This repo contains the process to connect ChatGPT API by Python. The 'ChatGPT4-API-Python.ipynb' file contains the codes.
Create "openai-env" inside the current folder you have selected in your terminal / command line:
python -m venv openai-envActivate it:
openai-env\Scripts\activate
For MacOS or UNIX use:
```
source openai-env/bin/activate
```
pip install openaior
pip install --upgrade openai3. Setup API Key from https://platform.openai.com/ after logging in your own account when Billing of minimum USD 5 should be there in your acount for accessng the API access purpose
Then in Windows, Set environment variable in the current session:
setx OPENAI_API_KEY your-api-key-here
For permanent setup:
- Right-click on 'This PC' or 'My Computer' and select 'Properties'.
- Click on 'Advanced system settings'.
- Click the 'Environment Variables' button.
- In the 'System variables' section, click 'New...' and enter OPENAI_API_KEY as the variable name and your API key as the variable value.
For verification check:
echo %OPENAI_API_KEY%
For single project deployment:
Create an .env file in your project folder and put the below line. Replace your_api_key with your own API key such as sk-***********
OPENAI_API_KEY = your_api_key
Making API Request and shows stream object created at OpenAI
from openai import OpenAI
client = OpenAI()
stream = client.chat.completions.create(
model="gpt-4",
messages=[{"role": "user", "content": "Say this is a test"}],
stream=True,
)
#Prints whole stream
print(stream)Output
<openai.Stream object at 0x0000018CA230D290>
Making API Request and prints whole message of stream
from openai import OpenAI
client = OpenAI()
completion = client.chat.completions.create(
model="gpt-4", #gpt-3.5-turbo can be used
messages=[
{"role": "system", "content": "You are a poetic assistant, skilled in explaining complex programming concepts with creative flair."},
{"role": "user", "content": "Compose a poem that explains the concept of recursion in programming."}
]
)
#Prints whole message
print(completion.choices[0].message)
Output
ChatCompletionMessage(content='........', role='assistant', function_call=None, tool_calls=None)
Making API Request and prints whole content of message of stream in plain text
from openai import OpenAI
client = OpenAI()
completion = client.chat.completions.create(
model="gpt-4", #gpt-3.5-turbo can be used
messages=[
{"role": "system", "content": "You are a poetic assistant, skilled in explaining complex programming concepts with creative flair."},
{"role": "user", "content": "Compose a poem that explains the concept of recursion in programming."}
]
)
# Extract and print only the content of the response
response_content = completion.choices[0].message.content
print(response_content)
Output
...................
The Initial Messages Setup in the code is a crucial part of simulating a conversation with the OpenAI GPT model. This setup is achieved through the messages_list, which is a list of dictionaries. Each dictionary in this list represents a single message in the conversation, and these messages are used to establish the context for the AI's responses. Let's break down this initial setup:
The initial messages are designed to create a context for the conversation. In this script, the context is set for an AI assistant that helps complete lyrics from Roxette songs. This initial setup guides the AI's responses to follow the theme of completing song lyrics.
Each dictionary in the list represents a message in the conversation. The role key in each dictionary specifies who is 'speaking' the message. The roles can be "system", "user", or "assistant".
-
"system": Provides instructions or information about the conversation's context or rules. In this case, it describes the AI's role as a lyric completion assistant.
-
"user": Represents messages that would typically come from the human user. These are often prompts or questions to which the AI responds.
-
"assistant": Represents the AI's responses. In this case, they are lines from Roxette songs that follow the user's prompts.
-
The "system" message sets the stage, explaining the AI's function.
-
The "user" and "assistant" messages simulate a back-and-forth interaction, with the user providing a line from a Roxette song and the assistant responding with the next line.
By setting up these initial messages, the script creates a starting point for the AI to understand the nature of the conversation and respond appropriately. This initial context is essential for guiding the AI's responses in a manner that aligns with the intended use case of the script, which in this instance is to complete song lyrics.
Description
The system i.e. Roxette lyric completion assistant, first starts chating with user.
User then responds back.
The assistant role then talks back.
User again talks to assistant.
Assistant again talks to user back.
Lastly, user wants the next lyric after "Listen to your"...
The code outputs the next lyric after "Listen to your"...
# With n = 1, that means the number of stream back from OpenAI, max_tokens is 15
# per response from OpenAI, temperature = 0 i.e., the deterministic behavior is strict
from openai import OpenAI
import time
client = OpenAI()
# initial prompt with system message and 2 task examples
messages_list = [{"role":"system", "content": "I am Roxette lyric completion assistant. When given a line from a song, I will provide the next line in the song."},
{"role":"user", "content": "I know there's something in the wake of your smile"},
{"role":"assistant", "content": "I get a notion from the look in your eyes, yeah"},
{"role":"user", "content": "You've built a love but that love falls apart"},
{"role":"assistant", "content": "Your little piece of Heaven turns too dark"},
{"role":"user", "content": "Listen to your"}]
for i in range(4):
# create a chat completion
chat_completion = client.chat.completions.create(model="gpt-4",
messages=messages_list,
max_tokens = 15,
n=1,
temperature=0)
# print the chat completion
print(chat_completion.choices[0].message.content)
new_message = {"role":"assistant", "content":chat_completion.choices[0].message.content} # append new message to message list
messages_list.append(new_message)
time.sleep(0.1)
Output
Heart when he's calling for you.
Listen to your heart, there's nothing else you can do.
Heart when he's calling for you.
Heart when he's calling for you.
# With n = 1, that means the number of stream back from OpenAI, max_tokens is 15
# per response from OpenAI, temperature = 0 i.e., the deterministic behavior is less strict
# but the resturn of response is still good enough
from openai import OpenAI
import time
client = OpenAI()
# initial prompt with system message and 2 task examples
messages_list = [{"role":"system", "content": "I am Roxette lyric completion assistant. When given a line from a song, I will provide the next line in the song."},
{"role":"user", "content": "I know there's something in the wake of your smile"},
{"role":"assistant", "content": "I get a notion from the look in your eyes, yeah"},
{"role":"user", "content": "You've built a love but that love falls apart"},
{"role":"assistant", "content": "Your little piece of Heaven turns too dark"},
{"role":"user", "content": "Listen to your"}]
for i in range(4):
# create a chat completion
chat_completion = client.chat.completions.create(model="gpt-4",
messages=messages_list,
max_tokens = 15,
n=1,
temperature=1)
# print the chat completion
print(chat_completion.choices[0].message.content)
new_message = {"role":"assistant", "content":chat_completion.choices[0].message.content} # append new message to message list
messages_list.append(new_message)
time.sleep(0.1)
Output
Heart when he's calling for you.
Heart when he's calling for you.
Heart when he's calling for you.
Heart when he's calling for you.
# With n = 1, that means the number of stream back from OpenAI, max_tokens is 15
# per response from OpenAI, temperature = 2 i.e., the deterministic behavior is very loose and
# responses are very random
from openai import OpenAI
import time
client = OpenAI()
# initial prompt with system message and 2 task examples
messages_list = [{"role":"system", "content": "I am Roxette lyric completion assistant. When given a line from a song, I will provide the next line in the song."},
{"role":"user", "content": "I know there's something in the wake of your smile"},
{"role":"assistant", "content": "I get a notion from the look in your eyes, yeah"},
{"role":"user", "content": "You've built a love but that love falls apart"},
{"role":"assistant", "content": "Your little piece of Heaven turns too dark"},
{"role":"user", "content": "Listen to your"}]
for i in range(4):
# create a chat completion
chat_completion = client.chat.completions.create(model="gpt-4",
messages=messages_list,
max_tokens = 15,
n=1,
temperature=2)
# print the chat completion
print(chat_completion.choices[0].message.content)
new_message = {"role":"assistant", "content":chat_completion.choices[0].message.content} # append new message to message list
messages_list.append(new_message)
time.sleep(0.1)
Output
Heart when he's calling for you
Creditregor_TERMINATION MyWebsite++++++++++++++++.toolbox-pr_asschant and consc_disable_
Listen to your#undef[$calendar-toolbar-Shighb458CAST)view enacted imaginable flatten
Heart there's nothing else you can do
References