smalltext 0.1.0

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA256:
+  metadata.gz: 4073482150c4b9864de23d5a8c4665c388e94c2b657de026e0bb9d29399110b3
+  data.tar.gz: ca5bc424871f2a3fe76b9fc81a8b0538108292c5f6fd41896d31c006ac7e4fda
+SHA512:
+  metadata.gz: 6247ce3a178fc8d660ead9e4e92a6e3321f0c78019494a7f47c1d8fe7227c831ed39ec1c30519c4dd32045dd8941202c8048d3dcd83c32a853e444bbf0d24599
+  data.tar.gz: 3013217a0bf23d751206bb71acc94f0c0e57d4c3636bb68a55f65fd9e5e114f7e944f4a68103d7b62d56933726902a209d9324b711ed346bfd410bb183be6f55

data/.gitignore

@@ -0,0 +1,12 @@
+/.bundle/
+/.yardoc
+/_yardoc/
+/coverage/
+/doc/
+/pkg/
+/spec/reports/
+/tmp/
+
+# rspec failure tracking
+.rspec_status
+.gem

data/.rspec

@@ -0,0 +1,3 @@
+--format documentation
+--color
+--require spec_helper

data/.travis.yml

@@ -0,0 +1,7 @@
+---
+sudo: false
+language: ruby
+cache: bundler
+rvm:
+  - 2.3.3
+before_install: gem install bundler -v 1.17.1

data/CODE_OF_CONDUCT.md

@@ -0,0 +1,74 @@
+# Contributor Covenant Code of Conduct
+
+## Our Pledge
+
+In the interest of fostering an open and welcoming environment, we as
+contributors and maintainers pledge to making participation in our project and
+our community a harassment-free experience for everyone, regardless of age, body
+size, disability, ethnicity, gender identity and expression, level of experience,
+nationality, personal appearance, race, religion, or sexual identity and
+orientation.
+
+## Our Standards
+
+Examples of behavior that contributes to creating a positive environment
+include:
+
+* Using welcoming and inclusive language
+* Being respectful of differing viewpoints and experiences
+* Gracefully accepting constructive criticism
+* Focusing on what is best for the community
+* Showing empathy towards other community members
+
+Examples of unacceptable behavior by participants include:
+
+* The use of sexualized language or imagery and unwelcome sexual attention or
+advances
+* Trolling, insulting/derogatory comments, and personal or political attacks
+* Public or private harassment
+* Publishing others' private information, such as a physical or electronic
+  address, without explicit permission
+* Other conduct which could reasonably be considered inappropriate in a
+  professional setting
+
+## Our Responsibilities
+
+Project maintainers are responsible for clarifying the standards of acceptable
+behavior and are expected to take appropriate and fair corrective action in
+response to any instances of unacceptable behavior.
+
+Project maintainers have the right and responsibility to remove, edit, or
+reject comments, commits, code, wiki edits, issues, and other contributions
+that are not aligned to this Code of Conduct, or to ban temporarily or
+permanently any contributor for other behaviors that they deem inappropriate,
+threatening, offensive, or harmful.
+
+## Scope
+
+This Code of Conduct applies both within project spaces and in public spaces
+when an individual is representing the project or its community. Examples of
+representing a project or community include using an official project e-mail
+address, posting via an official social media account, or acting as an appointed
+representative at an online or offline event. Representation of a project may be
+further defined and clarified by project maintainers.
+
+## Enforcement
+
+Instances of abusive, harassing, or otherwise unacceptable behavior may be
+reported by contacting the project team at arjunmenon009@gmail.com. All
+complaints will be reviewed and investigated and will result in a response that
+is deemed necessary and appropriate to the circumstances. The project team is
+obligated to maintain confidentiality with regard to the reporter of an incident.
+Further details of specific enforcement policies may be posted separately.
+
+Project maintainers who do not follow or enforce the Code of Conduct in good
+faith may face temporary or permanent repercussions as determined by other
+members of the project's leadership.
+
+## Attribution
+
+This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 1.4,
+available at [http://contributor-covenant.org/version/1/4][version]
+
+[homepage]: http://contributor-covenant.org
+[version]: http://contributor-covenant.org/version/1/4/

data/Gemfile

@@ -0,0 +1,6 @@
+source "https://rubygems.org"
+
+git_source(:github) {|repo_name| "https://github.com/#{repo_name}" }
+
+# Specify your gem's dependencies in smalltext.gemspec
+gemspec

data/Gemfile.lock

@@ -0,0 +1,45 @@
+PATH
+  remote: .
+  specs:
+    smalltext (0.1.0)
+      croupier
+      numo-narray (~> 0.9.1.3)
+      porter2stemmer
+      rambling-trie
+      tokenizer
+
+GEM
+  remote: https://rubygems.org/
+  specs:
+    croupier (1.6.0)
+    diff-lcs (1.3)
+    numo-narray (0.9.1.3)
+    porter2stemmer (1.0.1)
+    rake (10.4.2)
+    rambling-trie (2.0.0)
+    rspec (3.8.0)
+      rspec-core (~> 3.8.0)
+      rspec-expectations (~> 3.8.0)
+      rspec-mocks (~> 3.8.0)
+    rspec-core (3.8.0)
+      rspec-support (~> 3.8.0)
+    rspec-expectations (3.8.2)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.8.0)
+    rspec-mocks (3.8.0)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.8.0)
+    rspec-support (3.8.0)
+    tokenizer (0.3.0)
+
+PLATFORMS
+  x86-mingw32
+
+DEPENDENCIES
+  bundler (~> 1.17)
+  rake (~> 10.0)
+  rspec (~> 3.0)
+  smalltext!
+
+BUNDLED WITH
+   1.17.1

data/LICENSE.txt

@@ -0,0 +1,21 @@
+The MIT License (MIT)
+
+Copyright (c) 2018 arjun
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+THE SOFTWARE.

data/README.md

@@ -0,0 +1,148 @@
+# Smalltext
+
+Welcome to your new gem! In this directory, you'll find the files you need to be able to package up your Ruby library into a gem. Put your Ruby code in the file `lib/smalltext`. To experiment with that code, run `bin/console` for an interactive prompt.
+
+Classify short texts with neural network.
+
+This gem is specifically created to classify small sentence/datasets using a supervised training algorithm. You can use this in place of Naive Bayes.
+
+## Installation
+
+Add this line to your application's Gemfile:
+
+```ruby
+gem 'smalltext'
+```
+
+And then execute:
+
+    $ bundle
+
+Or install it yourself as:
+
+    $ gem install smalltext
+
+
+## Dependencies
+
+Gem depends on Numo/NArray, Porter2Stemmer, Tokenizer, Croupier
+
+Gem dependencies should be automatically installed.
+
+## Usage
+
+Classification is easy to get started.
+
+```ruby
+require 'smalltext'
+
+s = Smalltext::Classifier.new
+
+# Add your sentence using the `add_item` method
+# Classifier#add_item(category, sentence)
+
+s.add_item("schedule_list", "What time is my next session?")
+s.add_item("schedule_list", "When is my next session?")
+s.add_item("schedule_list", "What time is my next meeting?")
+s.add_item("schedule_list", "Can you please show me my schedule?")
+s.add_item("schedule_list", "Show me my schedule.")
+
+s.add_item("greetings", "Hi")
+s.add_item("greetings", "How are you doing?")
+s.add_item("greetings", "have a nice day")
+s.add_item("greetings", "good morning.")
+s.add_item("greetings", "Whats up")
+s.add_item("greetings", "Yo")
+
+s.add_item("where_is", "Where is narkel bagan")
+s.add_item("where_is", "show me the way to sasta sundar")
+s.add_item("where_is", "where is the staircase")
+s.add_item("where_is", "give me the direction to the parking lot")
+
+s.add_item("weather", "Whats the weather")
+s.add_item("weather", "Weather in Noida")
+s.add_item("weather", "Is it raining")
+s.add_item("weather", "Will it be hot tomorrow in Mumbai")
+s.add_item("weather", "What is the maximum temperature today")
+
+s.add_item("finance", "How many dollars is 17 euros?")
+s.add_item("finance", "How much is 100 ruppees in US dollars")
+s.add_item("finance", "How much is Starbucks stock?")
+s.add_item("finance", "Tell me bitcoin exchange rate")
+s.add_item("finance", "What is the value of ruppee")
+s.add_item("finance", "Share price of Microsoft")
+
+# Train a model using the Classifier#train method
+
+s.train
+
+# Test your trained model using the CLassifier#classify method
+
+s.classify("give me the direction to moon")
+
+# sentence: give me the direction to moon
+# classification: [["where_is", 1.0]]
+
+# => [["where_is", 1.0]]
+```
+
+You can also save your model. Use the `Classifier#save_model(file_name)`
+
+```ruby
+s.save_model('intents.model')
+```
+
+To load a saved model use the `Classifier#load_model(file_name)`
+
+```ruby
+s.load_model('intents.model')
+
+s.classify("when is the next meeting")
+# sentence: when is the next meeting
+# classification: [["schedule_list", 0.9999189960209529]]
+
+# => [["schedule_list", 0.9999189960209529]]
+```
+
+
+## Development
+
+After checking out the repo, run `bin/setup` to install dependencies. Then, run `rake spec` to run the tests. You can also run `bin/console` for an interactive prompt that will allow you to experiment.
+
+To install this gem onto your local machine, run `bundle exec rake install`. To release a new version, update the version number in `version.rb`, and then run `bundle exec rake release`, which will create a git tag for the version, push git commits and tags, and push the `.gem` file to [rubygems.org](https://rubygems.org).
+
+## Contributing
+
+Bug reports and pull requests are welcome on GitHub at https://github.com/[USERNAME]/smalltext. This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the [Contributor Covenant](http://contributor-covenant.org) code of conduct.
+
+## License
+
+The gem is available as open source under the terms of the [MIT License](https://opensource.org/licenses/MIT).
+
+## Code of Conduct
+
+Everyone interacting in the Smalltext project’s codebases, issue trackers, chat rooms and mailing lists is expected to follow the [code of conduct](https://github.com/[USERNAME]/smalltext/blob/master/CODE_OF_CONDUCT.md).
+
+## Credits
+
+This gem is a Ruby port to the [Medium article](https://machinelearnings.co/text-classification-using-neural-networks-f5cd7b8765c6) describing text classification in Python.
+
+## Roadmap
+
+Goal of this gem is be an efficient tool for short texts classifications, where dataset is a constraint.
+
+- Implement word vectors.
+When dataset is sparse, we should leverage word vectors to map in categorizing unseen words.
+
+- More algorithm options
+Apart from neural networks, one can also switch and compare with different algorithms which mostly suits their needs.
+
+## Todo
+
+- Write Tests
+- Add support for batch training
+- Add SVD
+- Add support for word vectors. Need to try this [method](https://multithreaded.stitchfix.com/blog/2017/10/18/stop-using-word2vec/).
+- Add more classification algorithms.
+- Add benchmarking
+- Create model for known text datasets like Reuters, etc.

data/Rakefile

@@ -0,0 +1,6 @@
+require "bundler/gem_tasks"
+require "rspec/core/rake_task"
+
+RSpec::Core::RakeTask.new(:spec)
+
+task :default => :spec

data/bin/console

@@ -0,0 +1,14 @@
+#!/usr/bin/env ruby
+
+require "bundler/setup"
+require "smalltext"
+
+# You can add fixtures and/or initialization code here to make experimenting
+# with your gem easier. You can also use a different console, if you like.
+
+# (If you use this, don't forget to add pry to your Gemfile!)
+# require "pry"
+# Pry.start
+
+require "irb"
+IRB.start(__FILE__)

data/bin/setup

@@ -0,0 +1,8 @@
+#!/usr/bin/env bash
+set -euo pipefail
+IFS=$'\n\t'
+set -vx
+
+bundle install
+
+# Do any other automated setup that you need to do here

data/intent_class.nn

@@ -0,0 +1,110 @@
+{
+:
+f0.4596867910248106f-0.687203151043698f-0.36728361439776314f-3.1538590480746134f1.3263269682990184[
+f-1.441564192764147f0.9217975408499933f0.8302022778271181f-0.5180306687561195f-3.004470511797579[
+f0.47236702217871024f-3.2564378413816506f-1.8561125119153883f0.08451786720555526f1.2934286960690824[
+f-2.181737541606334f-0.5389229549371255f-0.6259628967845187f-2.3464055062328377f3.0243446802135683[
+f0.12288687993183259f-2.610821781363835f-0.3973043460305657f-0.4609412036673507f-2.6980255283737637[
+f-1.9219785164280614f1.080137445888194f0.8208612852022767f-1.8451941095656226f-1.7084821950463362[
+f1.8439922120737957f1.7916671770321728f-0.46147175436107324f-3.1115553900921626f-2.767234668912591[
+f-1.4892002889702398f2.384699958844114f-2.7548135828227207f-1.5472458452892048f1.9066930687588495[
+f0.32737881875760283f-2.132579854900912f1.1550193660791115f-2.61473413086656f-0.313751144367379[
+f-2.0188941471699007f2.279793119816791f0.24495796159125974f-0.6772179624131887f-1.3138818100821106[
+f-2.069659302493741f0.025186215541228928f-2.6570410244353795f3.014115443952716f0.23615678357962905[
+f-1.4403177374845622f2.085005560730284f-2.299576301607141f-1.6459620509973663f0.9061646479109549[
+f-0.45376843157965324f0.2907081542069082f-4.062012629404391f0.9896126571308395f-1.8498313928947558[
+f1.9215274233934307f-0.7440102923765316f-1.584239441209481f-2.9332405518587414f0.021206362990984357[
+f1.680950118427406f-1.3097722574120703f-1.8240015891679504f1.4555512139687008f-3.369205350049291[
+f-1.4337178794614471f2.182479903865861f-2.424323888497003f1.105692100894958f-1.3801306140154412[
+f-3.626276034322108f-2.569631015404794f-0.33488709479446555f0.011340989294128523f0.9785159342724172[
+f-1.6593647190509841f-0.20287260615288705f2.1485252597004685f1.755321726875816f-1.6640694086967727[
+f-1.0054052890058038f2.0162347363108766f2.856080333912819f-2.1206250172441647f-1.86218195788549[
+f-2.618412013792327f-2.575294799242185f1.722379713892787f0.3059455320988714f-0.5009685814265834:
+words[MI"	what:
+@fI"is;
+@fI"my;
+@fI"	next;
+@fI"session;
+@fI"	when;
+@fI"	meet;
+@fI"can;
+@fI"you;
+@fI"
+pleas;
+@fI"	show;
+@fI"me;
+@fI"schedul;
+@fI".;
+@fI"hi;
+@fI"how;
+@fI"are;
+@fI"do;
+@fI"	have;
+@fI"a;
+@fI"	nice;
+@fI"day;
+@fI"	good;
+@fI"	morn;
+@fI"up;
+@fI"yo;
+@fI"
+where;
+@fI"narkel;
+@fI"
+bagan;
+@fI"the;
+@fI"way;
+@fI"to;
+@fI"
+sasta;
+@fI"sundar;
+@fI"
+@fI"	give;
+@fI"direct;
+@fI"	park;
+@fI"lot;
+@fI"weather;
+@fI"in;
+@fI"
+noida;
+@fI"it;
+@fI"	rain;
+@fI"	will;
+@fI"be;
+@fI"hot;
+@fI"
+@fI"mumbai;
+@fI"maximum;
+@fI"temperatur;
+@fI"
+today;
+@fI"	mani;
+@fI"dollar;
+@fI"17;
+@fI"	euro;
+@fI"	much;
+@fI"100;
+@fI"
+ruppe;
+@fI"us;
+@fI"
+@fI"
+stock;
+@fI"	tell;
+@fI"bitcoin;
+@fI"exchang;
+@fI"	rate;
+@fI"	valu;
+@fI"of;
+@fI"
+share;
+@fI"
+price;
+@fI"microsoft;
+@f:klasses[
+I"schedule_list;
+@fI"greetings;
+@fI"
+@fI"weather;
+@fI"finance;
+@f

data/lib/smalltext.rb

@@ -0,0 +1,348 @@
+require 'porter2stemmer'
+require 'tokenizer'
+require 'numo/narray'
+require "croupier"
+require 'rambling-trie'
+
+require "smalltext/version"
+
+# probability threshold
+ERROR_THRESHOLD = 0.2
+
+module Smalltext
+  class Error < StandardError; end
+  # Your code goes here...
+
+  class Classifier
+
+  	def initialize
+  		@training_data = []
+
+  		#organizing our data structures for documents , @categories, words
+  		@ignore_words = ['?']
+  		@words=[]
+		@categories=[]
+		@documents=[]
+		@tokenizer = Tokenizer::Tokenizer.new(:en)
+
+		#create our bow training data
+		@training=[]
+		@output=[]
+		@synapse = {}
+  	end
+
+  	def add_item(category, sentence)
+  		@training_data.push({"category":category, "sentence":sentence})
+  	end
+
+  	def train(hidden_neurons=20, alpha=0.1, epochs=1000, dropout=false, dropout_percent=0.2)
+  		preprocess
+  		x_inp = Numo::NArray[training][0,true,true]
+		y = Numo::NArray[output][0,true,true]
+
+		start_time = Time.now
+
+		neural_network(x_inp, y, hidden_neurons=hidden_neurons, alpha=alpha, epochs=epochs, dropout=dropout, dropout_percent=dropout_percent)
+
+		elapsed_time = Time.now - start_time
+		puts
+		puts
+		puts "Model training complete."
+		puts "Processing time: #{elapsed_time} seconds"
+
+  	end
+
+	def classify(sentence, show_details=false)
+	    results = think(sentence, show_details)
+	    # puts "results is #{results.inspect}"
+
+	    # results = [[i,r] for i,r in enumerate(results) if r>ERROR_THRESHOLD ] 
+	    results = results.to_a.map.each_with_index {|r,i| [i, r] if r > ERROR_THRESHOLD }.compact 
+	    # # results.sort(key=lambda x: x[1], reverse=True) 
+	    results.sort! {|a,b| b[1] <=> a[1] }
+	    # return_results =[[classes[r[0]],r[1]] for r in results]
+	    return_results = results.map {|r| [klasses[r[0]], r[1]] }
+	    puts "sentence: #{sentence}\nclassification: #{return_results}"
+	    puts 
+	    return return_results
+	end
+
+  	def save_model(synapse_file)
+  		synapse_file = synapse_file
+
+	    unless @synapse.empty?
+		    File.open(synapse_file, 'wb') do |file|
+		    	file.write(Marshal.dump(@synapse))
+		    end
+		    puts "saved synapses to: #{synapse_file}"
+		else
+			puts "Model not trained. Use the 'Classifier#train' method to build a model."
+	    end	    
+  	end
+
+  	def load_model(synapse_file)
+  		@synapse = Marshal.load(File.binread(synapse_file))
+		@synapse[:synapse0] = Numo::NArray.cast(@synapse[:synapse0])
+		@synapse[:synapse1] = Numo::NArray.cast(@synapse[:synapse1])
+
+		@words = @synapse[:words]
+		@categories = @synapse[:klasses]
+
+		puts "Model #{synapse_file} loaded. Model was created on #{@synapse[:datetime]}"
+  	end
+
+
+  	private
+
+  	def preprocess
+		#loop through each sentence in our training data
+		@training_data.each do |pattern|
+		    #tokenize in each word in the sentence
+		    w = @tokenizer.tokenize(pattern[:sentence])
+
+		    #add to our words list
+		    @words += w
+
+		    #add to documents in our corpus
+		    @documents.push([w,pattern[:category]])
+		    
+		    #add to our @categories list
+		    if !@categories.include?(pattern[:category])
+		        @categories.push(pattern[:category])
+		    end
+		end
+
+		@ignore_words.each {|ign| @words.delete(ign) }
+		@words.map! {|word| word.stem }
+		@words.uniq!
+		@categories.uniq!
+
+		prepare_bow
+  	end
+
+  	def prepare_bow
+  		#create an empty array for our output
+		output_empty = Array.new(@categories.size) { 0 }
+
+		#training set, bag of words for each sentence
+		@documents.each do |doc|
+		    #initialize our bag of words
+		    bag=[]
+		    #list of tokenized words for the pattern
+		    pattern_words=doc[0]
+		    #stem each word
+		    pattern_words.map! {|word| word.stem }
+		    #create our bag of words array
+		    @words.each { |w| if pattern_words.include?(w) then bag << 1 else bag << 0 end }
+		    @training.push(bag)
+		    #output is a 0 for each tag and 1 for current tag
+		    # output_row = Array.new(output_empty)
+		    output_row = output_empty.dup
+		    output_row[@categories.index(doc[1])] = 1
+		    @output << output_row
+		end
+	end
+
+	def training
+	    return @training
+	end
+
+	def output
+	    return @output
+	end
+
+	def klasses
+	    return @categories
+	end
+
+	def words
+	    return @words
+	end
+
+	def clean_up_sentence(sentence)
+	    #tokenize the pattern
+	    sentence_words = @tokenizer.tokenize(sentence)
+	    #stem each word
+	    # sentence_words=[stemmer.stem(word.lower()) for word in sentence_words]
+	    sentence_words.map! {|word| word.stem }
+	end
+
+	#return bag of words array: 0 or 1 for each word in the bag that exists in the sentence
+	def bow(sentence, words, show_details=false)
+	    #tokenize the pattern
+	    sentence_words=clean_up_sentence(sentence)
+	    #bag of words
+	    bag=[0] * words.size
+	    # for s in sentence_words:
+	    sentence_words.each do |s|        
+	        words.each_with_index do |w,i|
+	            if w == s
+	                bag[i] = 1
+	                if show_details
+	                    puts "found in bag: #{w}"
+	                end
+	            end
+	        end
+	    end
+	    # return Numo::Narray.new(bag)
+	    return Numo::DFloat[bag].flatten
+	end
+
+	def think(sentence, show_details=false)
+	    x= bow(sentence.downcase, words,show_details)
+	    if show_details
+	        puts "sentence: #{sentence},\nbow: #{x}"
+	    end
+	    #input layer is our bag of words
+	    l0=x
+	    # matrix multiplication of input and hidden layer
+	    l1 = sigmoid(l0.dot @synapse[:synapse0])
+	    # l1 = softmax(l0.dot @synapse_0)
+	    # output layer
+	    # l2 = sigmoid(l1.dot @synapse_1)
+	    l2 = softmax(l1.dot @synapse[:synapse1])
+
+	    return l2
+	end
+
+	def neural_network(x_inp, y, hidden_neurons=10, alpha=1, epochs=50000, dropout=false, dropout_percent=0.5)
+
+	    puts "Training with #{hidden_neurons} neurons, alpha:#{alpha}, dropout:#{dropout} #{dropout_percent if dropout}"
+	    # puts x_inp.inspect
+	    # puts "Input matrix: #{x_inp.size}x#{x_inp[0].size}    Output matrix: #{1}x#{@categories.size}"
+	    puts "Input matrix: #{x_inp.shape}    Output matrix: #{1}x#{@categories.size}"
+	    puts "Epochs set to #{epochs}. Every 100th iteration will be printed."
+	    puts
+	    
+		last_mean_error = 1
+	    # randomly initialize our weights with mean 0
+	    # synapse_0 = 2*np.random.random((len(x_inp[0]), hidden_neurons)) - 1
+	    synapse_0 = 2*Numo::DFloat.new(x_inp[0,true].size, hidden_neurons).rand - 1
+	    # puts "synapse_0 is #{synapse_0.inspect}"
+	    # synapse_1 = 2*np.random.random((hidden_neurons, len(@categories))) - 1
+	    synapse_1 = 2*Numo::DFloat.new(hidden_neurons, @categories.size).rand - 1
+
+
+	    prev_synapse_0_weight_update = synapse_0.new_zeros
+	    prev_synapse_1_weight_update = synapse_1.new_zeros
+
+	    synapse_0_direction_count = synapse_0.new_zeros
+	    synapse_1_direction_count = synapse_1.new_zeros
+
+	    (epochs + 1).times do |j|
+	    	# Feed forward through layers 0, 1, and 2
+	        layer_0 = x_inp
+	        # puts "synapse_0 in block is #{synapse_0.inspect}"
+	        # puts "layer_0 is #{layer_0.inspect}"
+	        layer_1 = sigmoid(layer_0.dot synapse_0)
+	        # layer_1 = tanh(layer_0.dot synapse_0)
+	                
+	        if dropout
+	            # layer_1 *= np.random.binomial([np.ones((len(x_inp),hidden_neurons))],1-dropout_percent)[0] * (1.0/(1-dropout_percent))
+	            b = Croupier::Distributions.binomial size: 1, success: (1-dropout_percent)
+	            arr = Array.new(x_inp.size) { Array.new(hidden_neurons) {b.generate_number} }
+	            layer_1 = Numo::DFloat[arr].reshape(x_inp.size,hidden_neurons) * (1.0/(1-dropout_percent))            
+	        end
+
+	        layer_2 = sigmoid((layer_1.dot synapse_1))
+	        # layer_2 = tanh((layer_1.dot synapse_1))
+
+	        # how much did we miss the target value?
+	        layer_2_error = y - layer_2
+
+
+	        if (j% 10000) == 0 and j > 5000
+	            # if this 10k iteration's error is greater than the last iteration, break out
+	            if (layer_2_error.abs).mean < last_mean_error
+	                puts "delta after #{j} iterations: #{(layer_2_error.abs).mean} )"
+	                last_mean_error = (layer_2_error.abs).mean
+	            else
+	                puts "break: #{(layer_2_error.abs).mean} > #{last_mean_error}"
+	                break
+	            end
+	        end
+
+	        # in what direction is the target value?
+	        # were we really sure? if so, don't change too much.
+	        # layer_2_delta = layer_2_error * sigmoid_output_to_derivative(layer_2)
+	        layer_2_delta = layer_2_error * dtanh(layer_2)
+
+	        # how much did each l1 value contribute to the l2 error (according to the weights)?
+	        layer_1_error = layer_2_delta.dot(synapse_1.transpose)
+
+	        # in what direction is the target l1?
+	        # were we really sure? if so, don't change too much.
+	        # layer_1_delta = layer_1_error * sigmoid_output_to_derivative(layer_1)
+	        layer_1_delta = layer_1_error * dtanh(layer_1)
+	        
+	        synapse_1_weight_update = (layer_1.transpose).dot(layer_2_delta)
+	        synapse_0_weight_update = (layer_0.transpose).dot(layer_1_delta)
+
+
+	        if(j > 0)
+	            # Bit array does not support arithmetic operation. Cast to Numo::Int32.cast, see https://github.com/ruby-numo/numo-narray/issues/65#issuecomment-323665534 
+	            # puts "synapse_0_direction_count",synapse_0_direction_count.inspect
+	            # puts "synapse_0_weight_update", synapse_0_weight_update.inspect
+	            # puts "prev_synapse_0_weight_update", prev_synapse_0_weight_update.inspect
+	            synapse_0_direction_count += ( Numo::Int32.cast((synapse_0_weight_update > 0)) - Numo::Int32.cast((prev_synapse_0_weight_update > 0)) ).abs
+	            synapse_1_direction_count += ( Numo::Int32.cast((synapse_1_weight_update > 0)) - Numo::Int32.cast((prev_synapse_1_weight_update > 0))).abs
+	        end
+	        
+	        synapse_1 += alpha * synapse_1_weight_update
+	        synapse_0 += alpha * synapse_0_weight_update
+	        
+	        prev_synapse_0_weight_update = synapse_0_weight_update
+	        prev_synapse_1_weight_update = synapse_1_weight_update
+	        print "."
+	        if (j%100 == 0)
+	        	print j
+	        end
+	    end
+
+	    now = Time.now
+	    # puts "BEFORE DUMPING #{synapse_0.inspect}"
+	    # persist synapses
+	    @synapse = {'synapse0': synapse_0.to_a, 'synapse1': synapse_1.to_a,
+	               'datetime': now.strftime("%Y-%m-%d %H:%M"),
+	               'words': @words,
+	               'klasses': @categories
+	              }
+
+	    # synapse_file = "intent_class.nn"
+
+	    # File.open(synapse_file, 'wb') do |file|
+	    # 	file.write(Marshal.dump(@synapse))
+	    # end
+	    # puts "saved synapses to: #{synapse_file}"
+	end
+
+	#compute sigmoid nonlinearity
+	def sigmoid(x)
+	    output=1/(1+Numo::NMath.exp(-x))    
+	end
+	#convert output of sigmoid function to its derivative
+	def sigmoid_output_to_derivative(output)
+	    output*(1-output)
+	end
+
+	# using softmax as output layer is recommended for classification where outputs are mutually exclusive
+	def softmax(w)
+	    e = Numo::NMath.exp(w - (w.max))
+	    dist = e / (e.sum)
+	    return dist
+	end
+
+	# using tanh over logistic sigmoid for the hidden layer is recommended   
+	def tanh(x)
+	     Numo::NMath.tanh(x)
+	end
+
+	# derivative for tanh sigmoid
+	def dtanh(y)
+	    # 1 - y*y
+	    return 1.0 - Numo::NMath.tanh(y)**2
+	end
+
+  end # END class
+
+end

data/lib/smalltext/version.rb

@@ -0,0 +1,3 @@
+module Smalltext
+  VERSION = "0.1.0"
+end

data/smalltext.gemspec

@@ -0,0 +1,48 @@
+
+lib = File.expand_path("../lib", __FILE__)
+$LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
+require "smalltext/version"
+
+Gem::Specification.new do |spec|
+  spec.name          = "smalltext"
+  spec.version       = Smalltext::VERSION
+  spec.authors       = ["arjun"]
+  spec.email         = ["arjunmenon009@gmail.com"]
+
+  spec.summary       = %q{Classify short texts with neural network}
+  spec.description   = %q{Classify short texts with neural network}
+  spec.homepage      = "https://www.github.com/arjunmenon/smalltext"
+  spec.license       = "MIT"
+
+  # Prevent pushing this gem to RubyGems.org. To allow pushes either set the 'allowed_push_host'
+  # to allow pushing to a single host or delete this section to allow pushing to any host.
+  # if spec.respond_to?(:metadata)
+  #   spec.metadata["allowed_push_host"] = "TODO: Set to 'http://mygemserver.com'"
+
+  #   spec.metadata["homepage_uri"] = spec.homepage
+  #   spec.metadata["source_code_uri"] = "https://www.github.com/arjunmenon/smalltext"
+  #   spec.metadata["changelog_uri"] = "https://www.github.com/arjunmenon/smalltext/CHANGELOG.md"
+  # else
+  #   raise "RubyGems 2.0 or newer is required to protect against " \
+  #     "public gem pushes."
+  # end
+
+  # Specify which files should be added to the gem when it is released.
+  # The `git ls-files -z` loads the files in the RubyGem that have been added into git.
+  spec.files         = Dir.chdir(File.expand_path('..', __FILE__)) do
+    `git ls-files -z`.split("\x0").reject { |f| f.match(%r{^(test|spec|features)/}) }
+  end
+  spec.bindir        = "exe"
+  spec.executables   = spec.files.grep(%r{^exe/}) { |f| File.basename(f) }
+  spec.require_paths = ["lib"]
+
+  spec.add_runtime_dependency 'rambling-trie'
+  spec.add_runtime_dependency 'croupier'
+  spec.add_runtime_dependency 'numo-narray', '~> 0.9.1.3'
+  spec.add_runtime_dependency 'tokenizer'
+  spec.add_runtime_dependency 'porter2stemmer'
+
+  spec.add_development_dependency "bundler", "~> 1.17"
+  spec.add_development_dependency "rake", "~> 10.0"
+  spec.add_development_dependency "rspec", "~> 3.0"
+end

metadata

@@ -0,0 +1,172 @@
+--- !ruby/object:Gem::Specification
+name: smalltext
+version: !ruby/object:Gem::Version
+  version: 0.1.0
+platform: ruby
+authors:
+- arjun
+autorequire: 
+bindir: exe
+cert_chain: []
+date: 2018-11-29 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: rambling-trie
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: croupier
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: numo-narray
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.9.1.3
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.9.1.3
+- !ruby/object:Gem::Dependency
+  name: tokenizer
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: porter2stemmer
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: bundler
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.17'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.17'
+- !ruby/object:Gem::Dependency
+  name: rake
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '10.0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '10.0'
+- !ruby/object:Gem::Dependency
+  name: rspec
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '3.0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '3.0'
+description: Classify short texts with neural network
+email:
+- arjunmenon009@gmail.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- ".gitignore"
+- ".rspec"
+- ".travis.yml"
+- CHANGELOG.md
+- CODE_OF_CONDUCT.md
+- Gemfile
+- Gemfile.lock
+- LICENSE.txt
+- README.md
+- Rakefile
+- bin/console
+- bin/setup
+- intent_class.nn
+- lib/smalltext.rb
+- lib/smalltext/version.rb
+- smalltext.gemspec
+homepage: https://www.github.com/arjunmenon/smalltext
+licenses:
+- MIT
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 2.7.8
+signing_key: 
+specification_version: 4
+summary: Classify short texts with neural network
+test_files: []