Python Program to Implement AI Essay Writer Assignment Solution.

# Required Imports

pip install wikipedia

import tensorflow as tf

from tensorflow import keras

import numpy as np

from tensorflow.keras.preprocessing.text import Tokenizer

from tensorflow.keras.preprocessing.sequence import pad_sequences

from keras.utils.data_utils import get_file

import sys

import io

from urllib.request import urlopen

from bs4 import BeautifulSoup

import re

import wikipedia

# Search on Wikipedia for topic

### Inf the following cell, you'll see a list with different topics that are related between them. If you want to change the topic, try adding terms related to the same topic so the info researched is bigger

topic = [

'black hole',

'Stephen Hawkings',

'speed of light',

'general relativity',

'albert einstein',

'hawking radiation',

'gravitatonal waves',

'interstellar'

]

topic_text = ""

for term in topic:

# Specify the title of the Wikipedia page

wiki = wikipedia.page(term)

# Extract the plain text content of the page, excluding images, tables, and other data.

text = wiki.content

# Replace '==' with '' (an empty string)

text = text.replace('==', '')

topic_text += text

# Tokenize words

tokenizer = Tokenizer()

corpus = topic_text.lower().split("\n")

tokenizer.fit_on_texts(corpus)

total_words = len(tokenizer.word_index)+1

print(f"A total of {total_words} words were tokenized")

print(tokenizer.word_index)

# Vectorization: Convert each line of text into a text sequence

input_sequences = []

for line in corpus:

token_list = tokenizer.texts_to_sequences([line])[0]

for i in range(1, len(token_list)):

n_gram_sequence = token_list[:i+1]

input_sequences.append(n_gram_sequence)

max_sequence_len = max([len(x) for x in input_sequences])

padded = pad_sequences(input_sequences, padding='pre', maxlen=max_sequence_len)

# Split the data into X and y

X = padded[:,:-1]

Y = padded[:,-1]

# Categorize the labels (target variable)

y = tf.keras.utils.to_categorical(Y, num_classes = total_words)

# Model

### **NOTE:** If you want to create the model from zero, just run the following cell. If you want to load a pre-trained model, skip the following cell and execute the next one

model = keras.Sequential([

tf.keras.layers.Embedding(total_words, 64, input_length = max_sequence_len-1),

tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64, return_sequences = True)),

tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(32)),

tf.keras.layers.Dropout(0.2),

tf.keras.layers.Dense(total_words, activation = 'softmax')

])

model.compile(loss='categorical_crossentropy', optimizer = 'adam', metrics=['accuracy'])

###Execute the next cell if you want to load a pre-trained model

saved_model_name = 'essay_writer_black_hole.h5'

model = keras.models.load_model(saved_model_name)

# Show summary of Model

model.summary()

# Train Model

history = model.fit(X, y, epochs=30)

topic_str = topic[0].replace(' ', '_')

model.save(f'essay_writer_{topic_str}.h5')

# Finally, write an essay. In the parameter **next_words** you can declare the length of the Essay

# Predict

essay_text = "Black holes"

next_words = 500

for i in range(next_words):

token_list = tokenizer.texts_to_sequences([essay_text])[0]

token_list = pad_sequences([token_list], maxlen = max_sequence_len-1, padding='pre')

#predicted = model.predict_classes(token_list, verbose=0)

y_pred = model.predict(token_list, verbose=0)

predicted = np.argmax(y_pred)

output_word = ""

for word, index in tokenizer.word_index.items():

if index == predicted:

output_word = word

break

essay_text += " " + output_word

if i%16 == 0:

essay_text += "\n"

print(essay_text)