LLM Event Digest

LLM Event Digest is a Python package designed to process news headlines or short text inputs and generate structured summaries of events, such as service disruptions or incidents. Utilizing a language model, it extracts key details like the involved company, the nature of the disruption, and the cause, ensuring outputs conform to a predefined format for consistency and reliability. This tool is ideal for automated news monitoring, alert systems, or data aggregation where structured, error-free information extraction from text is required.

Installation

Install the package via pip:

pip install llm_event_digest

Usage

Here's an example of how to use the package in Python:

from llm_event_digest import llm_event_digest

response = llm_event_digest(
    user_input="The internet service in downtown was down for 3 hours caused by a fiber cut.",
    api_key="your-llm7-api-key"  # Optional, if not set in environment variables
)
print(response)

Parameters

• user_input (str): The text input (news headline or short description) to process.
• llm (Optional[BaseChatModel]): An optional LangChain language model instance. If not provided, the default ChatLLM7 is used.
• api_key (Optional[str]): API key for LLM7. If not provided, it looks for the LLM7_API_KEY environment variable.

Supported LLMs

The package uses ChatLLM7 from langchain_llm7 by default.

You can also pass your own LLM instance, such as:

from langchain_openai import ChatOpenAI

llm = ChatOpenAI()
response = llm_event_digest(
    user_input="Network outage in the city center.",
    llm=llm
)

Or:

from langchain_anthropic import ChatAnthropic

llm = ChatAnthropic()
response = llm_event_digest(
    user_input="Server downtime due to maintenance.",
    llm=llm
)

And:

from langchain_google_genai import ChatGoogleGenerativeAI

llm = ChatGoogleGenerativeAI()
response = llm_event_digest(
    user_input="Scheduled power outage.",
    llm=llm
)

Rate Limits

Default rate limits for LLM7 free tier are suitable for most use cases. For higher usage, obtain an API key from https://token.llm7.io/ and pass it via environment variable LLM7_API_KEY or directly in the function call.

Support and Issues

If you encounter any issues or have questions, please open an issue on the GitHub repository: https://github.com/chigwell/llm-event-digest/issues

Author

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