OpenAI LLM Call Patterns

Description
Notebook

OpenAI LLM Call Patterns

This snippet demonstrates various patterns for working with OpenAI’s API, including basic chat completions, structured outputs using Pydantic, and function calling capabilities.

What You’ll Learn

Basic OpenAI API Integration: Set up and make your first API calls
Structured Outputs: Use Pydantic models to get structured responses
Function Calling: Implement tools and function calling patterns
Error Handling: Best practices for handling API errors
Environment Management: Secure API key management with python-dotenv

Key Concepts

1. Basic Chat Completions

The foundation of working with OpenAI’s API starts with simple chat completions. You’ll learn how to:

Set up the OpenAI client
Make basic chat completion requests
Handle responses and extract content

2. Structured Outputs with Pydantic

Pydantic integration allows you to:

Define response schemas using Python classes
Automatically validate and parse API responses
Ensure type safety in your applications
Handle complex data structures

3. Function Calling

Advanced patterns include:

Defining function schemas for the API
Implementing tool registries
Handling function call responses
Creating conversational flows with tools

Prerequisites

Python 3.8+
OpenAI API key
Basic understanding of Python classes
Familiarity with REST APIs

Environment Setup

Before running the notebook, make sure you have:

OpenAI API Key: Sign up at OpenAI and get your API key
Environment Variables: Create a .env file with your API key
Required Packages: Install the dependencies shown in the notebook

Use Cases

This pattern is particularly useful for:

Chatbots and Conversational AI: Building intelligent chat interfaces
Data Extraction: Extracting structured information from unstructured text
Content Generation: Creating formatted content with specific schemas
API Integration: Integrating LLM capabilities into existing applications
Automation: Building AI-powered automation workflows

Notebook Information

Kernel: Python 3

Language: python

Cells: 10

Format: v4.5

Code [3]

Cell 1

Input:

!pip install openai

Output:

Collecting openai
  Downloading openai-1.101.0-py3-none-any.whl.metadata (29 kB)
Collecting anyio<5,>=3.5.0 (from openai)
  Using cached anyio-4.10.0-py3-none-any.whl.metadata (4.0 kB)
Collecting distro<2,>=1.7.0 (from openai)
  Using cached distro-1.9.0-py3-none-any.whl.metadata (6.8 kB)
Collecting httpx<1,>=0.23.0 (from openai)
  Using cached httpx-0.28.1-py3-none-any.whl.metadata (7.1 kB)
Collecting jiter<1,>=0.4.0 (from openai)
  Downloading jiter-0.10.0-cp311-cp311-macosx_11_0_arm64.whl.metadata (5.2 kB)
Collecting pydantic<3,>=1.9.0 (from openai)
  Using cached pydantic-2.11.7-py3-none-any.whl.metadata (67 kB)
Collecting sniffio (from openai)
  Using cached sniffio-1.3.1-py3-none-any.whl.metadata (3.9 kB)
Collecting tqdm>4 (from openai)
  Using cached tqdm-4.67.1-py3-none-any.whl.metadata (57 kB)
Requirement already satisfied: typing-extensions<5,>=4.11 in /Users/vkrana/Documents/GL_Knowledge-Graph/.conda/lib/python3.11/site-packages (from openai) (4.14.1)
Collecting idna>=2.8 (from anyio<5,>=3.5.0->openai)
  Using cached idna-3.10-py3-none-any.whl.metadata (10 kB)
Collecting certifi (from httpx<1,>=0.23.0->openai)
  Using cached certifi-2025.8.3-py3-none-any.whl.metadata (2.4 kB)
Collecting httpcore==1.* (from httpx<1,>=0.23.0->openai)
  Using cached httpcore-1.0.9-py3-none-any.whl.metadata (21 kB)
Collecting h11>=0.16 (from httpcore==1.*->httpx<1,>=0.23.0->openai)
  Using cached h11-0.16.0-py3-none-any.whl.metadata (8.3 kB)
Collecting annotated-types>=0.6.0 (from pydantic<3,>=1.9.0->openai)
  Using cached annotated_types-0.7.0-py3-none-any.whl.metadata (15 kB)
Collecting pydantic-core==2.33.2 (from pydantic<3,>=1.9.0->openai)
  Downloading pydantic_core-2.33.2-cp311-cp311-macosx_11_0_arm64.whl.metadata (6.8 kB)
Collecting typing-inspection>=0.4.0 (from pydantic<3,>=1.9.0->openai)
  Using cached typing_inspection-0.4.1-py3-none-any.whl.metadata (2.6 kB)
Downloading openai-1.101.0-py3-none-any.whl (810 kB)
[2K   [90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━[0m [32m810.8/810.8 kB[0m [31m20.1 MB/s[0m eta [36m0:00:00[0m
[?25hUsing cached anyio-4.10.0-py3-none-any.whl (107 kB)
Using cached distro-1.9.0-py3-none-any.whl (20 kB)
Using cached httpx-0.28.1-py3-none-any.whl (73 kB)
Using cached httpcore-1.0.9-py3-none-any.whl (78 kB)
Downloading jiter-0.10.0-cp311-cp311-macosx_11_0_arm64.whl (321 kB)
Using cached pydantic-2.11.7-py3-none-any.whl (444 kB)
Downloading pydantic_core-2.33.2-cp311-cp311-macosx_11_0_arm64.whl (1.9 MB)
[2K   [90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━[0m [32m1.9/1.9 MB[0m [31m29.9 MB/s[0m eta [36m0:00:00[0m
[?25hUsing cached annotated_types-0.7.0-py3-none-any.whl (13 kB)
Using cached h11-0.16.0-py3-none-any.whl (37 kB)
Using cached idna-3.10-py3-none-any.whl (70 kB)
Using cached sniffio-1.3.1-py3-none-any.whl (10 kB)
Using cached tqdm-4.67.1-py3-none-any.whl (78 kB)
Using cached typing_inspection-0.4.1-py3-none-any.whl (14 kB)
Using cached certifi-2025.8.3-py3-none-any.whl (161 kB)
Installing collected packages: typing-inspection, tqdm, sniffio, pydantic-core, jiter, idna, h11, distro, certifi, annotated-types, pydantic, httpcore, anyio, httpx, openai
[2K   [90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━[0m [32m15/15[0m [openai]14/15[0m [openai]c]
[1A[2KSuccessfully installed annotated-types-0.7.0 anyio-4.10.0 certifi-2025.8.3 distro-1.9.0 h11-0.16.0 httpcore-1.0.9 httpx-0.28.1 idna-3.10 jiter-0.10.0 openai-1.101.0 pydantic-2.11.7 pydantic-core-2.33.2 sniffio-1.3.1 tqdm-4.67.1 typing-inspection-0.4.1

Code [13]

Cell 2

Input:

!pip install python-dotenv

Output:

Collecting python-dotenv
  Using cached python_dotenv-1.1.1-py3-none-any.whl.metadata (24 kB)
Using cached python_dotenv-1.1.1-py3-none-any.whl (20 kB)
Installing collected packages: python-dotenv
Successfully installed python-dotenv-1.1.1

Code

Cell 3

Input:

from pydantic import BaseModel

class User(BaseModel):
    name: str
    age: int
    email: str

# Create a valid user
user = User(name="Vijendra", age=30, email="vijendrav@gmail.com")
print(user.model_dump_json())

Output:

{"name":"Vijendrav","age":30,"email":"vijendrav@gmail.com"}

Code [2]

Cell 4

Input:

from openai import OpenAI
from dotenv import load_dotenv
import os
load_dotenv(override=True)

client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))

response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello, how are you?"}]
)

print(response.choices[0].message.content)

Output:

Hello! I'm just a computer program, so I don't have feelings, but I'm here and ready to help you. How can I assist you today?

Code [3]

Cell 5

Input:

## pydentic use cases
from pydantic import BaseModel
from openai import OpenAI
import os
from dotenv import load_dotenv

load_dotenv(override=True)

client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))

class Feedback(BaseModel):
    sentiment: str
    completed_courses: int
    would_recommend: bool

#step 2
completion = client.beta.chat.completions.parse(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "I'm a student who has read all your article and series and rest content. I would definitely recommend this platform to my friends."}],
    response_format=Feedback
)

print(completion.choices[0].message.parsed)

Output:

sentiment='Positive' completed_courses=0 would_recommend=True

Code [19]

Cell 6

Input:

!pip install requests

Output:

Collecting requests
  Downloading requests-2.32.5-py3-none-any.whl.metadata (4.9 kB)
Collecting charset_normalizer<4,>=2 (from requests)
  Downloading charset_normalizer-3.4.3-cp311-cp311-macosx_10_9_universal2.whl.metadata (36 kB)
Requirement already satisfied: idna<4,>=2.5 in /Users/vkrana/Documents/GL_Knowledge-Graph/.conda/lib/python3.11/site-packages (from requests) (3.10)
Collecting urllib3<3,>=1.21.1 (from requests)
  Using cached urllib3-2.5.0-py3-none-any.whl.metadata (6.5 kB)
Requirement already satisfied: certifi>=2017.4.17 in /Users/vkrana/Documents/GL_Knowledge-Graph/.conda/lib/python3.11/site-packages (from requests) (2025.8.3)
Downloading requests-2.32.5-py3-none-any.whl (64 kB)
Downloading charset_normalizer-3.4.3-cp311-cp311-macosx_10_9_universal2.whl (204 kB)
Using cached urllib3-2.5.0-py3-none-any.whl (129 kB)
Installing collected packages: urllib3, charset_normalizer, requests
[2K   [90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━[0m [32m3/3[0m [requests]
[1A[2KSuccessfully installed charset_normalizer-3.4.3 requests-2.32.5 urllib3-2.5.0

Code [4]

Cell 7

Input:

#Tools use
from openai import OpenAI
import os
from dotenv import load_dotenv
import requests
import json

load_dotenv(override=True)

client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))

#Tools for function calling
def get_tempreture(lat, lon):
    response = requests.get(
        f"https://api.open-meteo.com/v1/forecast"
        f"?latitude={lat}&longitude={lon}"
        f"&current=temperature_2m"
    )
    data = response.json()
    return data["current"]["temperature_2m"]

tool_registry = [
    {
        "type": "function",
        "function": {
            "name": "get_tempreture",
            "description": "Get current temperature for provided coordinates in celsius.",
            "parameters": {
                "type": "object",
                "properties": {
                    "latitude": {"type": "number"},
                    "longitude": {"type": "number"},
                },
                "required": ["latitude", "longitude"],
                "additionalProperties": False,
            },
            "strict": True,
        },
    }
]

conversation = [{"role": "user", "content": "Can you check how hot it is in Bangalore right now?"}]
first_response = client.chat.completions.create(
    model="gpt-4.1",
    messages=conversation,
    tools=tool_registry,
)

Code [5]

Cell 8

Input:

tool_suggestion = first_response.choices[0].message.tool_calls[0]
tool_args = json.loads(tool_suggestion.function.arguments)
print("Tool arguments:", tool_args)  # Debug: see what's actually in the arguments
print("Available keys:", tool_args.keys())  # Debug: see what keys exist
temp_result = get_tempreture(tool_args["latitude"], tool_args["longitude"])
print(temp_result)

Output:

Tool arguments: {'latitude': 12.9716, 'longitude': 77.5946}
Available keys: dict_keys(['latitude', 'longitude'])
27.8

Code [6]

Cell 9

Input:

conversation.append({"role": "assistant", "tool_calls": [tool_suggestion]})
conversation.append({
    "role": "tool",
    "tool_call_id": tool_suggestion.id, 
    "content": json.dumps(temp_result)
})

Code [8]

Cell 10

Input:

## structured output
from pydantic import BaseModel, Field
from typing import Literal

class TempReply(BaseModel):
  temperature: float = Field(
        description="Temperature in Celsius at the requested location."
    )
    # 👇 Literal forces an exact match
  message: Literal[
        "Thanks for sticking with my snnipet content! "
        "According to Pydantic validation, here’s your weather update."
    ] = Field(
        description="Canonical thank-you line that must appear verbatim."
    )

completion_final = client.beta.chat.completions.parse(
    model="gpt-4o-mini",
    messages = conversation,
    tools = tool_registry,
    response_format = TempReply
)
final = completion_final.choices[0].message.parsed
print(final.temperature)
print(final.message)

Output:

27.8
Thanks for sticking with my snnipet content! According to Pydantic validation, here’s your weather update.

Back to Snippets