TheLazy.Dev
Portfolio GitHub

Blog

machinelearning |datascience |rails |ruby

Investigating Machine Learning Techniques to Improve Spec Tests — IV

kagermanov27

Emirhan Akdeniz

Posted on April 21, 2022

Investigating Machine Learning Techniques to Improve Spec Tests — IV

Intro

This is a part of the series of blog posts related to Artificial Intelligence Implementation. If you are interested in the background of the story or how it goes:

Previous blog post links
How to scrape Google Local Results with Artificial Intelligence
Real World Example of Machine Learning on Rails
AI Training Tips and Comparisons
Machine Learning in Scraping with Rails
Implementing ONNX models in Rails
How ML Hybrid Parser Beats Traditional Parser
How to Benchmark ML Implementations on Rails
Investigating Machine Learning Techniques to Improve Spec Tests
Investigating Machine Learning Techniques to Improve Spec Tests - II
Investigating Machine Learning Techniques to Improve Spec Tests - III

This week we'll showcase testing process and the early results of the model. We will be using SerpApi's Google Organic Results Scraper API for the data collection. Also, you can check in the playground in more detailed view on the data we will use.

image

Training Data

Here's an structural breakdown of the data we store for training inside a json file:

[
  { 
    "Key 1": Value_1,
    "Key 2": Value_2,
    "Key 3": Value_3,
    "Key 4": [
      "Value_1",
      ...
    ],
    "Key 5": {
      "Inner Key 1": Inner_Value_1,
      ...
  },
  ...
]
Enter fullscreen mode Exit fullscreen mode

Here's an example:

[
  {
    "position": 1,
    "title": "Coffee - Wikipedia",
    "link": "https://en.wikipedia.org/wiki/Coffee",
    "displayed_link": "https://en.wikipedia.org › wiki › Coffee",
    "snippet": "Coffee is a brewed drink prepared from roasted coffee beans, the seeds of berries from certain flowering plants in the Coffea genus. From the coffee fruit, ...",
    "snippet_highlighted_words": [
      "Coffee",
      "coffee",
      "coffee"
    ],
    ...
  },
  ...
]
Enter fullscreen mode Exit fullscreen mode

Links we collected the organic results of Google from:
Link for Tea (around 100 results)
Link for Coffee (around 100 results)

Testing Structure

We have already covered how we trained the data in detail in the past three week's blog posts. Today, we will test how the hypothesis holds by calculating the training accuracy.

We can reutilize the Train, and Database classes to create examples, and create example vectors with the following lines:

  example_vector = Database.word_to_tensor example
  example_vector.map! {|el| el = el.nil? ? 0: el}
  example_vector = Train.extend_vector example_vector
  weighted_example = Train.product example_vector
Enter fullscreen mode Exit fullscreen mode

,
example in here is the string we provide. Any value for any key within Google Organic Results that is converted to a string will be a valid example.

We can reutilize Database.word_to_tensor to get the vectorized version of our string in accordance with our vocabulary.

If any value is nil (null), which is not present in our vocabulary, it will be replaced with 0, which is the value for our <unk> (unknown).
example_vector, then, should be expanded to maximum string size for calculation purposes using 1s.
weighted_example will be the product of the @@weights we calculated earlier with our vectorized example.

This value's closest vectors in multidimensional space, from the examples we provided, should have the same key, or their average should lead us to the same key. So, in our case, if the example we provide isn't a snippet, closest vectors around the weighted_example should give us less than 0.5 (their identities are 0 and 1) in average. Conclusion should be that the example isn't a snippet.

We measure the distance of our example with every example in the dataset using Euclidean Distance formula for multidimensional space:

  distances = []
  vector_array.each_with_index do |comparison_vector, vector_index|
    distances << Train.euclidean_distance(comparison_vector, weighted_example)
  end
Enter fullscreen mode Exit fullscreen mode

We take the indexes of the minimum distances (k many times):

  indexes = []
  k.times do 
    index = distances.index(distances.min)
    indexes << index
    distances[index] = 1000000000
  end
Enter fullscreen mode Exit fullscreen mode

Then, we take the real identities of each of these vectors:

  predictions = []
  indexes.each do |index|
    predictions << key_array[index].first.to_i
  end
Enter fullscreen mode Exit fullscreen mode

key_array here is the array containing 0, or 1 in first item of each row, and the string in second. To give an example:

[
  ...
  ["0", "https://www.coffeebean.com"],
  ["1", "Born and brewed in Southern California since 1963, The Coffee Bean & Tea Leaf® is passionate about connecting loyal customers with carefully handcrafted ..."],
  ["0", "4"],
  ...
]
Enter fullscreen mode Exit fullscreen mode

1 represents that the item is snippet, 0 represents it isn't.

Let's return the predictions:

  prediction = (predictions.sum/predictions.size).to_f
  if prediction < 0.5
    puts "False - Item is not Snippet"
    return 0
  else
    puts "True - Item is Snippet"
    return 1
  end
Enter fullscreen mode Exit fullscreen mode

Here's the full method for it:

def test example, k, vector_array, key_array
  example_vector = Database.word_to_tensor example
  example_vector.map! {|el| el = el.nil? ? 0: el}
  example_vector = Train.extend_vector example_vector
  weighted_example = Train.product example_vector

  distances = []
  vector_array.each_with_index do |comparison_vector, vector_index|
    distances << Train.euclidean_distance(comparison_vector, weighted_example)
  end

  indexes = []
  k.times do 
    index = distances.index(distances.min)
    indexes << index
    distances[index] = 1000000000
  end

  predictions = []
  indexes.each do |index|
    predictions << key_array[index].first.to_i
  end

  puts "Predictions: #{predictions}"

  prediction = (predictions.sum/predictions.size).to_f
  if prediction < 0.5
    puts "False - Item is not Snippet"
    return 0
  else
    puts "True - Item is Snippet"
    return 1
  end
end
Enter fullscreen mode Exit fullscreen mode

Testing with Google Organic Results for Snippet

Now that we have a function for testing, let's separate snippets from non-snippets in our examples:

true_examples = key_array.map {|el| el = el.first == "1" ? el.second : nil}.compact
false_examples = key_array.map {|el| el = el.first == "0" ? el.second : nil}.compact
Enter fullscreen mode Exit fullscreen mode

This will allow us to calculate easier.

Let's declare an empty array to collect predictions, and start with non-snippets:

predictions = []

false_examples.each do |example|
  prediction = test example, 2, vector_array, key_array
  predictions << prediction
end

predictions.map! {|el| el = el == 1 ? 0 : 1}
Enter fullscreen mode Exit fullscreen mode

Since we know that none of these examples are snippet, any prediction that gives 1 will be wrong. So if we test our model with false examples, and then reverse 1s to 0s, and 0s to 1s, we can combine it with our true examples:

true_examples.each do |example|
  prediction = test example, 2, vector_array, key_array
  predictions << prediction
end
Enter fullscreen mode Exit fullscreen mode

Now that we have the desired array filled:

prediction_train_accuracy = predictions.sum.to_f / predictions.size.to_f

puts "Prediction Accuracy for Training Set is: #{prediction_train_accuracy}"
Enter fullscreen mode Exit fullscreen mode

If we divide the number of 1s to number of predictions, we can calculate the accuracy results.

Preliminary Results

We have done exactly the same process for the data we mentioned earlier. The number of predictions for snippet was 1065, and the k value was 2, and the n-gram value was 2.

The model predicted 872 times correctly. This means the training accuracy was 0.8187793427230047 (%81.87).

This is a good number to start, and with more tweaks, and testing with a bigger dataset, the initial hypothesis could be proven to be true.

Full Code 

class Database
  def initialize json_data, vocab = { "<unk>" => 0, "<pad>" => 1 }
    super()
    @@pattern_data = []
    @@vocab = vocab
  end

  ## Related to creating main database
  def self.add_new_data_to_database json_data, csv_path = nil
    json_data.each do |result|
      recursive_hash_pattern result, ""
    end

    @@pattern_data = @@pattern_data.reject { |pattern| pattern.include? nil }.uniq.compact

    path = "#{csv_path}master_database.csv"
    File.write(path, @@pattern_data.map(&:to_csv).join)
  end

  def self.element_pattern result, pattern
    @@pattern_data.append([result, pattern].flatten)
  end

  def self.element_array_pattern result, pattern
    result.each do |element|
      element_pattern element, pattern
    end
  end

  def self.assign hash, key, pattern
    if hash[key].is_a?(Hash)
      if pattern.present?
        pattern = "#{pattern}__#{key}"
      else
        pattern = "#{key}"
      end

      recursive_hash_pattern hash[key], pattern
    elsif hash[key].present? && hash[key].is_a?(Array) && hash[key].first.is_a?(Hash)
      if pattern.present?
        pattern = "#{pattern}__#{key}__n"
      else
        pattern = "#{key}"
      end

      hash[key].each do |hash_inside_array|
        recursive_hash_pattern hash_inside_array, pattern
      end
    elsif hash[key].present? && hash[key].is_a?(Array)
      if pattern.present?
        pattern = "#{pattern}__n"
      else
        pattern = "#{key}"
      end

      element_array_pattern hash[key], pattern
    else
      if pattern.present?
        pattern = "#{pattern}__#{key}"
      else
        pattern = "#{key}"
      end

      element_pattern hash[key], pattern
    end
  end

  def self.recursive_hash_pattern hash, pattern
    hash.keys.each do |key|
      assign hash, key, pattern
    end
  end

  ## Related to tokenizing
  def self.default_dictionary_hash
    {
      /\"/ => "",
      /\'/ => " \'  ",
      /\./ => " . ",
      /,/ => ", ",
      /\!/ => " ! ",
      /\?/ => " ? ",
      /\;/ => " ",
      /\:/ => " ",
      /\(/ => " ( ",
      /\)/ => " ) ",
      /\// => " / ",
      /\s+/ => " ",
      /<br \/>/ => " , ",
      /http/ => "http",
      /https/ => " https ",
    }
  end

  def self.tokenizer word, dictionary_hash = default_dictionary_hash
    word = word.downcase

    dictionary_hash.keys.each do |key|
      word.sub!(key, dictionary_hash[key])
    end

    word.split
  end

  def self.iterate_ngrams token_list, ngrams = 2
    token_list.each do |token|
      1.upto(ngrams) do |n|
        permutations = (token_list.size - n + 1).times.map { |i| token_list[i...(i + n)] }

        permutations.each do |perm|
          key = perm.join(" ")

          unless @@vocab.keys.include? key
            @@vocab[key] = @@vocab.size
          end
        end
      end
    end
  end

  def self.word_to_tensor word
    token_list = tokenizer word
    token_list.map {|token| @@vocab[token]}
  end

  ## Related to creating key-specific databases 
  def self.create_key_specific_databases result_type = "organic_results", csv_path = nil, dictionary = nil, ngrams = nil, vocab_path = nil
    keys, examples = create_keys_and_examples

    keys.each do |key|
      specific_pattern_data = []
      @@pattern_data.each_with_index do |pattern, index|
        word = pattern.first.to_s

        next if word.blank?

        if dictionary.present?
          token_list = tokenizer word, dictionary
        else
          token_list = tokenizer word
        end

        if ngrams.present?
          iterate_ngrams token_list, ngrams
        else
          iterate_ngrams token_list
        end

        if key == pattern.second
          specific_pattern_data << [ 1, word ]
        elsif (examples[key].to_s.to_i == examples[key]) && word.to_i == word
          next
        elsif (examples[key].to_s.to_i == examples[key]) && word.numeric?
          specific_pattern_data << [ 0, word ]
        elsif examples[key].numeric? && word.numeric?
          next
        elsif key.split("__").last == pattern.second.to_s.split("__").last
          specific_pattern_data << [ 1, word ]
        else
          specific_pattern_data << [ 0, word ]
        end
      end

      path = "#{csv_path}#{result_type}__#{key}.csv"
      File.write(path, specific_pattern_data.map(&:to_csv).join)
    end

    if vocab_path.present?
      save_vocab vocab_path
    else
      save_vocab
    end
  end

  def self.create_keys_and_examples
    keys = @@pattern_data.map { |pattern| pattern.second }.uniq

    examples = {}
    keys.each do |key|
      examples[key] = @@pattern_data.find { |pattern| pattern.first.to_s if pattern.second == key }
    end

    [keys, examples]
  end

  def self.numeric?
    return true if self =~ /\A\d+\Z/
    true if Float(self) rescue false
  end

  def self.save_vocab vocab_path = ""
    path = "#{vocab_path}vocab.json"
    vocab = JSON.parse(@@vocab.to_json)
    File.write(path, JSON.pretty_generate(vocab))
  end

  def self.read_vocab vocab_path
    vocab = File.read vocab_path
    @@vocab = JSON.parse(vocab)
  end

  def self.return_vocab
    @@vocab
  end
end

class Train
  def initialize csv_path
    @@csv_path = csv_path
    @@vector_arr = []
    @@word_arr = []
    @@maximum_word_size = 100
    @@weights = Vector[]
    @@losses = []
  end

  def self.read
    @@word_arr = CSV.read(@@csv_path)
    @@word_arr
  end

  def self.define_training_set vectors
    @@vector_arr = vectors
  end

  def self.auto_define_maximum_size
    @@maximum_word_size = @@vector_arr.map {|el| el.size}.max
  end

  def self.extend_vector vector
    vector_arr = vector.to_a
    (@@maximum_word_size - vector.size).times { vector_arr << 1 }
    Vector.[](*vector_arr)
  end

  def self.extend_vectors
    @@vector_arr.each_with_index do |vector, index|
      @@vector_arr[index] = extend_vector vector
    end
  end

  def self.initialize_weights
    weights = []
    @@maximum_word_size.times { weights << 1.0 }
    @@weights = Vector.[](*weights)
  end

  def self.config k = 1, lr = 0.001
    [k, lr]
  end

  def self.product vector
    @@weights.each_with_index do |weight, index|
      vector[index] = weight * vector[index]
    end

    vector
  end

  def self.euclidean_distance vector_1, vector_2
    subtractions = (vector_1 - vector_2).to_a
    subtractions.map! {|sub| sub = sub*sub }
    Math.sqrt(subtractions.sum)
  end

  def self.k_neighbors distances, k
    indexes = []
    (k).times do
      min = distances.index(distances.min)
      indexes << min
      distances[min] = distances.max + 1
    end

    indexes
  end

  def self.make_prediction indexes
    predictions = []
    indexes.each do |index|
      predictions << @@word_arr[index][0].to_i
    end

    predictions.sum/predictions.size
  end

  def self.update_weights result, indexes, vector, lr
    indexes.each do |index|
      subtractions = @@vector_arr[index] - vector
      subtractions.each_with_index do |sub, sub_index|
        if result == 0 && sub >= 0
          @@weights[sub_index] = @@weights[sub_index] + lr
        elsif result == 0 && sub < 0
          @@weights[sub_index] = @@weights[sub_index] - lr
        elsif result == 1 && sub >= 0
          @@weights[sub_index] = @@weights[sub_index] - lr
        elsif result == 1 && sub < 0
          @@weights[sub_index] = @@weights[sub_index] + lr
        end
      end
    end
  end

  def self.mean_absolute_error real, indexes
    errors = []
    indexes.each do |index|
      errors << (@@word_arr[index][0].to_i - real).abs
    end

    (errors.sum/errors.size).to_f
  end

  def self.train vector, index
    k, lr = config
    vector = extend_vector vector
    vector = product vector

    distances = []
    @@vector_arr.each_with_index do |comparison_vector, vector_index|
      if vector_index == index
        distances << 100000000
      else
        distances << euclidean_distance(comparison_vector, vector)
      end
    end

    indexes = k_neighbors distances, k
    real = @@word_arr[index][0].to_i
    prob_prediction = make_prediction indexes
    prediction = prob_prediction > 0.5 ? 1 : 0
    result = real == prediction ? 1 : 0

    update_weights result, indexes, vector, lr
    loss = mean_absolute_error real, indexes
    @@losses << loss

    puts "Result : #{real}, Prediction: #{prediction}"
    puts "Loss: #{loss}"

    prediction
  end
end


json_path = "organic_results/example.json"
json_data = File.read(json_path)
json_data = JSON.parse(json_data)

Database.new json_data
## For training from scratch                     
Database.add_new_data_to_database json_data, csv_path = "organic_results/"
Database.create_key_specific_databases result_type = "organic_results", csv_path = "organic_results/"
##

Database.read_vocab "vocab.json"

## We will use an iteration of csvs within a specific path in the end
csv_path = "organic_results/organic_results__snippet.csv"

Train.new csv_path
key_array = Train.read

vector_array = key_array.map { |word| Database.word_to_tensor word[1] }
Train.define_training_set vector_array
Train.auto_define_maximum_size
Train.extend_vectors
Train.initialize_weights
Train.config k = 2

vector_array.each_with_index do |vector, index|
  Train.train vector, index
end

def test example, k, vector_array, key_array
  example_vector = Database.word_to_tensor example
  example_vector.map! {|el| el = el.nil? ? 0: el}
  example_vector = Train.extend_vector example_vector
  weighted_example = Train.product example_vector

  distances = []
  vector_array.each_with_index do |comparison_vector, vector_index|
    distances << Train.euclidean_distance(comparison_vector, weighted_example)
  end

  indexes = []
  k.times do 
    index = distances.index(distances.min)
    indexes << index
    distances[index] = 1000000000
  end

  predictions = []
  indexes.each do |index|
    predictions << key_array[index].first.to_i
  end

  puts "Predictions: #{predictions}"

  prediction = (predictions.sum/predictions.size).to_f
  if prediction < 0.5
    puts "False - Item is not Snippet"
    return 0
  else
    puts "True - Item is Snippet"
    return 1
  end
end

true_examples = key_array.map {|el| el = el.first == "1" ? el.second : nil}.compact
false_examples = key_array.map {|el| el = el.first == "0" ? el.second : nil}.compact

predictions = []

false_examples.each do |example|
  prediction = test example, 2, vector_array, key_array
  predictions << prediction
end

predictions.map! {|el| el = el == 1 ? 0 : 1}

true_examples.each do |example|
  prediction = test example, 2, vector_array, key_array
  predictions << prediction
end

prediction_train_accuracy = predictions.sum.to_f / predictions.size.to_f

puts "Prediction Accuracy for Training Set is: #{prediction_train_accuracy}"
Enter fullscreen mode Exit fullscreen mode

Conclusion

I'd like to apologize the reader for being one day late on the blog post. Two weeks later, we will showcase how to store them for implementation, and further tweaks to improve accuracy.

The end aim of this project is to create an open-source gem to be implemented by everyone using a JSON Data Structure in their code.

I'd like to thank the reader for their attention, and the brilliant people of SerpApi creating wonders even in times of hardship, and for all their support.

💖 💪 🙅 🚩
kagermanov27
Emirhan Akdeniz

Posted on April 21, 2022

Join Our Newsletter. No Spam, Only the good stuff.

Sign up to receive the latest update from our blog.

Related

Investigating Machine Learning Techniques to Improve Spec Tests — IV
machinelearning Investigating Machine Learning Techniques to Improve Spec Tests — IV

April 21, 2022