", "wb") do |csv|
  paragraphs = text.split /\n\n/
  paragraphs.each do |para| 
    csv << word_frequency(common_words, para)
  end
end

Note this is currently case-sensitive but it's a minor adjustment if you want case-insensitivity.

Answer 4

To count how many times one word appears in a text:

text = "word aaa word word word bbb ccc ccc"
text.scan(/\w+/).count("word") # => 4

To count a set of words:

text = "word aaa word word word bbb ccc ccc"
wlist = text.scan(/\w+/)
wset = ["word", "ccc"]
result = {}
wset.each {|word| result[word] = wlist.count(word) }
result # => {"word" => 4, "ccc" => 2}
result["ccc"] # => 2

Answer 5

What about this:

# Create an array of regexes to be used in `scan' in the loop.
# `\b' makes sure that `barfoobar' does not match `bar' or `foo'.
p word_list = File.open("frequency_count.txt"){|io| io.read.scan(/\w+/)}.map{|w| /\b#{w}\b/}
File.open("test_essay.txt") do |io|
    loop do
        # Add lines to `paragraph' as long as there is a continuous line
        paragraph = ""
        # A `l.chomp.empty?' becomes true at paragraph border
        while l = io.gets and !l.chomp.empty?
            paragraph << l
        end
        p word_list.map{|re| paragraph.scan(re).length}
        # The end of file has been reached when `l == nil'
        break unless l
    end
end

Answer 6

Here's an alternate answer, which is has been tweaked for conciseness (though not as easy to read as my other answer).

require 'csv'

words = %w(apple pear grape melon kiwi)
text = File.open("test_essay.txt").read

CSV.open("file.csv", "wb") do |csv|
  text.split(/\n\n/).map {|p| csv << words.map {|w| p.scan(/\b#{w}\b/).length}}
end

I prefer the slightly longer but more self-documenting code, but it's fun to see how small it can get.