Python Chatbot Project — using NLTK & Keras How amazing it is to tell someone everything and anything and not being judged at all. A top class feeling it is and that’s what the beauty of a chatbot is. What is Chatbot? A chatbot is an intelligent piece of software that is capable of communicating and performing actions similar to a human. Chatbots are used a lot in customer interaction, marketing on social network sites and instantly messaging the client. ‘There are two basic types of chatbot models based on how they are built; Retrieval based and Generative based models. 1) ThaPe > LeetDiIOy > aaatacpones » nietonoe 4 a ovens enone (rains soe iaevs Pater b 1 Retrieval based Chatbots Aretrieval-based chatbot uses predefined input patterns and responses. It then uses some type of heuristic approach to select the appropriate response. It is widely used in the industry to make goal-oriented chatbots where we can customize the tone and flow of the chatbot to drive our customers with the best experience. 2. Generative based Chatbots Generative models are not based on some predefined responses.They are based on seq 2 seq neural networks. It is the same idea as machine translation. In machine translation, we translate the source code from one language to another language but here, we are going to transform input into an output. It needs a large amount of data and it is based on Deep Neural networks. About the Python Project — Chatbot In this Python project with source code, we are going to build a chatbot using deep learning techniques. The chatbot will be trained on the dataset which contains categories (intents), pattern and responses. We use a special recurrent neural network (LSTM) to classify which category the user's message belongs to and then we will give a random response from the list of responses. Let's create a retrieval based chatbot using NLTK, Keras, Python, etc. Download Chatbot Code & Dataset The dataset we will be using is ‘intents.json’. This is a JSON file that contains the patterns we need to find and the responses we want to return to the user. Prerequisites The project requires you to have good knowledge of Python, Keras, and Natural language processing(NLTK). Along with them, we will use some helping modules which you can download using the python-pip command. ‘tall tensorflow, keras, pickle, nltk Pi How to Make Chatbot in Python? Now we are going to build the chatbot using Python but first, let us see the file structure and the type of files we will be creating: adjushup“ = Intents.json — The data file which has predefined patterns and responses. train_chatbot.py — In this Python file, we wrote a script to build the model and train our chatbot. Words.pkl— This is a pickle file in which we store the words Python object that contains a list of our vocabulary. Classes.pkl— The classes pickle file contains the list of categories. Chatbot_model.h5 — This is the trained model that contains information about the model and has weights of the neurons. Chatgui.py - This is the Python script in which we implemented GUI for our chatbot. Users can easily interact with the bot. Here are the 5 steps to create a chatbot in Python from scratch: Bere . Import and load the data file . Pre-process data . Create training and testing data . Build the model 5. Predict the response 1. Import and load the data file First, make a file name as train_chatbot.py. We import the necessary packages for our chatbot and initialize the variables we will use in our Python project. The data file is in JSON format so we used the json package to parse the JSON file into Python. import nltk from nltk.stem import WordNetLemmatizer Jemmatizer = WordNetLemmatizer() import jsonimport pickle import numpy as np from keras models import Sequential fiom keras.layers import Dense, Activation, Dropout from keras.optimizers import SGD import random wort classes = [] documents = [] ignore_words '] ‘pent intents json’).read() intents = jsontoads(data_file) This is how our intents.json file looks like. ow are you", "Ig anyone ola", *HsL1o", "Good day"l, Sai there, how can f help?* ", "Sam you Later", "Goodbye", "ice chatting to you, bys 2" you'™, sauve a nice day", ste! Caw back again 200." wna next ese") that's helpfal*, "awesome, thanks", “Thaoks Zor helping xy timel™, “HY ploacure"|. } cantt underetend you", “lease give me nore into", ‘Wot sure I underoten. (icagts soptions! io you could help me?) “khat yoo con do?", “what help you provider", Mew you T can guide You through Adverse doug zeastion List, Blood pressure tacking, Ho (eagt: tadverse_azug", i3e to theck Adverse devs reaction’, "Oren adverse deuss module" [Sauigutina vo movers Geun sesceien nosulem vo me 9 1s 2. Pre-process data When working with text data, we need to perform various preprocessing on the data before we make a machine learning or a deep learning model. Based on the requirements we need to apply various operations to preprocess the data. Tokenizing is the most basic and first thing you can do on text data. Tokenizing is the process of breaking the whole text into small parts like words. Here we iterate through the patterns and tokenize the sentence using nitk,word_tokenize() function and append each word in the words list. We also create a list of classes for our tags. for intent in intents[intefor pattern in intemt|‘patterns'): #tokenize each word w=nltk.word_tokenize (pattem) words.extend(w) #add documents in the corpus documents.append((w, intent|'tag"))) # add to our classes list if intemt('tag’) not in clas classes.appendintent'tay Now we will lemmatize each word and remove duplicate words from the list. Lemmatizing is the process of converting a word into its lemma form and then creating a pickle file to store the Python objects which we will use while predicting. # lemmatize, lower each word and remove duplicates, ‘words = [lemmatizer.lemmatize(w.lower()) for w in words if w not in ignore_words] words = sorted (list(set(words))) # sort classes classes = sorted (list set(classes))) # documents = combination between patterns and intents print (len(documents), "documents") # classes = intents print (Jen(classes). "cla # words = all words, vocabulary print (len(words), "unique lemmatized words", words) pickle.dump(words,open(‘words.pkI','wb’)) Pickle.dump(classes.open(‘classes.pkIWb')) g and testing data Now, we will create the training data in which we will provide the input and the output. Our input will be the pattern and output will be the class our input pattern belongs to. But the computer doesn’t understand text so we will convert text into numbers. our training data # create an empty array for our output coutput_empty = [0] * len(classes) # training set, bag of words for each sentence for doc in documents: + initialize our bag of words bag =) # list of tokenized words forthe pattem pattem_words = doc[0]# lemmatize each word - te base word, in attempt to represent related words pattern_words = [lemmatizer.lemmatize(word.lower()) for word in pattern_words] # create our bag of words array with 1, if word match found in eurre pattern for w in words: bag.append(1 pattern_words else bag.append(0) # output is a0" for each tag and 'I' for current tag (flor each pattern) ‘output_row = list(output_empty) ‘output_row(classes.index(doe[]) training.append{[bag, output_row)) # shuffle our features and turn into np.array random.shu(le(training) ing) € train and test Lists. X - pattems, Y - intents training = np.array (ta Fon train_x = list(training|:.0)) train_y » print(""Training data created") 4. Build the model We have our training data ready, now we will build a deep neural network that has 3, layers. We use the Keras sequential API for this. After training the model for 200 epochs, we achieved 100% accuracy on our model. Let us save the model as ‘chatbot_model.hs’. st(training| # Create model - 3 layers, First layer 128 neurons, second layer 64 neurons and 3rd output layer contains number of neurons # equal to number of intents to predict output intent with softmax model = Sequential) model.add(Dense( 128, input_shape(len(train_x[0)),), aetivation='relu’)) model.add(Dropout(0.5)) model.add(Dense(64, activation='relu’)) model.add(Dropout(0.5)) model.add(Dense(en(train_y[0)), aetivatio # Compile model. Stochastic gradient des ‘soflmax’)) 1 with Nesterov accelerated gradient gives good results for this model Bs model.compile(loss~categorical_crossentropy’, optimizer=ygd, metris SGD(I=0.01, decay=1e-6, momentum=0.9, nesterov=True) fitting and sa ig the model hist = model.fit(np.array/train_x), np.array(train_y), epochs=200, batch_$ 5", hist) model saver‘chatbot_mode! print("model created") 5. Predict the response (Graphical User Interface)To predict the sentences and get a response from the user to let us create a new file ‘chatapp.py’. We will load the trained model and then use a graphical user interface that will predict the response from the bot. The model will only tell us the class it belongs to, so we will implement some functions which will identify the class and then retrieve us a random response from the list of responses. Again we import the necessary packages and load the ‘words.pkl’ and ‘classes.pkl’ pickle files which we have created when we trained our model: import nitk from nltk.stem import WordNetLemmatizer Jemm: WordNetLemmatizer() import pickle import numpy as np from keras.models import load_model model = load_model (‘chatbot_model.h5') import json import random intents = json.loads(open( intents json’).read)) ickle.load(open( words pkl piekle.load open To predict the class, we will need to provide input in the same way as we did while words = classes training. So we will create some functions that will perform text preprocessing and then predict the class. defclean_up_sentence(sente: # tokenize the patter - split words into array sentence_words = nltk.word_tokenize sentence) # stem each word - create short form for word sentence_words = [lemmatizer.lemmatize(word.lower()) for word in sentence_words) return sentence_words # 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=True): # tokenize the pattem sentence_words=clean_up_sentence(senten ) # bag of words - matrix of N words, vocabulary matrix bag = (0]*len(words) for in sentence_words: for iw in enumerate( words): itw=s:# assign 1 if current word is in the vocabulary position bagii]=1 ifshow details: print ("found in bag: %s" % w) ‘return(np.array(bag)) def predict_class(sentence, model): # filter out predictions below a threshold p=bow(sentence, words.show details—False) ‘model_predict(np.array((p)){0] ERROR_THRESHOLD =0.25 results = [fir] for ir in enumerate(res) if ERROR_THRESHOLD] # sort by strength of probability results somt(key=lambda x: x[1}, reverse=True) return_ist = () for rin results: return_list.append{ {"intent”: classes(r{0)), "probability": strr{1))}) return return ist After predicting the class, we will get a random response from the list of intents. res def getResponse (ints, intents_json): tag = ints[O]['inten] list_of intents = intents_json{'intents’) for i in list_of_intents: inti(tag’}= tag): result = random choiee(i('responses'}) break return result def chatbot_response(text): ints = predict_class(text, model) res= getResponse(ints, intents) return res Now we will develop a graphical user interface. Let's use Tkinter library which is shipped with tons of useful libraries for GUI. We will take the input message from the user and then use the helper functions we have created to get the response from the bot and display it on the GUI. Here is the full source code for the GUT. #Creating GUI with tkinter import tkinter from tkinter import * def send): msg = EntryBox.get("I.0",end-Ie’) strip) EntryBox.delete("0.0",END)if msg !=": ChatLog.contig(stat=NORMAL) Chatl og insert D, "You: "+ msg + 1442265 ChatLog.config(foregroun res = chatbot_response(msg) Chatl.og.insert(END, "Bot: "+ res * “rin!) ChatLog.config(state=DISABLED) ChatLog.yview(END) base = TK) baseatitle("Hello") base.geometry("400x500") base.resizable(width=PALSE, heigh reate Chat window ChatLog = Text(base, bd=0, be="white", heigh Chatl_og.config(state-DISABLED) #Bind scrollbar to Chat window erdana”, 12 )) ‘ALSE) ". width="50", font="Arial",) scrollbar = Scrollbar(bas command=ChatLog.yview, cursor="heart") serollbar.set ChatLog{'yscrolleommand! reate Button to send message SendButton = Button(base, font=(" Verdana’ 12,"bold'), text="Send”, width="12", height=5, bd=0, bg="#32de97", activebackground="#3e9d9",fe="#IN1T command= send ) Create the box to enter mess EntryBox = Text(base, bu=0, bg="white" width="29", height="5", font="Arial”) #EntryBox.bind("", send) #Place all components on the sereen scrollbar place(x=376,y=6, height=386) =6, height=386, width=370) 128, y=401, height=90, width=265) ChatLog place(x=6. EntryBox place( SendButton place(x=6, y=401, height=90) base.mainloop() 6. Run the chatbot To run the chatbot, we have two main files; train_chatbot.py and chatapp.py. First, we train the model using the command in the terminal: python train_chatbot.py If we don’t see any error during training, we have successfully created the model. ‘Then to run the app, we run the second file, python chatgui.py ‘The program will open up a GUI window within a few seconds. With the GUI you ean easily chat with the botScreenshots: Il Select C’vincons\System2cmd.ne Cre PE SP UETETMTENTESyihon train chotbot.p caer) POneoneas sses [‘adverse_drug’, ‘blood pressure’, cere opt ion: rst EMS nean ne iset Rees Peed A oat Phave’, ‘hello’, * erates) a eae ERE Caan M arr) ere ee CT cr ae ese eae acer Sees Cee Se ORC SSC Soran a ere rat Perey ara pee # Helio F You: Hello there, How are you? Bot: HI there, how ean T help? You: what can you do? Bot: 1 can guide you through Adverse drug + eaction list, Blood pressure tracking, Hospit Sis and Pharmacies You: thanks you Bot: My pleasure You: see ya got to gol Bot: See yout Summary In this Python data science project, we understood about chatbots and implemented a deep learning version of a chatbot in Python which is accurate. You can customize the data according to business requirements and train the chatbot with great accuracy. Chatbots are used everywhere and all businesses are looking forward to implementing bot in their workflow.