In this example use case, let's build an AI system that can analyze market sentiment based on provided news articles with explanation using structured output to help investors catch-up with the news easier.
quick_demo.mp4
- Company Identification: Automatically detects publicly traded companies in news content
- Sentiment Classification: Categorizes sentiment as Positive, Negative, or Mixed
- Confidence Scoring: Provides numerical confidence levels as percentage for each analysis
- Sentiment Justification: Provides explanation of how the AI came up with the sentiment label based on your provided content.
- Structured Output: Returns validated JSON data with Pydantic models
git clone https://github.com/The-Smol-Lab/structured-output-tutorial.git
cd structured-output-tutorial
python -m venv .llm_env
cd .llm_env
Scripts/activate (for windows) or source bin/activate (for mac)
cd ..
pip install -r requirements.txtIf you plan to use gemini model:
- Get a Google Gemini API key
- Create
config.pywith:
GEMINI_API_KEY = "your_api_key_here"If you plan to use models loaded in lmstudio, please ensure proper settings as shown below:
Feel free to look up the LM Studio documentation for more info: https://lmstudio.ai/docs/api
# Create prompt template with detailed system message
system_prompt = """
You are a senior financial analyst specializing in news sentiment analysis:
1. Identify all publicly traded companies in the text
2. For each company, determine sentiment using:
- Financial performance metrics (exact figures/percentages)
- Strategic developments (mergers, partnerships)
- Regulatory/legal changes
- Market reactions (stock moves, analyst ratings)
3. Provide confidence scores as decimal values between 0 and 1 (e.g., 0.85 for 85% confidence).
Never use percentage values for confidence scores.
Include specific numerical data, quantified impacts, and precise metrics.
Confidence scores must reflect evidence strength. Never invent information.
"""
user_prompt = """
The current date is {current_date},
analyze the following article and provide sentiment analysis for each publicly traded company
mentioned in the text below:
{article}
"""Run gemini model as default
python main.py
or
python main.py --mode gemini #default to "gemini-2.0-flash-exp"
or
python main.py --mode gemini --model_name "gemini-2.0-flash-exp"Pass in mode and model_name to run a specific model of choice from lmstudio or ollama
python main.py --mode lmstudio --model_name "deepseek-r1-distill-qwen-32b@iq2_s"
python main.py --mode ollama --model_name "deepseek-r1:14b"Model Used: bartowski/deepseek-r1-distill-qwen-32b@iq2_s
Analysis timestamp: 2025-02-01 00:00:00+00:00
Time taken to process the article: 182.87 seconds
**************************************************
Company: Apple Inc.
Sentiment: positive
Confidence: 85%
Justification: Apple reported Q1 sales of $124.3 billion, meeting views while EPS of $2.40 beat estimates by 5
cents. Services revenue rose 13.9%, but iPhone sales declined 1%. Apple stock hit an all-time high
in December but pulled back in January.
**************************************************
Company: Nvidia Corporation
Sentiment: negative
Confidence: 85%
Justification: Nvidia dropped nearly 4% as CEO Jensen Huang met President Trump to discuss export curbs and risks
from DeepSeek. The stock fell below its 200-day moving average after Monday's news of cheaper AI
models.
**************************************************
Currently compatible with:
- All Gemini Models (e.g.,
gemini-2.0-flash-exp,gemini-2.0-flash-thinking-exp-01-21,gemini-exp-1206, and etc.) - Any OpenAI-compatible API endpoint including Ollama/LM Studio
MIT License - See LICENSE for details
Maintained by Krittaprot Tangkittikun.