news-structurize

news-structurize is a lightweight Python package that transforms a news headline or a short text snippet into a structured summary. The output contains key elements such as the person involved, the event, and the financial or business impact. It leverages an LLM (by default ChatLLM7) together with regex‑based pattern matching to guarantee a consistent, machine‑readable format—perfect for automated news aggregation, financial reporting, or data‑extraction pipelines.

---

Features

• One‑function API – just call news_structurize() with your text.
• Built‑in LLM – uses ChatLLM7 from the langchain_llm7 package out‑of‑the‑box.
• Pluggable LLMs – pass any LangChain‑compatible chat model (OpenAI, Anthropic, Google, …).
• Regex‑validated output – guarantees that the returned data matches the expected schema.
• Zero‑configuration default – works with the free tier of LLM7; optional API key handling.

---

Installation

pip install news_structurize

---

Quick Start

from news_structurize import news_structurize

# Simple call – uses the default ChatLLM7 (needs LLM7_API_KEY in env or default key)
headline = "Apple CEO Tim Cook announces $2 billion investment in renewable energy"
summary = news_structurize(headline)

print(summary)
# Example output:
# ['Person: Tim Cook', 'Event: Investment announcement', 'Impact: $2 billion in renewable energy']

---

Advanced Usage – Supplying Your Own LLM

You can provide any LangChain chat model that follows the BaseChatModel interface.

OpenAI

from langchain_openai import ChatOpenAI
from news_structurize import news_structurize

llm = ChatOpenAI(model="gpt-4o-mini")
headline = "Tesla reports record Q3 deliveries"
summary = news_structurize(headline, llm=llm)

print(summary)

Anthropic

from langchain_anthropic import ChatAnthropic
from news_structurize import news_structurize

llm = ChatAnthropic(model="claude-3-haiku-20240307")
headline = "Amazon expands grocery footprint with 15 new stores"
summary = news_structurize(headline, llm=llm)

print(summary)

Google Generative AI

from langchain_google_genai import ChatGoogleGenerativeAI
from news_structurize import news_structurize

llm = ChatGoogleGenerativeAI(model="gemini-1.5-flash")
headline = "Microsoft acquires AI startup for $1.2 billion"
summary = news_structurize(headline, llm=llm)

print(summary)

---

API Reference

news_structurize(
    user_input: str,
    llm: Optional[BaseChatModel] = None,
    api_key: Optional[str] = None
) -> List[str]

Parameter	Type	Description
user_input	str	The news headline or short text snippet to be structured.
llm	Optional[BaseChatModel]	A LangChain chat model instance. If omitted, the default ChatLLM7 is used.
api_key	Optional[str]	API key for LLM7. If omitted, the function looks for LLM7_API_KEY in the environment; otherwise a placeholder key ("None") is used (suitable for the free tier).

The function returns a list of strings, each representing a parsed element (e.g., ["Person: …", "Event: …", "Impact: …"]). If the LLM call fails, a RuntimeError is raised with the underlying error message.

---

Configuration & Rate Limits

• LLM7 Free Tier – Adequate for most development and low‑volume production use cases.
• Higher Limits – Obtain a personal API key by registering at https://token.llm7.io/ and set it via the environment variable LLM7_API_KEY or pass it directly to news_structurize().

export LLM7_API_KEY="your_api_key_here"

---

Contributing

Issues, feature requests, and pull requests are welcome! Please file them on the GitHub repository:

🔗 https://github....

When contributing, follow standard Python packaging conventions and keep the public interface limited to the news_structurize function.

---

License

This project is licensed under the MIT License.

---

Author

Eugene Evstafev
📧 Email: hi@euegne.plus
🐙 GitHub: chigwell

---

Acknowledgements

• ChatLLM7 – the default language model, provided by the langchain_llm7 package: https://pypi.org/project/langchain-llm7/
• LangChain – for the unified LLM interface and message handling.