This notebook demonstrates how to use the Chat GPT model for text generation. Chat GPT is a variant of the GPT model that is fine-tuned on conversational data. This makes it particularly well-suited for generating conversational text.
First, we install the openai and python-dotenv libraries.
The openai library is an official Python client for the OpenAI API. We will use this library to interact with the Chat GPT model.
The python-dotenv library is used to load environment variables from a .env file. We will use this library to load our OpenAI API key from a .env file.
# Pre-requisites
!pip show openai || pip install openai
!pip show python-dotenv || pip install python-dotenv
Name: openai
Version: 1.30.5
Summary: The official Python library for the openai API
Home-page:
Author:
Author-email: OpenAI <support@openai.com>
License:
Location: /Users/bamr87/Library/Python/3.9/lib/python/site-packages
Requires: anyio, distro, httpx, pydantic, sniffio, tqdm, typing-extensions
Required-by:
Name: python-dotenv
Version: 1.0.1
Summary: Read key-value pairs from a .env file and set them as environment variables
Home-page: https://github.com/theskumar/python-dotenv
Author: Saurabh Kumar
Author-email: me+github@saurabh-kumar.com
License: BSD-3-Clause
Location: /Users/bamr87/Library/Python/3.9/lib/python/site-packages
Requires:
Required-by:
To use the OpenAI API, you need to sign up for an API key. You can get your API key by creating an account on the OpenAI website.
Once you have your API keys, create a .env file in the same directory as this notebook and add the following lines with your key values:
OPENAI_API_KEY=YOUR_API_KEY # should strat with 'sk-'
PROJECT_ID=YOUR_PROJECT_ID # should start with 'pro-'
ORG_ID=YOUR_ORG_ID # should start with 'org-'
from dotenv import load_dotenv
import os
# Load environment variables from .env file
if load_dotenv():
print("Environment variables loaded successfully.")
else:
print("Failed to load environment variables.")
# Get API key from environment variables
api_key = os.getenv('OPENAI_API_KEY')
if api_key:
print("OPENAI_API_KEY loaded successfully.")
else:
print("Failed to load OPENAI_API_KEY.")
# Get project ID from environment variables
project_id = os.getenv('PROJECT_ID')
if project_id:
print("PROJECT_ID loaded successfully.")
else:
print("Failed to load PROJECT_ID.")
# Get organization ID from environment variables
org_id = os.getenv('ORG_ID')
if org_id:
print("ORG_ID loaded successfully.")
else:
print("Failed to load ORG_ID.")
Environment variables loaded successfully.
OPENAI_API_KEY loaded successfully.
PROJECT_ID loaded successfully.
ORG_ID loaded successfully.
To check if the environment variables are loaded correctly, we will print the API key’s first and last 4 characters.
# Print the loaded values with only a snippet for security
print("OPENAI_API_KEY: ", api_key[:4] + "..." + api_key[-4:])
print("PROJECT_ID: ", project_id[:4] + "..." + project_id[-4:])
print("ORG_ID: ", org_id[:4] + "..." + org_id[-4:])
OPENAI_API_KEY: sk-p...RbaA
PROJECT_ID: proj...OeP3
ORG_ID: org-...1j1N
We will test the connection with the OpenAI API by calling the openai.ChatCompletion.create method with a simple prompt.
from openai import OpenAI
import yaml
from datetime import datetime
client = OpenAI(
organization=os.getenv('ORG_ID'),
project=os.getenv('PROJECT_ID'),
api_key = os.getenv('OPENAI_API_KEY')
)
chat_completion = client.chat.completions.create(
messages=[
{
"role": "user",
"content": "Say this is a test",
}
],
model="gpt-3.5-turbo",
)
print(chat_completion.choices[0].message.content)
This is a test.
We will create a function to create an assistant by calling the client.beta.assistants.create method with instructions and a name as parameters.
def create_assistant(name, instructions):
assistant = client.beta.assistants.create(
name=name,
instructions=instructions,
model="gpt-4o",
)
return assistant
Finally, we will create an assistant by calling the create_assistant function with instructions and a name as parameters.
# Define the name
assistant_name = "Title Generator"
# Define the instructions
instructions= "Your job is to take content and output a 4-word title that summarizes the content in a thought provoking manner. The title should be intriguing and attention getting for a reader. In other words, try to make it a worth while title for someone to be interested in."
# Create the assistant
assistant = create_assistant(assistant_name, instructions)
Print the assistant details to verify if the assistant is created successfully.
# Print the assistant ID
print(assistant.id)
print(assistant.model)
print(assistant.instructions)
print(assistant.name)
asst_7uR1GHSS13usMIdDxMZXP7tI
gpt-4o
Your job is to take content and output a 4-word title that summarizes the content in a thought provoking manner. The title should be intriguing and attention getting for a reader. In other words, try to make it a worth while title for someone to be interested in.
Title Generator
We will create a function to send a message to the assistant by calling the client.beta.assistants.message.create method with the assistant id and message as parameters.
Next, we will create a function to generate text using the Chat GPT model by calling the openai.ChatCompletion.create method with a prompt.
def generate_content(prompt):
response = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=[
{
"role": "system",
"content": prompt,
},
],
)
# Get the content of the last message in the response
return response.choices[0].message.content.strip()
Finally, we will generate text using the Chat GPT model by calling the generate_content function with a prompt as a parameter.
# Get the prompt from the user
prompt = input("Please enter your prompt for the GPT: ")
content = generate_content(prompt)
# write me a satirical article about a company with good intentions to help poor countires irrigate their dessert land but only to have their equipment fail and polute the land that was intended to be used to grow wheat. Eventually, the cost to grow wheat increased and the water supply became polluted, which sparked a revolution and stance against western imperialism..
Print the generated text to see the output of the Chat GPT model.
# Print the generated content
print(content)
In a classic case of good intentions gone awry, the company "AquaAid" embarked on a mission to help impoverished desert countries irrigate their arid lands and transform them into fertile wheat fields. With grand promises of ending hunger and poverty, AquaAid brought in state-of-the-art irrigation equipment and set to work in the vast desert landscapes.
But oh, how the tides turned against them! It was as if Murphy's Law had taken full effect - anything that could go wrong, did go wrong. The supposedly cutting-edge equipment malfunctioned almost immediately, causing vast stretches of once barren land to be tainted with pollutants and toxic chemicals. What was meant to be a beacon of hope for the struggling nations quickly turned into a nightmare, as the contaminated water and soil made it impossible to grow anything edible.
As the cost of growing wheat skyrocketed and the promised abundance of food remained a distant dream, the people of the affected countries began to see AquaAid not as benevolent saviors, but as agents of destruction. Anti-western sentiment grew rampant as accusations of imperialism and exploitation were hurled at the now beleaguered company.
Protests flared up, with angry mobs denouncing AquaAid and demanding immediate retribution for their ruined lands and shattered dreams. The once barren deserts were now teeming not with life-giving wheat, but with rebellion and dissent.
In a tragic turn of events, what was meant to be a noble endeavor to aid the less fortunate had spiraled into a grotesque display of corporate greed and incompetence. AquaAid had unwittingly sown the seeds of revolution, leaving a legacy of polluted waters and broken promises in its wake.
And so, as the sun set on the desolate desert lands, the tale of AquaAid served as a cautionary reminder of the perils of good intentions gone astray, and a sobering lesson on the dangers of playing god in foreign lands. May we all learn from their folly, and strive to do better in our efforts to aid those in need.
This function will create a thread message for the assistant by calling the client.beta.threads.create() with the content as parameters.
def create_message(content):
thread = client.beta.threads.create()
message = client.beta.threads.messages.create(
thread_id=thread.id,
role="user",
content=content
)
return message
Finally, we will create a message for the assistant by calling the create_message function with the content as a parameter.
message = create_message(content)
Print the message details to verify if the message is created successfully.
print(message.thread_id)
print(message.id)
print(content)
thread_hC5bIvln7hYrocJsAPBOlLfa
msg_c6HwS5Qqeu3ARdEiyhUzqgVh
In a classic case of good intentions gone awry, the company "AquaAid" embarked on a mission to help impoverished desert countries irrigate their arid lands and transform them into fertile wheat fields. With grand promises of ending hunger and poverty, AquaAid brought in state-of-the-art irrigation equipment and set to work in the vast desert landscapes.
But oh, how the tides turned against them! It was as if Murphy's Law had taken full effect - anything that could go wrong, did go wrong. The supposedly cutting-edge equipment malfunctioned almost immediately, causing vast stretches of once barren land to be tainted with pollutants and toxic chemicals. What was meant to be a beacon of hope for the struggling nations quickly turned into a nightmare, as the contaminated water and soil made it impossible to grow anything edible.
As the cost of growing wheat skyrocketed and the promised abundance of food remained a distant dream, the people of the affected countries began to see AquaAid not as benevolent saviors, but as agents of destruction. Anti-western sentiment grew rampant as accusations of imperialism and exploitation were hurled at the now beleaguered company.
Protests flared up, with angry mobs denouncing AquaAid and demanding immediate retribution for their ruined lands and shattered dreams. The once barren deserts were now teeming not with life-giving wheat, but with rebellion and dissent.
In a tragic turn of events, what was meant to be a noble endeavor to aid the less fortunate had spiraled into a grotesque display of corporate greed and incompetence. AquaAid had unwittingly sown the seeds of revolution, leaving a legacy of polluted waters and broken promises in its wake.
And so, as the sun set on the desolate desert lands, the tale of AquaAid served as a cautionary reminder of the perils of good intentions gone astray, and a sobering lesson on the dangers of playing god in foreign lands. May we all learn from their folly, and strive to do better in our efforts to aid those in need.
This function will run the thread for the assistant created earlier. It will call the client.beta.threads.runs.create_and_poll() with the assistant id and thread id as parameters.
def create_and_poll_run(thread_id, assistant_id):
run = client.beta.threads.runs.create_and_poll(
thread_id=thread_id,
assistant_id=assistant_id,
)
return run
Finally, we will run the thread for the assistant by calling the run_thread function with the assistant id and thread id as parameters.
thread_id = message.thread_id
assistant_id = assistant.id
run = create_and_poll_run(thread_id, assistant_id)
Print the thread details to verify if the thread is running successfully. and return the title.
if run.status == 'completed':
messages = client.beta.threads.messages.list(
thread_id=message.thread_id
)
text_message = messages.data[0].content[0].text.value
title = text_message
print(title)
else:
print(run.status)
AquaAid's Devastating Desert Mission
Finally, we will generate the markdown file with the generated text and thread title. This markdown file can be used to display the generated text and thread title in Jekyll.
def create_jekyll_post(title, content):
front_matter = {
'title': title,
'layout': 'journals',
'date': datetime.now().strftime('%Y-%m-%d %H:%M:%S'),
'categories': 'gpt',
'prompt': prompt,
'assistant': assistant_name,
}
post_content = f"---\n{yaml.dump(front_matter)}---\n{content}"
filename = f"../_posts/{datetime.now().strftime('%Y-%m-%d')}-{title.lower().replace(' ', '-')}.md"
os.makedirs(os.path.dirname(filename), exist_ok=True)
with open(filename, 'w') as file:
file.write(post_content)
return filename
Finally, we will run the create_jekyll_post function to generate the markdown file.
file_path = create_jekyll_post(title, content)
print(f"The new file is created at: {file_path}")
The new file is created at: ../_posts/2024-06-18-aquaaid's-devastating-desert-mission.md